You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

8588 lines
441 KiB

  1. systemd System and Service Manager
  2. CHANGES WITH 239:
  3. * NETWORK INTERFACE DEVICE NAMING CHANGES: elogind-udevd's "net_id"
  4. builtin will name network interfaces differently than in previous
  5. versions for virtual network interfaces created with SR-IOV and NPAR
  6. and for devices where the PCI network controller device does not have
  7. a slot number associated.
  8. SR-IOV virtual devices are now named based on the name of the parent
  9. interface, with a suffix of "v<N>", where <N> is the virtual device
  10. number. Previously those virtual devices were named as if completely
  11. independent.
  12. The ninth and later NPAR virtual devices will be named following the
  13. scheme used for the first eight NPAR partitions. Previously those
  14. devices were not renamed and the kernel default (eth<n>) was used.
  15. "net_id" will also generate names for PCI devices where the PCI
  16. network controller device does not have an associated slot number
  17. itself, but one of its parents does. Previously those devices were
  18. not renamed and the kernel default (eth<n>) was used.
  19. * AF_INET and AF_INET6 are dropped from RestrictAddressFamilies= in
  20. elogind.service. Since v235, IPAddressDeny=any has been set to
  21. the unit. So, it is expected that the default behavior of
  22. elogind is not changed. However, if distribution packagers or
  23. administrators disabled or modified IPAddressDeny= setting by a
  24. drop-in config file, then it may be necessary to update the file to
  25. re-enable AF_INET and AF_INET6 to support network user name services,
  26. e.g. NIS.
  27. * When the RestrictNamespaces= unit property is specified multiple
  28. times, then the specified types are merged now. Previously, only the
  29. last assignment was used. So, if distribution packagers or
  30. administrators modified the setting by a drop-in config file, then it
  31. may be necessary to update the file.
  32. * When OnFailure= is used in combination with Restart= on a service
  33. unit, then the specified units will no longer be triggered on
  34. failures that result in restarting. Previously, the specified units
  35. would be activated each time the unit failed, even when the unit was
  36. going to be restarted automatically. This behaviour contradicted the
  37. documentation. With this release the code is adjusted to match the
  38. documentation.
  39. * elogind-tmpfiles will now print a notice whenever it encounters
  40. tmpfiles.d/ lines referencing the /var/run/ directory. It will
  41. recommend reworking them to use the /run/ directory instead (for
  42. which /var/run/ is simply a symlinked compatibility alias). This way
  43. elogind-tmpfiles can properly detect line conflicts and merge lines
  44. referencing the same file by two paths, without having to access
  45. them.
  46. * systemctl disable/unmask/preset/preset-all cannot be used with
  47. --runtime. Previously this was allowed, but resulted in unintuitive
  48. behaviour that wasn't useful. systemctl disable/unmask will now undo
  49. both runtime and persistent enablement/masking, i.e. it will remove
  50. any relevant symlinks both in /run and /etc.
  51. * Note that all long-running system services shipped with elogind will
  52. now default to a system call whitelist (rather than a blacklist, as
  53. before). In particular, elogind-udevd will now enforce one too. For
  54. most cases this should be safe, however downstream distributions
  55. which disabled sandboxing of elogind-udevd (specifically the
  56. MountFlags= setting), might want to disable this security feature
  57. too, as the default whitelisting will prohibit all mount, swap,
  58. reboot and clock changing operations from udev rules.
  59. * sd-boot acquired new loader configuration settings to optionally turn
  60. off Windows and MacOS boot partition discovery as well as
  61. reboot-into-firmware menu items. It is also able to pick a better
  62. screen resolution for HiDPI systems, and now provides loader
  63. configuration settings to change the resolution explicitly.
  64. * The elogind-resolve tool has been renamed to resolvectl (it also
  65. * elogind-resolved now supports DNS-over-TLS. It's still
  66. turned off by default, use DNSOverTLS=opportunistic to turn it on in
  67. resolved.conf. We intend to make this the default as soon as couple
  68. of additional techniques for optimizing the initial latency caused by
  69. establishing a TLS/TCP connection are implemented.
  70. * elogind-resolved.service and elogind-networkd.service now set
  71. DynamicUser=yes. The users elogind-resolve and elogind-network are
  72. not created by elogind-sysusers.
  73. remains available under the old name, for compatibility), and its
  74. interface is now verb-based, similar in style to the other <xyz>ctl
  75. tools, such as systemctl or loginctl.
  76. * The resolvectl/elogind-resolve tool also provides 'resolvconf'
  77. compatibility. It may be symlinked under the 'resolvconf' name, in
  78. which case it will take arguments and input compatible with the
  79. Debian and FreeBSD resolvconf tool.
  80. * Support for suspend-then-hibernate has been added, i.e. a sleep mode
  81. where the system initially suspends, and after a time-out resumes and
  82. hibernates again.
  83. * networkd's ClientIdentifier= now accepts a new option "duid-only". If
  84. set the client will only send a DUID as client identifier.
  85. * The nss-elogind glibc NSS module will now enumerate dynamic users and
  86. groups in effect. Previously, it could resolve UIDs/GIDs to user
  87. names/groups and vice versa, but did not support enumeration.
  88. * journald's Compress= configuration setting now optionally accepts a
  89. byte threshold value. All journal objects larger than this threshold
  90. will be compressed, smaller ones will not. Previously this threshold
  91. was not configurable and set to 512.
  92. * A new system.conf setting NoNewPrivileges= is now available which may
  93. be used to turn off acquisition of new privileges system-wide
  94. (i.e. set Linux' PR_SET_NO_NEW_PRIVS for PID 1 itself, and thus also
  95. for all its children). Note that turning this option on means setuid
  96. binaries and file system capabilities lose their special powers.
  97. While turning on this option is a big step towards a more secure
  98. system, doing so is likely to break numerous pre-existing UNIX tools,
  99. in particular su and sudo.
  100. * A new service elogind-time-sync-wait.service has been added. If
  101. enabled it will delay the unit at boot until time
  102. synchronization has been received from the network. This
  103. functionality is useful on systems lacking a local RTC or where it is
  104. acceptable that the boot process shall be delayed by external network
  105. services.
  106. * When hibernating, elogind will now inform the kernel of the image
  107. write offset, on kernels new enough to support this. This means swap
  108. files should work for hibernation now.
  109. * When loading unit files, elogind will now look for drop-in unit files
  110. extensions in additional places. Previously, for a unit file name
  111. "foo-bar-baz.service" it would look for dropin files in
  112. "foo-bar-baz.service.d/*.conf". Now, it will also look in
  113. "foo-bar-.service.d/*.conf" and "foo-.service.d/", i.e. at the
  114. service name truncated after all inner dashes. This scheme allows
  115. writing drop-ins easily that apply to a whole set of unit files at
  116. once. It's particularly useful for mount and slice units (as their
  117. naming is prefix based), but is also useful for service and other
  118. units, for packages that install multiple unit files at once,
  119. following a strict naming regime of beginning the unit file name with
  120. the package's name. Two new specifiers are now supported in unit
  121. files to match this: %j and %J are replaced by the part of the unit
  122. name following the last dash.
  123. * Unit files and other configuration files that support specifier
  124. expansion now understand another three new specifiers: %T and %V will
  125. resolve to /tmp and /var/tmp respectively, or whatever temporary
  126. directory has been set for the calling user. %E will expand to either
  127. /etc (for system units) or $XDG_CONFIG_HOME (for user units).
  128. * The ExecStart= lines of unit files are no longer required to
  129. reference absolute paths. If non-absolute paths are specified the
  130. specified binary name is searched within the service manager's
  131. built-in $PATH, which may be queried with 'elogind-path
  132. search-binaries-default'. It's generally recommended to continue to
  133. use absolute paths for all binaries specified in unit files.
  134. * Units gained a new load state "bad-setting", which is used when a
  135. unit file was loaded, but contained fatal errors which prevent it
  136. from being started (for example, a service unit has been defined
  137. lacking both ExecStart= and ExecStop= lines).
  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
  141. new --debugger= switch or the $SYSTEMD_DEBUGGER environment variable
  142. to pick an alternative debugger instead of the default gdb.
  143. * systemctl and the other tools will now output escape sequences that
  144. generate proper clickable hyperlinks in various terminal emulators
  145. where useful (for example, in the "systemctl status" output you can
  146. now click on the unit file name to quickly open it in the
  147. editor/viewer of your choice). Note that not all terminal emulators
  148. support this functionality yet, but many do. Unfortunately, the
  149. "less" pager doesn't support this yet, hence this functionality is
  150. currently automatically turned off when a pager is started (which
  151. happens quite often due to auto-paging). We hope to remove this
  152. limitation as soon as "less" learns these escape sequences. This new
  153. behaviour may also be turned off explicitly with the $SYSTEMD_URLIFY
  154. environment variable. For details on these escape sequences see:
  156. * networkd's .network files now support a new IPv6MTUBytes= option for
  157. setting the MTU used by IPv6 explicitly as well as a new MTUBytes=
  158. option in the [Route] section to configure the MTU to use for
  159. specific routes. It also gained support for configuration of the DHCP
  160. "UserClass" option through the new UserClass= setting. It gained
  161. three new options in the new [CAN] section for configuring CAN
  162. networks. The MULTICAST and ALLMULTI interface flags may now be
  163. controlled explicitly with the new Multicast= and AllMulticast=
  164. settings.
  165. * networkd will now automatically make use of the kernel's route
  166. expiration feature, if it is available.
  167. * udevd's .link files now support setting the number of receive and
  168. transmit channels, using the RxChannels=, TxChannels=,
  169. OtherChannels=, CombinedChannels= settings.
  170. * Support for UDPSegmentationOffload= has been removed, given its
  171. limited support in hardware, and waning software support.
  172. * networkd's .netdev files now support creating "netdevsim" interfaces.
  173. * PID 1 learnt a new bus call GetUnitByControlGroup() which may be used
  174. to query the unit belonging to a specific kernel control group.
  175. * elogind-analyze gained a new verb "cat-config", which may be used to
  176. dump the contents of any configuration file, with all its matching
  177. drop-in files added in, and honouring the usual search and masking
  178. logic applied to elogind configuration files. For example use
  179. "elogind-analyze cat-config elogind/system.conf" to get the complete
  180. system configuration file of elogind how it would be loaded by PID 1
  181. itself. Similar to this, various tools such as elogind-tmpfiles or
  182. elogind-sysusers, gained a new option "--cat-config", which does the
  183. corresponding operation for their own configuration settings. For
  184. example, "elogind-tmpfiles --cat-config" will now output the full
  185. list of tmpfiles.d/ lines in place.
  186. * timedatectl gained three new verbs: "show" shows bus properties of
  187. elogind-timedated, "timesync-status" shows the current NTP
  188. synchronization state of elogind-timesyncd, and "show-timesync"
  189. shows bus properties of elogind-timesyncd.
  190. * elogind-timesyncd gained a bus interface on which it exposes details
  191. about its state.
  192. * elogind-nspawn gained a new --rlimit= switch for setting initial
  193. * A new environment variable $SYSTEMD_TIMEDATED_NTP_SERVICES is now
  194. understood by elogind-timedated. It takes a colon-separated list of
  195. unit names of NTP client services. The list is used by
  196. "timedatectl set-ntp".
  197. resource limits for the container payload. There's a new switch
  198. --hostname= to explicitly override the container's hostname. A new
  199. --no-new-privileges= switch may be used to control the
  200. PR_SET_NO_NEW_PRIVS flag for the container payload. A new
  201. --oom-score-adjust= switch controls the OOM scoring adjustment value
  202. for the payload. The new --cpu-affinity= switch controls the CPU
  203. affinity of the container payload. The new --resolv-conf= switch
  204. allows more detailed control of /etc/resolv.conf handling of the
  205. container. Similarly, the new --timezone= switch allows more detailed
  206. control of /etc/localtime handling of the container.
  207. * elogind-detect-virt gained a new --list switch, which will print a
  208. list of all currently known VM and container environments.
  209. * Support for "Portable Services" has been added, see
  210. doc/ for details. Currently, the support is still
  211. experimental, but this is expected to change soon. Reflecting this
  212. experimental state, the "portablectl" binary is not installed into
  213. /usr/bin yet. The binary has to be called with the full path
  214. /usr/lib/elogind/portablectl instead.
  215. * journalctl's and systemctl's -o switch now knows a new log output
  216. mode "with-unit". The output it generates is very similar to the
  217. regular "short" mode, but displays the unit name instead of the
  218. syslog tag for each log line. Also, the date is shown with timezone
  219. information. This mode is probably more useful than the classic
  220. "short" output mode for most purposes, except where pixel-perfect
  221. compatibility with classic /var/log/messages formatting is required.
  222. * A new --dump-bus-properties switch has been added to the elogind
  223. binary, which may be used to dump all supported D-Bus properties.
  224. (Options which are still supported, but are deprecated, are *not*
  225. shown.)
  226. * sd-bus gained a set of new calls:
  227. sd_bus_slot_set_floating()/sd_bus_slot_get_floating() may be used to
  228. enable/disable the "floating" state of a bus slot object,
  229. i.e. whether the slot object pins the bus it is allocated for into
  230. memory or if the bus slot object gets disconnected when the bus goes
  231. away. sd_bus_open_with_description(),
  232. sd_bus_open_user_with_description(),
  233. sd_bus_open_system_with_description() may be used to allocate bus
  234. objects and set their description string already during allocation.
  235. * sd-event gained support for watching inotify events from the event
  236. loop, in an efficient way, sharing inotify handles between multiple
  237. users. For this a new function sd_event_add_inotify() has been added.
  238. * sd-event and sd-bus gained support for calling special user-supplied
  239. destructor functions for userdata pointers associated with
  240. sd_event_source, sd_bus_slot, and sd_bus_track objects. For this new
  241. functions sd_bus_slot_set_destroy_callback,
  242. sd_bus_slot_get_destroy_callback, sd_bus_track_set_destroy_callback,
  243. sd_bus_track_get_destroy_callback,
  244. sd_event_source_set_destroy_callback,
  245. sd_event_source_get_destroy_callback have been added.
  246. * The "net.ipv4.tcp_ecn" sysctl will now be turned on by default.
  247. * PID 1 will now automatically reschedule .timer units whenever the
  248. local timezone changes. (They previously got rescheduled
  249. automatically when the system clock changed.)
  250. * New documentation has been added to document cgroups delegation,
  251. portable services and the various code quality tools we have set up:
  255. * The Boot Loader Specification has been added to the source tree.
  257. While moving it into our source tree we have updated it and further
  258. changes are now accepted through the usual github PR workflow.
  259. * pam_elogind will now look for PAM userdata fields elogind.memory_max,
  260. elogind.tasks_max, elogind.cpu_weight, elogind.io_weight set by
  261. earlier PAM modules. The data in these fields is used to initialize
  262. the session scope's resource properties. Thus external PAM modules
  263. may now configure per-session limits, for example sourced from
  264. external user databases.
  265. * socket units with Accept=yes will now maintain a "refused" counter in
  266. addition to the existing "accepted" counter, counting connections
  267. refused due to the enforced limits.
  268. * The "elogind-path search-binaries-default" command may now be use to
  269. query the default, built-in $PATH PID 1 will pass to the services it
  270. manages.
  271. * A new unit file setting PrivateMounts= has been added. It's a boolean
  272. option. If enabled the unit's processes are invoked in their own file
  273. system namespace. Note that this behaviour is also implied if any
  274. other file system namespacing options (such as PrivateTmp=,
  275. PrivateDevices=, ProtectSystem=, …) are used. This option is hence
  276. primarily useful for services that do not use any of the other file
  277. system namespacing options. One such service is elogind-udevd.service
  278. wher this is now used by default.
  279. * ConditionSecurity= gained a new value "uefi-secureboot" that is true
  280. when the system is booted in UEFI "secure mode".
  281. * A new unit "" is added, which defines an
  282. optional synchronization point for offline system updates, as
  283. implemented by the pre-existing "" unit. It
  284. allows ordering services before the service that executes the actual
  285. update process in a generic way.
  286. Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale,
  287. Alexander Kurtz, Alex Gartrell, Anssi Hannula, Arnaud Rebillout, Brian
  288. J. Murrell, Bruno Vernay, Chris Lamb, Chris Lesiak, Christian Brauner,
  289. Christian Hesse, Christian Rebischke, Colin Guthrie, Daniel Dao, Daniel
  290. Lin, Danylo Korostil, Davide Cavalca, David Tardon, Dimitri John
  291. Ledkov, Dmitriy Geels, Douglas Christman, Elia Geretto, emelenas, Emil
  292. Velikov, Evgeny Vereshchagin, Felipe Sateler, Feng Sun, Filipe
  293. Brandenburger, Franck Bui, futpib, Giuseppe Scrivano, Guillem Jover,
  294. guixxx, Hannes Reinecke, Hans de Goede, Harald Hoyer, Henrique Dante de
  295. Almeida, Hiram van Paassen, Ian Miell, Igor Gnatenko, Ivan Shapovalov,
  296. Iwan Timmer, James Cowgill, Jan Janssen, Jan Synacek, Jared Kazimir,
  297. Jérémy Rosen, João Paulo Rechi Vita, Joost Heitbrink, Jui-Chi Ricky
  298. Liang, Jürg Billeter, Kai-Heng Feng, Karol Augustin, Kay Sievers,
  299. Krzysztof Nowicki, Lauri Tirkkonen, Lennart Poettering, Leonard König,
  300. Long Li, Luca Boccassi, Lucas Werkmeister, Marcel Hoppe, Marc
  301. Kleine-Budde, Mario Limonciello, Martin Jansa, Martin Wilck, Mathieu
  302. Malaterre, Matteo F. Vescovi, Matthew McGinn, Matthias-Christian Ott,
  303. Michael Biebl, Michael Olbrich, Michael Prokop, Michal Koutný, Michal
  304. Sekletar, Mike Gilbert, Mikhail Kasimov, Milan Broz, Milan Pässler,
  305. Mladen Pejaković, Muhammet Kara, Nicolas Boichat, Omer Katz, Paride
  306. Legovini, Paul Menzel, Paul Milliken, Pavel Hrdina, Peter A. Bigot,
  307. Peter D'Hoye, Peter Hutterer, Peter Jones, Philip Sequeira, Philip
  308. Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de Araujo,
  309. Ronny Chevalier, Rosen Penev, Rubén Suárez Alvarez, Ryan Gonzalez,
  310. Salvo Tomaselli, Sebastian Reichel, Sergey Ptashnick, Sergio Lindo
  311. Mansilla, Stefan Schweter, Stephen Hemminger, Stuart Hayes, Susant
  312. Sahani, Sylvain Plantefève, Thomas H. P. Andersen, Tobias Jungel,
  313. Tomasz Torcz, Vito Caputo, Will Dietz, Will Thompson, Wim van Mourik,
  314. Yu Watanabe, Zbigniew Jędrzejewski-Szmek
  315. — Berlin, 2018-06-22
  316. CHANGES WITH 238:
  317. * The MemoryAccounting= unit property now defaults to on. After
  318. discussions with the upstream control group maintainers we learnt
  319. that the negative impact of cgroup memory accounting on current
  320. kernels is finally relatively minimal, so that it should be safe to
  321. enable this by default without affecting system performance. Besides
  322. memory accounting only task accounting is turned on by default, all
  323. other forms of resource accounting (CPU, IO, IP) remain off for now,
  324. because it's not clear yet that their impact is small enough to move
  325. from opt-in to opt-out. We recommend downstreams to leave memory
  326. accounting on by default if kernel 4.14 or higher is primarily
  327. used. On very resource constrained systems or when support for old
  328. kernels is a necessity, -Dmemory-accounting-default=false can be used
  329. to revert this change.
  330. * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update,
  331. %udev_rules_update) and the journal catalog (%journal_catalog_update)
  332. from the upgrade scriptlets of individual packages now do nothing.
  333. Transfiletriggers have been added which will perform those updates
  334. once at the end of the transaction.
  335. Similar transfiletriggers have been added to execute any sysctl.d
  336. and binfmt.d rules. Thus, it should be unnecessary to provide any
  337. scriptlets to execute this configuration from package installation
  338. scripts.
  339. * elogind-sysusers gained a mode where the configuration to execute is
  340. specified on the command line, but this configuration is not executed
  341. directly, but instead it is merged with the configuration on disk,
  342. and the result is executed. This is useful for package installation
  343. scripts which want to create the user before installing any files on
  344. disk (in case some of those files are owned by that user), while
  345. still allowing local admin overrides.
  346. This functionality is exposed to rpm scriptlets through a new
  347. %sysusers_create_package macro. Old %sysusers_create and
  348. %sysusers_create_inline macros are deprecated.
  349. A transfiletrigger for sysusers.d configuration is now installed,
  350. which means that it should be unnecessary to call elogind-sysusers from
  351. package installation scripts, unless the package installs any files
  352. owned by those newly-created users, in which case
  353. %sysusers_create_package should be used.
  354. * Analogous change has been done for elogind-tmpfiles: it gained a mode
  355. where the command-line configuration is merged with the configuration
  356. on disk. This is exposed as the new %tmpfiles_create_package macro,
  357. and %tmpfiles_create is deprecated. A transfiletrigger is installed
  358. for tmpfiles.d, hence it should be unnecessary to call elogind-tmpfiles
  359. from package installation scripts.
  360. * sysusers.d configuration for a user may now also specify the group
  361. number, in addition to the user number ("u username 123:456"), or
  362. without the user number ("u username -:456").
  363. * Configution items for elogind-sysusers can now be specified as
  364. positional arguments when the new --inline switch is used.
  365. * The login shell of users created through sysusers.d may now be
  366. specified (previously, it was always /bin/sh for root and
  367. /sbin/nologin for other users).
  368. * elogind-analyze gained a new --global switch to look at global user
  369. configuration. It also gained a unit-paths verb to list the unit load
  370. paths that are compiled into elogind (which can be used with
  371. --elogind, --user, or --global).
  372. * udevadm trigger gained a new --settle/-w option to wait for any
  373. triggered events to finish (but just those, and not any other events
  374. which are triggered meanwhile).
  375. * The action that elogind takes when the lid is closed and the
  376. machine is connected to external power can now be configured using
  377. HandleLidSwitchExternalPower= in logind.conf. Previously, this action
  378. was determined by HandleLidSwitch=, and, for backwards compatibility,
  379. is still is, if HandleLidSwitchExternalPower= is not explicitly set.
  380. * journalctl will periodically call sd_journal_process() to make it
  381. resilient against inotify queue overruns when journal files are
  382. rotated very quickly.
  383. * Two new functions in libelogind — sd_bus_get_n_queued_read and
  384. sd_bus_get_n_queued_write — may be used to check the number of
  385. pending bus messages.
  386. * elogind gained a new
  387. org.freedesktop.elogind1.Manager.AttachProcessesToUnit dbus call
  388. which can be used to migrate foreign processes to scope and service
  389. units. The primary user for this new API is elogind itself: the
  390. elogind --user instance uses this call of the elogind --system
  391. instance to migrate processes if it itself gets the request to
  392. migrate processes and the kernel refuses this due to access
  393. restrictions. Thanks to this "elogind-run --scope --user …" works
  394. again in pure cgroups v2 environments when invoked from the user
  395. session scope.
  396. * A new TemporaryFileSystem= setting can be used to mask out part of
  397. the real file system tree with tmpfs mounts. This may be combined
  398. with BindPaths= and BindReadOnlyPaths= to hide files or directories
  399. not relevant to the unit, while still allowing some paths lower in
  400. the tree to be accessed.
  401. ProtectHome=tmpfs may now be used to hide user home and runtime
  402. directories from units, in a way that is mostly equivalent to
  403. "TemporaryFileSystem=/home /run/user /root".
  404. * Non-service units are now started with KeyringMode=shared by default.
  405. This means that mount and swapon and other mount tools have access
  406. to keys in the main keyring.
  407. * /sys/fs/bpf is now mounted automatically.
  408. * QNX virtualization is now detected by elogind-detect-virt and may
  409. be used in ConditionVirtualization=.
  410. * IPAccounting= may now be enabled also for slice units.
  411. * A new -Dsplit-bin= build configuration switch may be used to specify
  412. whether bin and sbin directories are merged, or if they should be
  413. included separately in $PATH and various listings of executable
  414. directories. The build configuration scripts will try to autodetect
  415. the proper values of -Dsplit-usr= and -Dsplit-bin= based on build
  416. system, but distributions are encouraged to configure this
  417. explicitly.
  418. * A new -Dok-color= build configuration switch may be used to change
  419. the colour of "OK" status messages.
  420. * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with
  421. PrivateNetwork=yes was buggy in previous versions of elogind. This
  422. means that after the upgrade and daemon-reexec, any such units must
  423. be restarted.
  424. * INCOMPATIBILITY: as announced in the NEWS for 237, elogind-tmpfiles
  425. will not exclude read-only files owned by root from cleanup.
  426. Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet,
  427. Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337,
  428. Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo
  429. de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel
  430. Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny
  431. Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib,
  432. Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer,
  433. Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld,
  434. Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas
  435. Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt,
  436. MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren,
  437. Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert
  438. Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
  439. Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain
  440. Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe,
  441. Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić)
  442. — Warsaw, 2018-03-05
  443. CHANGES WITH 237:
  444. * Some keyboards come with a zoom see-saw or rocker which until now got
  445. mapped to the Linux "zoomin/out" keys in hwdb. However, these
  446. keycodes are not recognized by any major desktop. They now produce
  447. Up/Down key events so that they can be used for scrolling.
  448. * INCOMPATIBILITY: elogind-tmpfiles' "f" lines changed behaviour
  449. slightly: previously, if an argument was specified for lines of this
  450. type (i.e. the right-most column was set) this string was appended to
  451. existing files each time elogind-tmpfiles was run. This behaviour was
  452. different from what the documentation said, and not particularly
  453. useful, as repeated elogind-tmpfiles invocations would not be
  454. idempotent and grow such files without bounds. With this release
  455. behaviour has been altered slightly, to match what the documentation
  456. says: lines of this type only have an effect if the indicated files
  457. don't exist yet, and only then the argument string is written to the
  458. file.
  459. * FUTURE INCOMPATIBILITY: In elogind v238 we intend to slightly change
  460. elogind-tmpfiles behaviour: previously, read-only files owned by root
  461. were always excluded from the file "aging" algorithm (i.e. the
  462. automatic clean-up of directories like /tmp based on
  463. atime/mtime/ctime). We intend to drop this restriction, and age files
  464. by default even when owned by root and read-only. This behaviour was
  465. inherited from older tools, but there have been requests to remove
  466. it, and it's not obvious why this restriction was made in the first
  467. place. Please speak up now, if you are aware of software that reqires
  468. this behaviour, otherwise we'll remove the restriction in v238.
  469. * A new environment variable $SYSTEMD_OFFLINE is now understood by
  470. systemctl. It takes a boolean argument. If on, systemctl assumes it
  471. operates on an "offline" OS tree, and will not attempt to talk to the
  472. service manager. Previously, this mode was implicitly enabled if a
  473. chroot() environment was detected, and this new environment variable
  474. now provides explicit control.
  475. * .path and .socket units may now be created transiently, too.
  476. Previously only service, mount, automount and timer units were
  477. supported as transient units. The elogind-run tool has been updated
  478. to expose this new functionality, you may hence use it now to bind
  479. arbitrary commands to path or socket activation on-the-fly from the
  480. command line. Moreover, almost all properties are now exposed for the
  481. unit types that already supported transient operation.
  482. * The elogind-mount command gained support for a new --owner= parameter
  483. which takes a user name, which is then resolved and included in uid=
  484. and gid= mount options string of the file system to mount.
  485. * A new unit condition ConditionControlGroupController= has been added
  486. that checks whether a specific cgroup controller is available.
  487. * Unit files, udev's .link files, and elogind-networkd's .netdev and
  488. .network files all gained support for a new condition
  489. ConditionKernelVersion= for checking against specific kernel
  490. versions.
  491. * In elogind-networkd, the [IPVLAN] section in .netdev files gained
  492. support for configuring device flags in the Flags= setting. In the
  493. same files, the [Tunnel] section gained support for configuring
  494. AllowLocalRemote=. The [Route] section in .network files gained
  495. support for configuring InitialCongestionWindow=,
  496. InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now
  497. understands RapidCommit=.
  498. * elogind-networkd's DHCPv6 support gained support for Prefix
  499. Delegation.
  500. * sd-bus gained support for a new "watch-bind" feature. When this
  501. feature is enabled, an sd_bus connection may be set up to connect to
  502. an AF_UNIX socket in the file system as soon as it is created. This
  503. functionality is useful for writing early-boot services that
  504. automatically connect to the system bus as soon as it is started,
  505. without ugly time-based polling. elogind-networkd and
  506. elogind-resolved have been updated to make use of this
  507. functionality. busctl exposes this functionality in a new
  508. --watch-bind= command line switch.
  509. * sd-bus will now optionally synthesize a local "Connected" signal as
  510. soon as a D-Bus connection is set up fully. This message mirrors the
  511. already existing "Disconnected" signal which is synthesized when the
  512. connection is terminated. This signal is generally useful but
  513. particularly handy in combination with the "watch-bind" feature
  514. described above. Synthesizing of this message has to be requested
  515. explicitly through the new API call sd_bus_set_connected_signal(). In
  516. addition a new call sd_bus_is_ready() has been added that checks
  517. whether a connection is fully set up (i.e. between the "Connected" and
  518. "Disconnected" signals).
  519. * sd-bus gained two new calls sd_bus_request_name_async() and
  520. sd_bus_release_name_async() for asynchronously registering bus
  521. names. Similar, there is now sd_bus_add_match_async() for installing
  522. a signal match asynchronously. All of elogind's own services have
  523. been updated to make use of these calls. Doing these operations
  524. asynchronously has two benefits: it reduces the risk of deadlocks in
  525. case of cyclic dependencies between bus services, and it speeds up
  526. service initialization since synchronization points for bus
  527. round-trips are removed.
  528. * sd-bus gained two new calls sd_bus_match_signal() and
  529. sd_bus_match_signal_async(), which are similar to sd_bus_add_match()
  530. and sd_bus_add_match_async() but instead of taking a D-Bus match
  531. string take match fields as normal function parameters.
  532. * sd-bus gained two new calls sd_bus_set_sender() and
  533. sd_bus_message_set_sender() for setting the sender name of outgoing
  534. messages (either for all outgoing messages or for just one specific
  535. one). These calls are only useful in direct connections as on
  536. brokered connections the broker fills in the sender anyway,
  537. overwriting whatever the client filled in.
  538. * sd-event gained a new pseudo-handle that may be specified on all API
  539. calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When
  540. used this refers to the default event loop object of the calling
  541. thread. Note however that this does not implicitly allocate one —
  542. which has to be done prior by using sd_event_default(). Similarly
  543. sd-bus gained three new pseudo-handles SD_BUS_DEFAULT,
  544. SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer
  545. to the default bus of the specified type of the calling thread. Here
  546. too this does not implicitly allocate bus connection objects, this
  547. has to be done prior with sd_bus_default() and friends.
  548. * sd-event gained a new call pair
  549. sd_event_source_{get|set}_io_fd_own(). This may be used to request
  550. automatic closure of the file descriptor an IO event source watches
  551. when the event source is destroyed.
  552. * elogind-networkd gained support for natively configuring WireGuard
  553. connections.
  554. * In previous versions elogind synthesized user records both for the
  555. "nobody" (UID 65534) and "root" (UID 0) users in nss-elogind and
  556. internally. In order to simplify distribution-wide renames of the
  557. "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a
  558. new transitional flag file has been added: if
  559. /etc/elogind/dont-synthesize-nobody exists synthesizing of the 65534
  560. user and group record within the elogind codebase is disabled.
  561. * elogind-notify gained a new --uid= option for selecting the source
  562. user/UID to use for notification messages sent to the service
  563. manager.
  564. * journalctl gained a new --grep= option to list only entries in which
  565. the message matches a certain pattern. By default matching is case
  566. insensitive if the pattern is lowercase, and case sensitive
  567. otherwise. Option --case-sensitive=yes|no can be used to override
  568. this an specify case sensitivity or case insensitivity.
  569. * There's now a "elogind-analyze service-watchdogs" command for printing
  570. the current state of the service runtime watchdog, and optionally
  571. enabling or disabling the per-service watchdogs system-wide if given a
  572. boolean argument (i.e. the concept you configure in WatchdogSec=), for
  573. debugging purposes. There's also a kernel command line option
  574. elogind.service_watchdogs= for controlling the same.
  575. * Two new "log-level" and "log-target" options for elogind-analyze were
  576. added that merge the now deprecated get-log-level, set-log-level and
  577. get-log-target, set-log-target pairs. The deprecated options are still
  578. understood for backwards compatibility. The two new options print the
  579. current value when no arguments are given, and set them when a
  580. level/target is given as an argument.
  581. * sysusers.d's "u" lines now optionally accept both a UID and a GID
  582. specification, separated by a ":" character, in order to create users
  583. where UID and GID do not match.
  584. Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov,
  585. Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman
  586. Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton
  587. Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov,
  588. Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui,
  589. Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian
  590. Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander
  591. Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen,
  592. Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg
  593. Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering,
  594. Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt,
  595. Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy,
  596. Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał
  597. Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf
  598. Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer,
  599. Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer,
  600. Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani,
  601. Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz
  602. Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary
  603. Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян
  604. Палаузов
  605. — Brno, 2018-01-28
  606. CHANGES WITH 236:
  607. * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
  608. in v235 has been extended to also set the dummy.ko module option
  609. numdummies=0, preventing the kernel from automatically creating
  610. dummy0. All dummy interfaces must now be explicitly created.
  611. * Unknown '%' specifiers in configuration files are now rejected. This
  612. applies to units and tmpfiles.d configuration. Any percent characters
  613. that are followed by a letter or digit that are not supposed to be
  614. interpreted as the beginning of a specifier should be escaped by
  615. doubling ("%%"). (So "size=5%" is still accepted, as well as
  616. "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
  617. valid specifiers today.)
  618. * systemd-resolved now maintains a new dynamic
  619. /run/systemd/resolve/stub-resolv.conf compatibility file. It is
  620. recommended to make /etc/resolv.conf a symlink to it. This file
  621. points at the systemd-resolved stub DNS resolver and
  622. includes dynamically acquired search domains, achieving more correct
  623. DNS resolution by software that bypasses local DNS APIs such as NSS.
  624. * The "uaccess" udev tag has been dropped from /dev/kvm and
  625. /dev/dri/renderD*. These devices now have the 0666 permissions by
  626. default (but this may be changed at build-time). /dev/dri/renderD*
  627. will now be owned by the "render" group along with /dev/kfd.
  628. * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
  629. systemd-journal-gatewayd.service and
  630. systemd-journal-upload.service. This means "nss-systemd" must be
  631. enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
  632. services are resolved properly.
  633. * In /etc/fstab two new mount options are now understood:
  634. x-systemd.makefs and x-systemd.growfs. The former has the effect that
  635. the configured file system is formatted before it is mounted, the
  636. latter that the file system is resized to the full block device size
  637. after it is mounted (i.e. if the file system is smaller than the
  638. partition it resides on, it's grown). This is similar to the fsck
  639. logic in /etc/fstab, and pulls in systemd-makefs@.service and
  640. systemd-growfs@.service as necessary, similar to
  641. systemd-fsck@.service. Resizing is currently only supported on ext4
  642. and btrfs.
  643. * In systemd-networkd, the IPv6 RA logic now optionally may announce
  644. DNS server and domain information.
  645. * Support for the LUKS2 on-disk format for encrypted partitions has
  646. been added. This requires libcryptsetup2 during compilation and
  647. runtime.
  648. * The systemd --user instance will now signal "readiness" when its
  649. unit has been reached, instead of when the run queue ran
  650. empty for the first time.
  651. * Tmpfiles.d with user configuration are now also supported.
  652. systemd-tmpfiles gained a new --user switch, and snippets placed in
  653. ~/.config/user-tmpfiles.d/ and corresponding directories will be
  654. executed by systemd-tmpfiles --user running in the new
  655. systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
  656. running in the user session.
  657. * Unit files and tmpfiles.d snippets learnt three new % specifiers:
  658. %S resolves to the top-level state directory (/var/lib for the system
  659. instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
  660. top-level cache directory (/var/cache for the system instance,
  661. $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
  662. logs directory (/var/log for the system instance,
  663. $XDG_CONFIG_HOME/log/ for the user instance). This matches the
  664. existing %t specifier, that resolves to the top-level runtime
  665. directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
  666. user instance).
  667. * journalctl learnt a new parameter --output-fields= for limiting the
  668. set of journal fields to output in verbose and JSON output modes.
  669. * systemd-timesyncd's configuration file gained a new option
  670. RootDistanceMaxSec= for setting the maximum root distance of servers
  671. it'll use, as well as the new options PollIntervalMinSec= and
  672. PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
  673. * bootctl gained a new command "list" for listing all available boot
  674. menu items on systems that follow the boot loader specification.
  675. * systemctl gained a new --dry-run switch that shows what would be done
  676. instead of doing it, and is currently supported by the shutdown and
  677. sleep verbs.
  678. * ConditionSecurity= can now detect the TOMOYO security module.
  679. * Unit file [Install] sections are now also respected in unit drop-in
  680. files. This is intended to be used by drop-ins under /usr/lib/.
  681. * systemd-firstboot may now also set the initial keyboard mapping.
  682. * Udev "changed" events for devices which are exposed as systemd
  683. .device units are now propagated to units specified in
  684. ReloadPropagatedFrom= as reload requests.
  685. * If a udev device has a SYSTEMD_WANTS= property containing a systemd
  686. unit template name (i.e. a name in the form of 'foobar@.service',
  687. without the instance component between the '@' and - the '.'), then
  688. the escaped sysfs path of the device is automatically used as the
  689. instance.
  690. * SystemCallFilter= in unit files has been extended so that an "errno"
  691. can be specified individually for each system call. Example:
  692. SystemCallFilter=~uname:EILSEQ.
  693. * The cgroup delegation logic has been substantially updated. Delegate=
  694. now optionally takes a list of controllers (instead of a boolean, as
  695. before), which lists the controllers to delegate at least.
  696. * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
  697. * A new LogLevelMax= setting configures the maximum log level any
  698. process of the service may log at (i.e. anything with a lesser
  699. priority than what is specified is automatically dropped). A new
  700. LogExtraFields= setting allows configuration of additional journal
  701. fields to attach to all log records generated by any of the unit's
  702. processes.
  703. * New StandardInputData= and StandardInputText= settings along with the
  704. new option StandardInput=data may be used to configure textual or
  705. binary data that shall be passed to the executed service process via
  706. standard input, encoded in-line in the unit file.
  707. * StandardInput=, StandardOutput= and StandardError= may now be used to
  708. connect stdin/stdout/stderr of executed processes directly with a
  709. file or AF_UNIX socket in the file system, using the new "file:" option.
  710. * A new unit file option CollectMode= has been added, that allows
  711. tweaking the garbage collection logic for units. It may be used to
  712. tell systemd to garbage collect units that have failed automatically
  713. (normally it only GCs units that exited successfully). systemd-run
  714. and systemd-mount expose this new functionality with a new -G option.
  715. * "machinectl bind" may now be used to bind mount non-directories
  716. (i.e. regularfiles, devices, fifos, sockets).
  717. * systemd-analyze gained a new verb "calendar" for validating and
  718. testing calendar time specifications to use for OnCalendar= in timer
  719. units. Besides validating the expression it will calculate the next
  720. time the specified expression would elapse.
  721. * In addition to the pre-existing FailureAction= unit file setting
  722. there's now SuccessAction=, for configuring a shutdown action to
  723. execute when a unit completes successfully. This is useful in
  724. particular inside containers that shall terminate after some workload
  725. has been completed. Also, both options are now supported for all unit
  726. types, not just services.
  727. * networkds's IP rule support gained two new options
  728. IncomingInterface= and OutgoingInterface= for configuring the incoming
  729. and outgoing interfaces of configured rules. systemd-networkd also
  730. gained support for "vxcan" network devices.
  731. * networkd gained a new setting RequiredForOnline=, taking a
  732. boolean. If set, systemd-wait-online will take it into consideration
  733. when determining that the system is up, otherwise it will ignore the
  734. interface for this purpose.
  735. * The sd_notify() protocol gained support for a new operation: with
  736. FDSTOREREMOVE=1 file descriptors may be removed from the per-service
  737. store again, ahead of POLLHUP or POLLERR when they are removed
  738. anyway.
  739. * A new document doc/ has been added to the source tree,
  740. that documents the UID/GID range and assignment assumptions and
  741. requirements of systemd.
  742. * The watchdog device PID 1 will ping may now be configured through the
  743. WatchdogDevice= configuration file setting, or by setting the
  744. systemd.watchdog_service= kernel commandline option.
  745. * systemd-resolved's gained support for registering DNS-SD services on
  746. the local network using MulticastDNS. Services may either be
  747. registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
  748. the same dir below /run, /usr/lib), or through its D-Bus API.
  749. * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
  750. extend the effective start, runtime, and stop time. The service must
  751. continue to send EXTEND_TIMEOUT_USEC within the period specified to
  752. prevent the service manager from making the service as timedout.
  753. * elogind-resolved's DNSSEC support gained support for RFC 8080
  754. (Ed25519 keys and signatures).
  755. * The elogind-resolve command line tool gained a new set of options
  756. --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
  757. --set-nta= and --revert to configure per-interface DNS configuration
  758. dynamically during runtime. It's useful for pushing DNS information
  759. into elogind-resolved from DNS hook scripts that various interface
  760. managing software supports (such as pppd).
  761. * elogind-nspawn gained a new --network-namespace-path= command line
  762. option, which may be used to make a container join an existing
  763. network namespace, by specifying a path to a "netns" file.
  764. Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
  765. Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
  766. Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
  767. Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
  768. John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
  769. Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
  770. Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
  771. Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
  772. Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
  773. Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
  774. Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
  775. Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
  776. Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
  777. Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
  778. Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
  779. Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
  780. Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
  781. Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
  782. Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
  783. Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
  784. Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
  785. Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
  786. Jędrzejewski-Szmek, Zeal Jagannatha
  787. — Berlin, 2017-12-14
  788. CHANGES WITH 235:
  789. * INCOMPATIBILITY: systemd-logind.service and other long-running
  790. services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
  791. communication with the outside. This generally improves security of
  792. the system, and is in almost all cases a safe and good choice, as
  793. these services do not and should not provide any network-facing
  794. functionality. However, systemd-logind uses the glibc NSS API to
  795. query the user database. This creates problems on systems where NSS
  796. is set up to directly consult network services for user database
  797. lookups. In particular, this creates incompatibilities with the
  798. "nss-nis" module, which attempts to directly contact the NIS/YP
  799. network servers it is configured for, and will now consistently
  800. fail. In such cases, it is possible to turn off IP sandboxing for
  801. systemd-logind.service (set IPAddressDeny= in its [Service] section
  802. to the empty string, via a .d/ unit file drop-in). Downstream
  803. distributions might want to update their nss-nis packaging to include
  804. such a drop-in snippet, accordingly, to hide this incompatibility
  805. from the user. Another option is to make use of glibc's nscd service
  806. to proxy such network requests through a privilege-separated, minimal
  807. local caching daemon, or to switch to more modern technologies such
  808. sssd, whose NSS hook-ups generally do not involve direct network
  809. access. In general, we think it's definitely time to question the
  810. implementation choices of nss-nis, i.e. whether it's a good idea
  811. today to embed a network-facing loadable module into all local
  812. processes that need to query the user database, including the most
  813. trivial and benign ones, such as "ls". For more details about
  814. IPAddressDeny= see below.
  815. * A new modprobe.d drop-in is now shipped by default that sets the
  816. bonding module option max_bonds=0. This overrides the kernel default,
  817. to avoid conflicts and ambiguity as to whether or not bond0 should be
  818. managed by systemd-networkd or not. This resolves multiple issues
  819. with bond0 properties not being applied, when bond0 is configured
  820. with systemd-networkd. Distributors may choose to not package this,
  821. however in that case users will be prevented from correctly managing
  822. bond0 interface using systemd-networkd.
  823. * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
  824. which print the logging level and target of the system manager. They
  825. complement the existing "set-log-level" and "set-log-target" verbs
  826. used to change those values.
  827. * journald.conf gained a new boolean setting ReadKMsg= which defaults
  828. to on. If turned off kernel log messages will not be read by
  829. systemd-journald or included in the logs. It also gained a new
  830. setting LineMax= for configuring the maximum line length in
  831. STDOUT/STDERR log streams. The new default for this value is 48K, up
  832. from the previous hardcoded 2048.
  833. * A new unit setting RuntimeDirectoryPreserve= has been added, which
  834. allows more detailed control of what to do with a runtime directory
  835. configured with RuntimeDirectory= (i.e. a directory below /run or
  836. $XDG_RUNTIME_DIR) after a unit is stopped.
  837. * The RuntimeDirectory= setting for units gained support for creating
  838. deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
  839. one top-level directory.
  840. * Units gained new options StateDirectory=, CacheDirectory=,
  841. LogsDirectory= and ConfigurationDirectory= which are closely related
  842. to RuntimeDirectory= but manage per-service directories below
  843. /var/lib, /var/cache, /var/log and /etc. By making use of them it is
  844. possible to write unit files which when activated automatically gain
  845. properly owned service specific directories in these locations, thus
  846. making unit files self-contained and increasing compatibility with
  847. stateless systems and factory reset where /etc or /var are
  848. unpopulated at boot. Matching these new settings there's also
  849. StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
  850. ConfigurationDirectoryMode= for configuring the access mode of these
  851. directories. These settings are particularly useful in combination
  852. with DynamicUser=yes as they provide secure, properly-owned,
  853. writable, and stateful locations for storage, excluded from the
  854. sandbox that such services live in otherwise.
  855. * Automake support has been removed from this release. systemd is now
  856. Meson-only.
  857. * systemd-journald will now aggressively cache client metadata during
  858. runtime, speeding up log write performance under pressure. This comes
  859. at a small price though: as much of the metadata is read
  860. asynchronously from /proc/ (and isn't implicitly attached to log
  861. datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
  862. metadata stored alongside a log entry might be slightly
  863. out-of-date. Previously it could only be slightly newer than the log
  864. message. The time window is small however, and given that the kernel
  865. is unlikely to be improved anytime soon in this regard, this appears
  866. acceptable to us.
  867. * nss-myhostname/systemd-resolved will now by default synthesize an
  868. A/AAAA resource record for the "_gateway" hostname, pointing to the
  869. current default IP gateway. Previously it did that for the "gateway"
  870. name, hampering adoption, as some distributions wanted to leave that
  871. host name open for local use. The old behaviour may still be
  872. requested at build time.
  873. * systemd-networkd's [Address] section in .network files gained a new
  874. Scope= setting for configuring the IP address scope. The [Network]
  875. section gained a new boolean setting ConfigureWithoutCarrier= that
  876. tells systemd-networkd to ignore link sensing when configuring the
  877. device. The [DHCP] section gained a new Anonymize= boolean option for
  878. turning on a number of options suggested in RFC 7844. A new
  879. [RoutingPolicyRule] section has been added for configuring the IP
  880. routing policy. The [Route] section has gained support for a new
  881. Type= setting which permits configuring
  882. blackhole/unreachable/prohibit routes.
  883. * The [VRF] section in .netdev files gained a new Table= setting for
  884. configuring the routing table to use. The [Tunnel] section gained a
  885. new Independent= boolean field for configuring tunnels independent of
  886. an underlying network interface. The [Bridge] section gained a new
  887. GroupForwardMask= option for configuration of propagation of link
  888. local frames between bridge ports.
  889. * The WakeOnLan= setting in .link files gained support for a number of
  890. new modes. A new TCP6SegmentationOffload= setting has been added for
  891. configuring TCP/IPv6 hardware segmentation offload.
  892. * The IPv6 RA sender implementation may now optionally send out RDNSS
  893. and RDNSSL records to supply DNS configuration to peers.
  894. * systemd-nspawn gained support for a new --system-call-filter= command
  895. line option for adding and removing entries in the default system
  896. call filter it applies. Moreover systemd-nspawn has been changed to
  897. implement a system call whitelist instead of a blacklist.
  898. * systemd-run gained support for a new --pipe command line option. If
  899. used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
  900. are directly passed on to the activated transient service
  901. executable. This allows invoking arbitrary processes as systemd
  902. services (for example to take benefit of dependency management,
  903. accounting management, resource management or log management that is
  904. done automatically for services) — while still allowing them to be
  905. integrated in a classic UNIX shell pipeline.
  906. * When a service sends RELOAD=1 via sd_notify() and reload propagation
  907. using ReloadPropagationTo= is configured, a reload is now propagated
  908. to configured units. (Previously this was only done on explicitly
  909. requested reloads, using "systemctl reload" or an equivalent
  910. command.)
  911. * For each service unit a restart counter is now kept: it is increased
  912. each time the service is restarted due to Restart=, and may be
  913. queried using "systemctl show -p NRestarts …".
  914. * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
  915. @signal and @timer have been added, for usage with SystemCallFilter=
  916. in unit files and the new --system-call-filter= command line option
  917. of systemd-nspawn (see above).
  918. * ExecStart= lines in unit files gained two new modifiers: when a
  919. command line is prefixed with "!" the command will be executed as
  920. configured, except for the credentials applied by
  921. setuid()/setgid()/setgroups(). It is very similar to the pre-existing
  922. "+", but does still apply namespacing options unlike "+". There's
  923. also "!!" now, which is mostly identical, but becomes a NOP on
  924. systems that support ambient capabilities. This is useful to write
  925. unit files that work with ambient capabilities where possible but
  926. automatically fall back to traditional privilege dropping mechanisms
  927. on systems where this is not supported.
  928. * ListenNetlink= settings in socket units now support RDMA netlink
  929. sockets.
  930. * A new unit file setting LockPersonality= has been added which permits
  931. locking down the chosen execution domain ("personality") of a service
  932. during runtime.
  933. * A new special target "" has been added, which is
  934. ordered before all text logins, and may be used to order services
  935. before textual logins acquire access to the console.
  936. * systemd will now attempt to load the virtio-rng.ko kernel module very
  937. early on if a VM environment supporting this is detected. This should
  938. improve entropy during early boot in virtualized environments.
  939. * A _netdev option is now supported in /etc/crypttab that operates in a
  940. similar way as the same option in /etc/fstab: it permits configuring
  941. encrypted devices that need to be ordered after the network is up.
  942. Following this logic, two new special targets
  943. and have been
  944. added that are to what and
  945. are to
  946. * Service units gained a new UnsetEnvironment= setting which permits
  947. unsetting specific environment variables for services that are
  948. normally passed to it (for example in order to mask out locale
  949. settings for specific services that can't deal with it).
  950. * Units acquired a new boolean option IPAccounting=. When turned on, IP
  951. traffic accounting (packet count as well as byte count) is done for
  952. the service, and shown as part of "systemctl status" or "systemd-run
  953. --wait".
  954. * Service units acquired two new options IPAddressAllow= and
  955. IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
  956. for configuring a simple IP access control list for all sockets of
  957. the unit. These options are available also on .slice and .socket
  958. units, permitting flexible access list configuration for individual
  959. services as well as groups of services (as defined by a slice unit),
  960. including system-wide. Note that IP ACLs configured this way are
  961. enforced on every single IPv4 and IPv6 socket created by any process
  962. of the service unit, and apply to ingress as well as egress traffic.
  963. * If CPUAccounting= or IPAccounting= is turned on for a unit a new
  964. structured log message is generated each time the unit is stopped,
  965. containing information about the consumed resources of this
  966. invocation.
  967. * A new setting KeyringMode= has been added to unit files, which may be
  968. used to control how the kernel keyring is set up for executed
  969. processes.
  970. * "systemctl poweroff", "systemctl reboot", "systemctl halt",
  971. "systemctl kexec" and "systemctl exit" are now always asynchronous in
  972. behaviour (that is: these commands return immediately after the
  973. operation was enqueued instead of waiting for the operation to
  974. complete). Previously, "systemctl poweroff" and "systemctl reboot"
  975. were asynchronous on systems using systemd-logind (i.e. almost
  976. always, and like they were on sysvinit), and the other three commands
  977. were unconditionally synchronous. With this release this is cleaned
  978. up, and callers will see the same asynchronous behaviour on all
  979. systems for all five operations.
  980. * systemd-logind gained new Halt() and CanHalt() bus calls for halting
  981. the system.
  982. * .timer units now accept calendar specifications in other timezones
  983. than UTC or the local timezone.
  984. * The tmpfiles snippet var.conf has been changed to create
  985. /var/log/btmp with access mode 0660 instead of 0600. It was owned by
  986. the "utmp" group already, and it appears to be generally understood
  987. that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
  988. databases. Previously this was implemented correctly for all these
  989. databases excepts btmp, which has been opened up like this now
  990. too. Note that while the other databases are world-readable
  991. (i.e. 0644), btmp is not and remains more restrictive.
  992. * The systemd-resolve tool gained a new --reset-server-features
  993. switch. When invoked like this systemd-resolved will forget
  994. everything it learnt about the features supported by the configured
  995. upstream DNS servers, and restarts the feature probing logic on the
  996. next resolver look-up for them at the highest feature level
  997. again.
  998. * The status dump systemd-resolved sends to the logs upon receiving
  999. SIGUSR1 now also includes information about all DNS servers it is
  1000. configured to use, and the features levels it probed for them.
  1001. Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
  1002. Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
  1003. Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
  1004. Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
  1005. Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
  1006. Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
  1007. ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
  1008. Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
  1009. Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
  1010. John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
  1011. Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
  1012. Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
  1013. Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
  1014. Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
  1015. Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
  1016. Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
  1017. Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
  1018. Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
  1019. Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
  1020. Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
  1021. — Berlin, 2017-10-06
  1022. CHANGES WITH 234:
  1023. * Meson is now supported as build system in addition to Automake. It is
  1024. our plan to remove Automake in one of our next releases, so that
  1025. Meson becomes our exclusive build system. Hence, please start using
  1026. the Meson build system in your downstream packaging. There's plenty
  1027. of documentation around how to use Meson, the extremely brief
  1028. summary:
  1029. ./ && ./configure && make && sudo make install
  1030. becomes:
  1031. meson build && ninja -C build && sudo ninja -C build install
  1032. * Unit files gained support for a new JobRunningTimeoutUSec= setting,
  1033. which permits configuring a timeout on the time a job is
  1034. running. This is particularly useful for setting timeouts on jobs for
  1035. .device units.
  1036. * Unit files gained two new options ConditionUser= and ConditionGroup=
  1037. for conditionalizing units based on the identity of the user/group
  1038. running a systemd user instance.
  1039. * systemd-networkd now understands a new FlowLabel= setting in the
  1040. [VXLAN] section of .network files, as well as a Priority= in
  1041. [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
  1042. and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
  1043. gained support for configuration of GENEVE links, and IPv6 address
  1044. labels. The [Network] section gained the new IPv6ProxyNDP= setting.
  1045. * .link files now understand a new Port= setting.
  1046. * systemd-networkd's DHCP support gained support for DHCP option 119
  1047. (domain search list).
  1048. * systemd-networkd gained support for serving IPv6 address ranges using
  1049. the Router Advertisement protocol. The new .network configuration
  1050. section [IPv6Prefix] may be used to configure the ranges to
  1051. serve. This is implemented based on a new, minimal, native server
  1052. implementation of RA.
  1053. * journalctl's --output= switch gained support for a new parameter
  1054. "short-iso-precise" for a mode where timestamps are shown as precise
  1055. ISO date values.
  1056. * systemd-udevd's "net_id" builtin may now generate stable network
  1057. interface names from IBM PowerVM VIO devices as well as ACPI platform
  1058. devices.
  1059. * MulticastDNS support in systemd-resolved may now be explicitly
  1060. enabled/disabled using the new MulticastDNS= configuration file
  1061. option.
  1062. * systemd-resolved may now optionally use libidn2 instead of the libidn
  1063. for processing internationalized domain names. Support for libidn2
  1064. should be considered experimental and should not be enabled by
  1065. default yet.
  1066. * "machinectl pull-tar" and related call may now do verification of
  1067. downloaded images using SUSE-style .sha256 checksum files in addition
  1068. to the already existing support for validating using Ubuntu-style
  1069. SHA256SUMS files.
  1070. * sd-bus gained support for a new sd_bus_message_appendv() call which
  1071. is va_list equivalent of sd_bus_message_append().
  1072. * sd-boot gained support for validating images using SHIM/MOK.
  1073. * The SMACK code learnt support for "onlycap".
  1074. * systemd-mount --umount is now much smarter in figuring out how to
  1075. properly unmount a device given its mount or device path.
  1076. * The code to call libnss_dns as a fallback from libnss_resolve when
  1077. the communication with systemd-resolved fails was removed. This
  1078. fallback was redundant and interfered with the [!UNAVAIL=return]
  1079. suffix. See nss-resolve(8) for the recommended configuration.
  1080. * systemd-logind may now be restarted without losing state. It stores
  1081. the file descriptors for devices it manages in the system manager
  1082. using the FDSTORE= mechanism. Please note that further changes in
  1083. other components may be required to make use of this (for example
  1084. Xorg has code to listen for stops of systemd-logind and terminate
  1085. itself when logind is stopped or restarted, in order to avoid using
  1086. stale file descriptors for graphical devices, which is now
  1087. counterproductive and must be reverted in order for restarts of
  1088. systemd-logind to be safe. See
  1090. * All kernel install plugins are called with the environment variable
  1091. KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
  1092. /etc/machine-id. If the file is missing or empty, the variable is
  1093. empty and BOOT_DIR_ABS is the path of a temporary directory which is
  1094. removed after all the plugins exit. So, if KERNEL_INSTALL_MACHINE_ID
  1095. is empty, all plugins should not put anything in BOOT_DIR_ABS.
  1096. Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
  1097. Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
  1098. Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
  1099. Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
  1100. Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
  1101. Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
  1102. Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
  1103. Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
  1104. Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
  1105. Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
  1106. hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
  1107. Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
  1108. Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
  1109. Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
  1110. Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
  1111. Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
  1112. Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
  1113. Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
  1114. Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
  1115. Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
  1116. Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
  1117. Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
  1118. Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
  1119. Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
  1120. Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
  1121. H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
  1122. Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
  1123. userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
  1124. Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
  1125. Георгиевски
  1126. — Berlin, 2017-07-12
  1127. CHANGES WITH 233:
  1128. * The "hybrid" control group mode has been modified to improve
  1129. compatibility with "legacy" cgroups-v1 setups. Specifically, the
  1130. "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
  1131. "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
  1132. cgroups-v1 hierarchy), the only externally visible change being that
  1133. the cgroups-v2 hierarchy is also mounted, to
  1134. /sys/fs/cgroup/unified. This should provide a large degree of
  1135. compatibility with "legacy" cgroups-v1, while taking benefit of the
  1136. better management capabilities of cgroups-v2.
  1137. * The default control group setup mode may be selected both a boot-time
  1138. via a set of kernel command line parameters (specifically:
  1139. systemd.unified_cgroup_hierarchy= and
  1140. systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
  1141. default selected on the configure command line
  1142. (--with-default-hierarchy=). The upstream default is "hybrid"
  1143. (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
  1144. this will change in a future systemd version to be "unified" (pure
  1145. cgroups-v2 mode). The third option for the compile time option is
  1146. "legacy", to enter pure cgroups-v1 mode. We recommend downstream
  1147. distributions to default to "hybrid" mode for release distributions,
  1148. starting with v233. We recommend "unified" for development
  1149. distributions (specifically: distributions such as Fedora's rawhide)
  1150. as that's where things are headed in the long run. Use "legacy" for
  1151. greatest stability and compatibility only.
  1152. * Note one current limitation of "unified" and "hybrid" control group
  1153. setup modes: the kernel currently does not permit the systemd --user
  1154. instance (i.e. unprivileged code) to migrate processes between two
  1155. disconnected cgroup subtrees, even if both are managed and owned by
  1156. the user. This effectively means "systemd-run --user --scope" doesn't
  1157. work when invoked from outside of any "systemd --user" service or
  1158. scope. Specifically, it is not supported from session scopes. We are
  1159. working on fixing this in a future systemd version. (See #3388 for
  1160. further details about this.)
  1161. * DBus policy files are now installed into /usr rather than /etc. Make
  1162. sure your system has dbus >= 1.9.18 running before upgrading to this
  1163. version, or override the install path with --with-dbuspolicydir= .
  1164. * All python scripts shipped with systemd (specifically: the various
  1165. tests written in Python) now require Python 3.
  1166. * systemd unit tests can now run standalone (without the source or
  1167. build directories), and can be installed into /usr/lib/systemd/tests/
  1168. with 'make install-tests'.
  1169. * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
  1170. CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
  1171. kernel.
  1172. * Support for the %c, %r, %R specifiers in unit files has been
  1173. removed. Specifiers are not supposed to be dependent on configuration
  1174. in the unit file itself (so that they resolve the same regardless
  1175. where used in the unit files), but these specifiers were influenced
  1176. by the Slice= option.
  1177. * The shell invoked by debug-shell.service now defaults to /bin/sh in
  1178. all cases. If distributions want to use a different shell for this
  1179. purpose (for example Fedora's /sbin/sushell) they need to specify
  1180. this explicitly at configure time using --with-debug-shell=.
  1181. * The confirmation spawn prompt has been reworked to offer the
  1182. following choices:
  1183. (c)ontinue, proceed without asking anymore
  1184. (D)ump, show the state of the unit
  1185. (f)ail, don't execute the command and pretend it failed
  1186. (h)elp
  1187. (i)nfo, show a short summary of the unit
  1188. (j)obs, show jobs that are in progress
  1189. (s)kip, don't execute the command and pretend it succeeded
  1190. (y)es, execute the command
  1191. The 'n' choice for the confirmation spawn prompt has been removed,
  1192. because its meaning was confusing.
  1193. The prompt may now also be redirected to an alternative console by
  1194. specifying the console as parameter to systemd.confirm_spawn=.
  1195. * Services of Type=notify require a READY=1 notification to be sent
  1196. during startup. If no such message is sent, the service now fails,
  1197. even if the main process exited with a successful exit code.
  1198. * Services that fail to start up correctly now always have their
  1199. ExecStopPost= commands executed. Previously, they'd enter "failed"
  1200. state directly, without executing these commands.
  1201. * The option MulticastDNS= of network configuration files has acquired
  1202. an actual implementation. With MulticastDNS=yes a host can resolve
  1203. names of remote hosts and reply to mDNS A and AAAA requests.
  1204. * When units are about to be started an additional check is now done to
  1205. ensure that all dependencies of type BindsTo= (when used in
  1206. combination with After=) have been started.
  1207. * systemd-analyze gained a new verb "syscall-filter" which shows which
  1208. system call groups are defined for the SystemCallFilter= unit file
  1209. setting, and which system calls they contain.
  1210. * A new system call filter group "@filesystem" has been added,
  1211. consisting of various file system related system calls. Group
  1212. "@reboot" has been added, covering reboot, kexec and shutdown related
  1213. calls. Finally, group "@swap" has been added covering swap
  1214. configuration related calls.
  1215. * A new unit file option RestrictNamespaces= has been added that may be
  1216. used to restrict access to the various process namespace types the
  1217. Linux kernel provides. Specifically, it may be used to take away the
  1218. right for a service unit to create additional file system, network,
  1219. user, and other namespaces. This sandboxing option is particularly
  1220. relevant due to the high amount of recently discovered namespacing
  1221. related vulnerabilities in the kernel.
  1222. * systemd-udev's .link files gained support for a new AutoNegotiation=
  1223. setting for configuring Ethernet auto-negotiation.
  1224. * systemd-networkd's .network files gained support for a new
  1225. ListenPort= setting in the [DHCP] section to explicitly configure the
  1226. UDP client port the DHCP client shall listen on.
  1227. * .network files gained a new Unmanaged= boolean setting for explicitly
  1228. excluding one or more interfaces from management by systemd-networkd.
  1229. * The systemd-networkd ProxyARP= option has been renamed to
  1230. IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
  1231. renamed to ReduceARPProxy=. The old names continue to be available
  1232. for compatibility.
  1233. * systemd-networkd gained support for configuring IPv6 Proxy NDP
  1234. addresses via the new IPv6ProxyNDPAddress= .network file setting.
  1235. * systemd-networkd's bonding device support gained support for two new
  1236. configuration options ActiveSlave= and PrimarySlave=.
  1237. * The various options in the [Match] section of .network files gained
  1238. support for negative matching.
  1239. * New systemd-specific mount options are now understood in /etc/fstab:
  1240. x-systemd.mount-timeout= may be used to configure the maximum
  1241. permitted runtime of the mount command.
  1242. x-systemd.device-bound may be set to bind a mount point to its
  1243. backing device unit, in order to automatically remove a mount point
  1244. if its backing device is unplugged. This option may also be
  1245. configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
  1246. on the block device, which is now automatically set for all CDROM
  1247. drives, so that mounted CDs are automatically unmounted when they are
  1248. removed from the drive.
  1249. x-systemd.after= and x-systemd.before= may be used to explicitly
  1250. order a mount after or before another unit or mount point.
  1251. * Enqueued start jobs for device units are now automatically garbage
  1252. collected if there are no jobs waiting for them anymore.
  1253. * systemctl list-jobs gained two new switches: with --after, for every
  1254. queued job the jobs it's waiting for are shown; with --before the
  1255. jobs which it's blocking are shown.
  1256. * systemd-nspawn gained support for ephemeral boots from disk images
  1257. (or in other words: --ephemeral and --image= may now be
  1258. combined). Moreover, ephemeral boots are now supported for normal
  1259. directories, even if the backing file system is not btrfs. Of course,
  1260. if the file system does not support file system snapshots or
  1261. reflinks, the initial copy operation will be relatively expensive, but
  1262. this should still be suitable for many use cases.
  1263. * Calendar time specifications in .timer units now support
  1264. specifications relative to the end of a month by using "~" instead of
  1265. "-" as separator between month and day. For example, "*-02~03" means
  1266. "the third last day in February". In addition a new syntax for
  1267. repeated events has been added using the "/" character. For example,
  1268. "9..17/2:00" means "every two hours from 9am to 5pm".
  1269. * systemd-socket-proxyd gained a new parameter --connections-max= for
  1270. configuring the maximum number of concurrent connections.
  1271. * sd-id128 gained a new API for generating unique IDs for the host in a
  1272. way that does not leak the machine ID. Specifically,
  1273. sd_id128_get_machine_app_specific() derives an ID based on the
  1274. machine ID a in well-defined, non-reversible, stable way. This is
  1275. useful whenever an identifier for the host is needed but where the
  1276. identifier shall not be useful to identify the system beyond the
  1277. scope of the application itself. (Internally this uses HMAC-SHA256 as
  1278. keyed hash function using the machine ID as input.)
  1279. * NotifyAccess= gained a new supported value "exec". When set
  1280. notifications are accepted from all processes systemd itself invoked,
  1281. including all control processes.
  1282. * .nspawn files gained support for defining overlay mounts using the
  1283. Overlay= and OverlayReadOnly= options. Previously this functionality
  1284. was only available on the systemd-nspawn command line.
  1285. * systemd-nspawn's --bind= and --overlay= options gained support for
  1286. bind/overlay mounts whose source lies within the container tree by
  1287. prefixing the source path with "+".
  1288. * systemd-nspawn's --bind= and --overlay= options gained support for
  1289. automatically allocating a temporary source directory in /var/tmp
  1290. that is removed when the container dies. Specifically, if the source
  1291. directory is specified as empty string this mechanism is selected. An
  1292. example usage is --overlay=+/var::/var, which creates an overlay
  1293. mount based on the original /var contained in the image, overlayed
  1294. with a temporary directory in the host's /var/tmp. This way changes
  1295. to /var are automatically flushed when the container shuts down.
  1296. * systemd-nspawn --image= option does now permit raw file system block
  1297. devices (in addition to images containing partition tables, as
  1298. before).
  1299. * The disk image dissection logic in systemd-nspawn gained support for
  1300. automatically setting up LUKS encrypted as well as Verity protected
  1301. partitions. When a container is booted from an encrypted image the
  1302. passphrase is queried at start-up time. When a container with Verity
  1303. data is started, the root hash is search in a ".roothash" file
  1304. accompanying the disk image (alternatively, pass the root hash via
  1305. the new --root-hash= command line option).
  1306. * A new tool /usr/lib/systemd/systemd-dissect has been added that may
  1307. be used to dissect disk images the same way as systemd-nspawn does
  1308. it, following the Bootable Partition Specification. It may even be
  1309. used to mount disk images with complex partition setups (including
  1310. LUKS and Verity partitions) to a local host directory, in order to
  1311. inspect them. This tool is not considered public API (yet), and is
  1312. thus not installed into /usr/bin. Please do not rely on its
  1313. existence, since it might go away or be changed in later systemd
  1314. versions.
  1315. * A new generator "systemd-verity-generator" has been added, similar in
  1316. style to "systemd-cryptsetup-generator", permitting automatic setup of
  1317. Verity root partitions when systemd boots up. In order to make use of
  1318. this your partition setup should follow the Discoverable Partitions
  1319. Specification, and the GPT partition ID of the root file system
  1320. partition should be identical to the upper 128bit of the Verity root
  1321. hash. The GPT partition ID of the Verity partition protecting it
  1322. should be the lower 128bit of the Verity root hash. If the partition
  1323. image follows this model it is sufficient to specify a single
  1324. "roothash=" kernel command line argument to both configure which root
  1325. image and verity partition to use as well as the root hash for
  1326. it. Note that systemd-nspawn's Verity support follows the same
  1327. semantics, meaning that disk images with proper Verity data in place
  1328. may be booted in containers with systemd-nspawn as well as on
  1329. physical systems via the verity generator. Also note that the "mkosi"
  1330. tool available at has been updated
  1331. to generate Verity protected disk images following this scheme. In
  1332. fact, it has been updated to generate disk images that optionally
  1333. implement a complete UEFI SecureBoot trust chain, involving a signed
  1334. kernel and initrd image that incorporates such a root hash as well as
  1335. a Verity-enabled root partition.
  1336. * The hardware database (hwdb) udev supports has been updated to carry
  1337. accelerometer quirks.
  1338. * All system services are now run with a fresh kernel keyring set up
  1339. for them. The invocation ID is stored by default in it, thus
  1340. providing a safe, non-overridable way to determine the invocation
  1341. ID of each service.
  1342. * Service unit files gained new BindPaths= and BindReadOnlyPaths=
  1343. options for bind mounting arbitrary paths in a service-specific
  1344. way. When these options are used, arbitrary host or service files and
  1345. directories may be mounted to arbitrary locations in the service's
  1346. view.
  1347. * Documentation has been added that lists all of systemd's low-level
  1348. environment variables:
  1350. * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
  1351. whether a specific socket file descriptor matches a specified socket
  1352. address.
  1353. * systemd-firstboot has been updated to check for the
  1354. systemd.firstboot= kernel command line option. It accepts a boolean
  1355. and when set to false the first boot questions are skipped.
  1356. * systemd-fstab-generator has been updated to check for the
  1357. systemd.volatile= kernel command line option, which either takes an
  1358. optional boolean parameter or the special value "state". If used the
  1359. system may be booted in a "volatile" boot mode. Specifically,
  1360. "systemd.volatile" is used, the root directory will be mounted as
  1361. tmpfs, and only /usr is mounted from the actual root file system. If
  1362. "systemd.volatile=state" is used, the root directory will be mounted
  1363. as usual, but /var is mounted as tmpfs. This concept provides similar
  1364. functionality as systemd-nspawn's --volatile= option, but provides it
  1365. on physical boots. Use this option for implementing stateless
  1366. systems, or testing systems with all state and/or configuration reset
  1367. to the defaults. (Note though that many distributions are not
  1368. prepared to boot up without a populated /etc or /var, though.)
  1369. * systemd-gpt-auto-generator gained support for LUKS encrypted root
  1370. partitions. Previously it only supported LUKS encrypted partitions
  1371. for all other uses, except for the root partition itself.
  1372. * Socket units gained support for listening on AF_VSOCK sockets for
  1373. communication in virtualized QEMU environments.
  1374. * The "configure" script gained a new option --with-fallback-hostname=
  1375. for specifying the fallback hostname to use if none is configured in
  1376. /etc/hostname. For example, by specifying
  1377. --with-fallback-hostname=fedora it is possible to default to a
  1378. hostname of "fedora" on pristine installations.
  1379. * systemd-cgls gained support for a new --unit= switch for listing only
  1380. the control groups of a specific unit. Similar --user-unit= has been
  1381. added for listing only the control groups of a specific user unit.
  1382. * systemd-mount gained a new --umount switch for unmounting a mount or
  1383. automount point (and all mount/automount points below it).
  1384. * systemd will now refuse full configuration reloads (via systemctl
  1385. daemon-reload and related calls) unless at least 16MiB of free space
  1386. are available in /run. This is a safety precaution in order to ensure
  1387. that generators can safely operate after the reload completed.
  1388. * A new unit file option RootImage= has been added, which has a similar
  1389. effect as RootDirectory= but mounts the service's root directory from
  1390. a disk image instead of plain directory. This logic reuses the same
  1391. image dissection and mount logic that systemd-nspawn already uses,
  1392. and hence supports any disk images systemd-nspawn supports, including
  1393. those following the Discoverable Partition Specification, as well as
  1394. Verity enabled images. This option enables systemd to run system
  1395. services directly off disk images acting as resource bundles,
  1396. possibly even including full integrity data.
  1397. * A new MountAPIVFS= unit file option has been added, taking a boolean
  1398. argument. If enabled /proc, /sys and /dev (collectively called the
  1399. "API VFS") will be mounted for the service. This is only relevant if
  1400. RootDirectory= or RootImage= is used for the service, as these mounts
  1401. are of course in place in the host mount namespace anyway.
  1402. * systemd-nspawn gained support for a new --pivot-root= switch. If
  1403. specified the root directory within the container image is pivoted to
  1404. the specified mount point, while the original root disk is moved to a
  1405. different place. This option enables booting of ostree images
  1406. directly with systemd-nspawn.
  1407. * The systemd build scripts will no longer complain if the NTP server
  1408. addresses are not changed from the defaults. Google now supports
  1409. these NTP servers officially. We still recommend downstreams to
  1410. properly register an NTP pool with the NTP pool project though.
  1411. * coredumpctl gained a new "--reverse" option for printing the list
  1412. of coredumps in reverse order.
  1413. * coredumpctl will now show additional information about truncated and
  1414. inaccessible coredumps, as well as coredumps that are still being
  1415. processed. It also gained a new --quiet switch for suppressing
  1416. additional informational message in its output.
  1417. * coredumpctl gained support for only showing coredumps newer and/or
  1418. older than specific timestamps, using the new --since= and --until=
  1419. options, reminiscent of journalctl's options by the same name.
  1420. * The systemd-coredump logic has been improved so that it may be reused
  1421. to collect backtraces in non-compiled languages, for example in
  1422. scripting languages such as Python.
  1423. * machinectl will now show the UID shift of local containers, if user
  1424. namespacing is enabled for them.
  1425. * systemd will now optionally run "environment generator" binaries at
  1426. configuration load time. They may be used to add environment
  1427. variables to the environment block passed to services invoked. One
  1428. user environment generator is shipped by default that sets up
  1429. environment variables based on files dropped into /etc/environment.d
  1430. and ~/.config/environment.d/.
  1431. * systemd-resolved now includes the new, recently published 2017 DNSSEC
  1432. root key (KSK).
  1433. * hostnamed has been updated to report a new chassis type of
  1434. "convertible" to cover "foldable" laptops that can both act as a
  1435. tablet and as a laptop, such as various Lenovo Yoga devices.
  1436. Contributions from: Adrián López, Alexander Galanin, Alexander
  1437. Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
  1438. Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
  1439. Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
  1440. Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
  1441. David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
  1442. Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
  1443. Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
  1444. Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
  1445. Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
  1446. Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
  1447. Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
  1448. Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
  1449. Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
  1450. Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
  1451. Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
  1452. Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
  1453. Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
  1454. Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
  1455. Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
  1456. Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
  1457. Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
  1458. Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
  1459. Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
  1460. Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
  1461. Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
  1462. Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
  1463. YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
  1464. Тихонов
  1465. — Berlin, 2017-03-01
  1466. CHANGES WITH 232:
  1467. * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
  1468. RestrictAddressFamilies= enabled. These sandboxing options should
  1469. generally be compatible with the various external udev call-out
  1470. binaries we are aware of, however there may be exceptions, in
  1471. particular when exotic languages for these call-outs are used. In
  1472. this case, consider turning off these settings locally.
  1473. * The new RemoveIPC= option can be used to remove IPC objects owned by
  1474. the user or group of a service when that service exits.
  1475. * The new ProtectKernelModules= option can be used to disable explicit
  1476. load and unload operations of kernel modules by a service. In
  1477. addition access to /usr/lib/modules is removed if this option is set.
  1478. * ProtectSystem= option gained a new value "strict", which causes the
  1479. whole file system tree with the exception of /dev, /proc, and /sys,
  1480. to be remounted read-only for a service.
  1481. * The new ProtectKernelTunables= option can be used to disable
  1482. modification of configuration files in /sys and /proc by a service.
  1483. Various directories and files are remounted read-only, so access is
  1484. restricted even if the file permissions would allow it.
  1485. * The new ProtectControlGroups= option can be used to disable write
  1486. access by a service to /sys/fs/cgroup.
  1487. * Various systemd services have been hardened with
  1488. ProtectKernelTunables=yes, ProtectControlGroups=yes,
  1489. RestrictAddressFamilies=.
  1490. * Support for dynamically creating users for the lifetime of a service
  1491. has been added. If DynamicUser=yes is specified, user and group IDs
  1492. will be allocated from the range 61184..65519 for the lifetime of the
  1493. service. They can be resolved using the new NSS
  1494. module. The module must be enabled in /etc/nsswitch.conf. Services
  1495. started in this way have PrivateTmp= and RemoveIPC= enabled, so that
  1496. any resources allocated by the service will be cleaned up when the
  1497. service exits. They also have ProtectHome=read-only and
  1498. ProtectSystem=strict enabled, so they are not able to make any
  1499. permanent modifications to the system.
  1500. * The nss-systemd module also always resolves root and nobody, making
  1501. it possible to have no /etc/passwd or /etc/group files in minimal
  1502. container or chroot environments.
  1503. * Services may be started with their own user namespace using the new
  1504. boolean PrivateUsers= option. Only root, nobody, and the uid/gid
  1505. under which the service is running are mapped. All other users are
  1506. mapped to nobody.
  1507. * Support for the cgroup namespace has been added to systemd-nspawn. If
  1508. supported by kernel, the container system started by systemd-nspawn
  1509. will have its own view of the cgroup hierarchy. This new behaviour
  1510. can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
  1511. * The new MemorySwapMax= option can be used to limit the maximum swap
  1512. usage under the unified cgroup hierarchy.
  1513. * Support for the CPU controller in the unified cgroup hierarchy has
  1514. been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
  1515. options. This controller requires out-of-tree patches for the kernel
  1516. and the support is provisional.
  1517. * Mount and automount units may now be created transiently
  1518. (i.e. dynamically at runtime via the bus API, instead of requiring
  1519. unit files in the file system).
  1520. * systemd-mount is a new tool which may mount file systems – much like
  1521. mount(8), optionally pulling in additional dependencies through
  1522. transient .mount and .automount units. For example, this tool
  1523. automatically runs fsck on a backing block device before mounting,
  1524. and allows the automount logic to be used dynamically from the
  1525. command line for establishing mount points. This tool is particularly
  1526. useful when dealing with removable media, as it will ensure fsck is
  1527. run – if necessary – before the first access and that the file system
  1528. is quickly unmounted after each access by utilizing the automount
  1529. logic. This maximizes the chance that the file system on the
  1530. removable media stays in a clean state, and if it isn't in a clean
  1531. state is fixed automatically.
  1532. * LazyUnmount=yes option for mount units has been added to expose the
  1533. umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
  1534. option.
  1535. * /efi will be used as the mount point of the EFI boot partition, if
  1536. the directory is present, and the mount point was not configured
  1537. through other means (e.g. fstab). If /efi directory does not exist,
  1538. /boot will be used as before. This makes it easier to automatically
  1539. mount the EFI partition on systems where /boot is used for something
  1540. else.
  1541. * When operating on GPT disk images for containers, systemd-nspawn will
  1542. now mount the ESP to /boot or /efi according to the same rules as PID
  1543. 1 running on a host. This allows tools like "bootctl" to operate
  1544. correctly within such containers, in order to make container images
  1545. bootable on physical systems.
  1546. * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
  1547. * Two new user session targets have been added to support running
  1548. graphical sessions under the systemd --user instance:
  1549. and See
  1550. systemd.special(7) for a description of how those targets should be
  1551. used.
  1552. * The vconsole initialization code has been significantly reworked to
  1553. use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
  1554. support unicode keymaps. Font and keymap configuration will now be
  1555. copied to all allocated virtual consoles.
  1556. * FreeBSD's bhyve virtualization is now detected.
  1557. * Information recorded in the journal for core dumps now includes the
  1558. contents of /proc/mountinfo and the command line of the process at
  1559. the top of the process hierarchy (which is usually the init process
  1560. of the container).
  1561. * systemd-journal-gatewayd learned the --directory= option to serve
  1562. files from the specified location.
  1563. * journalctl --root=… can be used to peruse the journal in the
  1564. /var/log/ directories inside of a container tree. This is similar to
  1565. the existing --machine= option, but does not require the container to
  1566. be active.
  1567. * The hardware database has been extended to support
  1568. ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
  1569. trackball devices.
  1570. MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
  1571. specify the click rate for mice which include a horizontal wheel with
  1572. a click rate that is different than the one for the vertical wheel.
  1573. * systemd-run gained a new --wait option that makes service execution
  1574. synchronous. (Specifically, the command will not return until the
  1575. specified service binary exited.)
  1576. * systemctl gained a new --wait option that causes the start command to
  1577. wait until the units being started have terminated again.
  1578. * A new journal output mode "short-full" has been added which displays
  1579. timestamps with abbreviated English day names and adds a timezone
  1580. suffix. Those timestamps include more information than the default
  1581. "short" output mode, and can be passed directly to journalctl's
  1582. --since= and --until= options.
  1583. * /etc/resolv.conf will be bind-mounted into containers started by
  1584. systemd-nspawn, if possible, so any changes to resolv.conf contents
  1585. are automatically propagated to the container.
  1586. * The number of instances for socket-activated services originating
  1587. from a single IP address can be limited with
  1588. MaxConnectionsPerSource=, extending the existing setting of
  1589. MaxConnections=.
  1590. * systemd-networkd gained support for vcan ("Virtual CAN") interface
  1591. configuration.
  1592. * .netdev and .network configuration can now be extended through
  1593. drop-ins.
  1594. * UDP Segmentation Offload, TCP Segmentation Offload, Generic
  1595. Segmentation Offload, Generic Receive Offload, Large Receive Offload
  1596. can be enabled and disabled using the new UDPSegmentationOffload=,
  1597. TCPSegmentationOffload=, GenericSegmentationOffload=,
  1598. GenericReceiveOffload=, LargeReceiveOffload= options in the
  1599. [Link] section of .link files.
  1600. * The Spanning Tree Protocol, Priority, Aging Time, and the Default
  1601. Port VLAN ID can be configured for bridge devices using the new STP=,
  1602. Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
  1603. section of .netdev files.
  1604. * The route table to which routes received over DHCP or RA should be
  1605. added can be configured with the new RouteTable= option in the [DHCP]
  1606. and [IPv6AcceptRA] sections of .network files.
  1607. * The Address Resolution Protocol can be disabled on links managed by
  1608. systemd-networkd using the ARP=no setting in the [Link] section of
  1609. .network files.
  1610. * New environment variables $SERVICE_RESULT, $EXIT_CODE and
  1611. $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
  1612. encode information about the result and exit codes of the current
  1613. service runtime cycle.
  1614. * systemd-sysctl will now configure kernel parameters in the order
  1615. they occur in the configuration files. This matches what sysctl
  1616. has been traditionally doing.
  1617. * kernel-install "plugins" that are executed to perform various
  1618. tasks after a new kernel is added and before an old one is removed
  1619. can now return a special value to terminate the procedure and
  1620. prevent any later plugins from running.
  1621. * Journald's SplitMode=login setting has been deprecated. It has been
  1622. removed from documentation, and its use is discouraged. In a future
  1623. release it will be completely removed, and made equivalent to current
  1624. default of SplitMode=uid.
  1625. * Storage=both option setting in /etc/systemd/coredump.conf has been
  1626. removed. With fast LZ4 compression storing the core dump twice is not
  1627. useful.
  1628. * The --share-system systemd-nspawn option has been replaced with an
  1629. (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
  1630. this functionality is discouraged. In addition the variables
  1632. $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
  1633. individual namespaces.
  1634. * "machinectl list" now shows the IP address of running containers in
  1635. the output, as well as OS release information.
  1636. * "loginctl list" now shows the TTY of each session in the output.
  1637. * sd-bus gained new API calls sd_bus_track_set_recursive(),
  1638. sd_bus_track_get_recursive(), sd_bus_track_count_name(),
  1639. sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
  1640. tracking objects in a "recursive" mode, where a single client can be
  1641. counted multiple times, if it takes multiple references.
  1642. * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
  1643. sd_bus_get_exit_on_disconnect(). They may be used to make a
  1644. process using sd-bus automatically exit if the bus connection is
  1645. severed.
  1646. * Bus clients of the service manager may now "pin" loaded units into
  1647. memory, by taking an explicit reference on them. This is useful to
  1648. ensure the client can retrieve runtime data about the service even
  1649. after the service completed execution. Taking such a reference is
  1650. available only for privileged clients and should be helpful to watch
  1651. running services in a race-free manner, and in particular collect
  1652. information about exit statuses and results.
  1653. * The nss-resolve module has been changed to strictly return UNAVAIL
  1654. when communication via D-Bus with resolved failed, and NOTFOUND when
  1655. a lookup completed but was negative. This means it is now possible to
  1656. neatly configure fallbacks using nsswitch.conf result checking
  1657. expressions. Taking benefit of this, the new recommended
  1658. configuration line for the "hosts" entry in /etc/nsswitch.conf is:
  1659. hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
  1660. * A new setting CtrlAltDelBurstAction= has been added to
  1661. /etc/systemd/system.conf which may be used to configure the precise
  1662. behaviour if the user on the console presses Ctrl-Alt-Del more often
  1663. than 7 times in 2s. Previously this would unconditionally result in
  1664. an expedited, immediate reboot. With this new setting the precise
  1665. operation may be configured in more detail, and also turned off
  1666. entirely.
  1667. * In .netdev files two new settings RemoteChecksumTx= and
  1668. RemoteChecksumRx= are now understood that permit configuring the
  1669. remote checksumming logic for VXLAN networks.
  1670. * The service manager learnt a new "invocation ID" concept for invoked
  1671. services. Each runtime cycle of a service will get a new invocation
  1672. ID (a 128bit random UUID) assigned that identifies the current
  1673. run of the service uniquely and globally. A new invocation ID
  1674. is generated each time a service starts up. The journal will store
  1675. the invocation ID of a service along with any logged messages, thus
  1676. making the invocation ID useful for matching the online runtime of a
  1677. service with the offline log data it generated in a safe way without
  1678. relying on synchronized timestamps. In many ways this new service
  1679. invocation ID concept is similar to the kernel's boot ID concept that
  1680. uniquely and globally identifies the runtime of each boot. The
  1681. invocation ID of a service is passed to the service itself via an
  1682. environment variable ($INVOCATION_ID). A new bus call
  1683. GetUnitByInvocationID() has been added that is similar to GetUnit()
  1684. but instead of retrieving the bus path for a unit by its name
  1685. retrieves it by its invocation ID. The returned path is valid only as
  1686. long as the passed invocation ID is current.
  1687. * systemd-resolved gained a new "DNSStubListener" setting in
  1688. resolved.conf. It either takes a boolean value or the special values
  1689. "udp" and "tcp", and configures whether to enable the stub DNS
  1690. listener on
  1691. * IP addresses configured via networkd may now carry additional
  1692. configuration settings supported by the kernel. New options include:
  1693. HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
  1694. PrefixRoute=, AutoJoin=.
  1695. * The PAM configuration fragment file for "user@.service" shipped with
  1696. systemd (i.e. the --user instance of systemd) has been stripped to
  1697. the minimum necessary to make the system boot. Previously, it
  1698. contained Fedora-specific stanzas that did not apply to other
  1699. distributions. It is expected that downstream distributions add
  1700. additional configuration lines, matching their needs to this file,
  1701. using it only as rough template of what systemd itself needs. Note
  1702. that this reduced fragment does not even include an invocation of
  1703. pam_limits which most distributions probably want to add, even though
  1704. systemd itself does not need it. (There's also the new build time
  1705. option --with-pamconfdir=no to disable installation of the PAM
  1706. fragment entirely.)
  1707. * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
  1708. capability is now also dropped from its set (in addition to
  1709. CAP_SYS_MKNOD as before).
  1710. * In service unit files it is now possible to connect a specific named
  1711. file descriptor with stdin/stdout/stdout of an executed service. The
  1712. name may be specified in matching .socket units using the
  1713. FileDescriptorName= setting.
  1714. * A number of journal settings may now be configured on the kernel
  1715. command line. Specifically, the following options are now understood:
  1716. systemd.journald.max_level_console=,
  1717. systemd.journald.max_level_store=,
  1718. systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
  1719. systemd.journald.max_level_wall=.
  1720. * "systemctl is-enabled --full" will now show by which symlinks a unit
  1721. file is enabled in the unit dependency tree.
  1722. * Support for VeraCrypt encrypted partitions has been added to the
  1723. "cryptsetup" logic and /etc/crypttab.
  1724. * systemd-detect-virt gained support for a new --private-users switch
  1725. that checks whether the invoking processes are running inside a user
  1726. namespace. Similar, a new special value "private-users" for the
  1727. existing ConditionVirtualization= setting has been added, permitting
  1728. skipping of specific units in user namespace environments.
  1729. Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
  1730. Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
  1731. Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
  1732. Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
  1733. Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
  1734. Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
  1735. Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
  1736. Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
  1737. Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
  1738. Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
  1739. Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
  1740. Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
  1741. Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
  1742. Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
  1743. Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
  1744. Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
  1745. Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
  1746. Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
  1747. Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
  1748. Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
  1749. Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
  1750. Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
  1751. Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
  1752. Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
  1753. E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
  1754. Jędrzejewski-Szmek, Zeal Jagannatha
  1755. — Santa Fe, 2016-11-03
  1756. CHANGES WITH 231:
  1757. * In service units the various ExecXYZ= settings have been extended
  1758. with an additional special character as first argument of the
  1759. assigned value: if the character '+' is used the specified command
  1760. line it will be run with full privileges, regardless of User=,
  1761. Group=, CapabilityBoundingSet= and similar options. The effect is
  1762. similar to the existing PermissionsStartOnly= option, but allows
  1763. configuration of this concept for each executed command line
  1764. independently.
  1765. * Services may now alter the service watchdog timeout at runtime by
  1766. sending a WATCHDOG_USEC= message via sd_notify().
  1767. * MemoryLimit= and related unit settings now optionally take percentage
  1768. specifications. The percentage is taken relative to the amount of
  1769. physical memory in the system (or in case of containers, the assigned
  1770. amount of memory). This allows scaling service resources neatly with
  1771. the amount of RAM available on the system. Similarly, systemd-logind's
  1772. RuntimeDirectorySize= option now also optionally takes percentage
  1773. values.
  1774. * In similar fashion TasksMax= takes percentage values now, too. The
  1775. value is taken relative to the configured maximum number of processes
  1776. on the system. The per-service task maximum has been changed to 15%
  1777. using this functionality. (Effectively this is an increase of 512 →
  1778. 4915 for service units, given the kernel's default pid_max setting.)
  1779. * Calendar time specifications in .timer units now understand a ".."
  1780. syntax for time ranges. Example: "4..7:10" may now be used for
  1781. defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
  1782. 7:10am every day.
  1783. * The InaccessableDirectories=, ReadOnlyDirectories= and
  1784. ReadWriteDirectories= unit file settings have been renamed to
  1785. InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
  1786. applied to all kinds of file nodes, and not just directories, with
  1787. the exception of symlinks. Specifically these settings may now be
  1788. used on block and character device nodes, UNIX sockets and FIFOS as
  1789. well as regular files. The old names of these settings remain
  1790. available for compatibility.
  1791. * systemd will now log about all service processes it kills forcibly
  1792. (using SIGKILL) because they remained after the clean shutdown phase
  1793. of the service completed. This should help identifying services that
  1794. shut down uncleanly. Moreover if KillUserProcesses= is enabled in
  1795. systemd-logind's configuration a similar log message is generated for
  1796. processes killed at the end of each session due to this setting.
  1797. * systemd will now set the $JOURNAL_STREAM environment variable for all
  1798. services whose stdout/stderr are connected to the Journal (which
  1799. effectively means by default: all services). The variable contains
  1800. the device and inode number of the file descriptor used for
  1801. stdout/stderr. This may be used by invoked programs to detect whether
  1802. their stdout/stderr is connected to the Journal, in which case they
  1803. can switch over to direct Journal communication, thus being able to
  1804. pass extended, structured metadata along with their log messages. As
  1805. one example, this is now used by glib's logging primitives.
  1806. * When using systemd's default tmp.mount unit for /tmp, the mount point
  1807. will now be established with the "nosuid" and "nodev" options. This
  1808. avoids privilege escalation attacks that put traps and exploits into
  1809. /tmp. However, this might cause problems if you e. g. put container
  1810. images or overlays into /tmp; if you need this, override tmp.mount's
  1811. "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
  1812. desired options.
  1813. * systemd now supports the "memory" cgroup controller also on
  1814. cgroupsv2.
  1815. * The systemd-cgtop tool now optionally takes a control group path as
  1816. command line argument. If specified, the control group list shown is
  1817. limited to subgroups of that group.
  1818. * The SystemCallFilter= unit file setting gained support for
  1819. pre-defined, named system call filter sets. For example
  1820. SystemCallFilter=@clock is now an effective way to make all clock
  1821. changing-related system calls unavailable to a service. A number of
  1822. similar pre-defined groups are defined. Writing system call filters
  1823. for system services is simplified substantially with this new
  1824. concept. Accordingly, all of systemd's own, long-running services now
  1825. enable system call filtering based on this, by default.
  1826. * A new service setting MemoryDenyWriteExecute= has been added, taking
  1827. a boolean value. If turned on, a service may no longer create memory
  1828. mappings that are writable and executable at the same time. This
  1829. enhances security for services where this is enabled as it becomes
  1830. harder to dynamically write and then execute memory in exploited
  1831. service processes. This option has been enabled for all of systemd's
  1832. own long-running services.
  1833. * A new RestrictRealtime= service setting has been added, taking a
  1834. boolean argument. If set the service's processes may no longer
  1835. acquire realtime scheduling. This improves security as realtime
  1836. scheduling may otherwise be used to easily freeze the system.
  1837. * systemd-nspawn gained a new switch --notify-ready= taking a boolean
  1838. value. This may be used for requesting that the system manager inside
  1839. of the container reports start-up completion to nspawn which then
  1840. propagates this notification further to the service manager
  1841. supervising nspawn itself. A related option NotifyReady= in .nspawn
  1842. files has been added too. This functionality allows ordering of the
  1843. start-up of multiple containers using the usual systemd ordering
  1844. primitives.
  1845. * machinectl gained a new command "stop" that is an alias for
  1846. "terminate".
  1847. * systemd-resolved gained support for contacting DNS servers on
  1848. link-local IPv6 addresses.
  1849. * If systemd-resolved receives the SIGUSR2 signal it will now flush all
  1850. its caches. A method call for requesting the same operation has been
  1851. added to the bus API too, and is made available via "systemd-resolve
  1852. --flush-caches".
  1853. * systemd-resolve gained a new --status switch. If passed a brief
  1854. summary of the used DNS configuration with per-interface information
  1855. is shown.
  1856. * resolved.conf gained a new Cache= boolean option, defaulting to
  1857. on. If turned off local DNS caching is disabled. This comes with a
  1858. performance penalty in particular when DNSSEC is enabled. Note that
  1859. resolved disables its internal caching implicitly anyway, when the
  1860. configured DNS server is on a host-local IP address such as ::1 or
  1861., thus automatically avoiding double local caching.
  1862. * systemd-resolved now listens on the local IP address
  1863. for DNS requests. This improves compatibility with local programs
  1864. that do not use the libc NSS or systemd-resolved's bus APIs for name
  1865. resolution. This minimal DNS service is only available to local
  1866. programs and does not implement the full DNS protocol, but enough to
  1867. cover local DNS clients. A new, static resolv.conf file, listing just
  1868. this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
  1869. now recommended to make /etc/resolv.conf a symlink to this file in
  1870. order to route all DNS lookups to systemd-resolved, regardless if
  1871. done via NSS, the bus API or raw DNS packets. Note that this local
  1872. DNS service is not as fully featured as the libc NSS or
  1873. systemd-resolved's bus APIs. For example, as unicast DNS cannot be
  1874. used to deliver link-local address information (as this implies
  1875. sending a local interface index along), LLMNR/mDNS support via this
  1876. interface is severely restricted. It is thus strongly recommended for
  1877. all applications to use the libc NSS API or native systemd-resolved
  1878. bus API instead.
  1879. * systemd-networkd's bridge support learned a new setting
  1880. VLANFiltering= for controlling VLAN filtering. Moreover a new section
  1881. in .network files has been added for configuring VLAN bridging in
  1882. more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
  1883. * systemd-networkd's IPv6 Router Advertisement code now makes use of
  1884. the DNSSL and RDNSS options. This means IPv6 DNS configuration may
  1885. now be acquired without relying on DHCPv6. Two new options
  1886. UseDomains= and UseDNS= have been added to configure this behaviour.
  1887. * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
  1888. renamed IPv6AcceptRA=, without altering its behaviour. The old
  1889. setting name remains available for compatibility reasons.
  1890. * The systemd-networkd VTI/VTI6 tunneling support gained new options
  1891. Key=, InputKey= and OutputKey=.
  1892. * systemd-networkd gained support for VRF ("Virtual Routing Function")
  1893. interface configuration.
  1894. * "systemctl edit" may now be used to create new unit files by
  1895. specifying the --force switch.
  1896. * sd-event gained a new function sd_event_get_iteration() for
  1897. requesting the current iteration counter of the event loop. It starts
  1898. at zero and is increased by one with each event loop iteration.
  1899. * A new rpm macro %systemd_ordering is provided by the macros.systemd
  1900. file. It can be used in lieu of %systemd_requires in packages which
  1901. don't use any systemd functionality and are intended to be installed
  1902. in minimal containers without systemd present. This macro provides
  1903. ordering dependencies to ensure that if the package is installed in
  1904. the same rpm transaction as systemd, systemd will be installed before
  1905. the scriptlets for the package are executed, allowing unit presets
  1906. to be handled.
  1907. New macros %_systemdgeneratordir and %_systemdusergeneratordir have
  1908. been added to simplify packaging of generators.
  1909. * The os-release file gained VERSION_CODENAME field for the
  1910. distribution nickname (e.g. VERSION_CODENAME=woody).
  1912. can be set to disable parsing of metadata and the creation
  1913. of persistent symlinks for that device.
  1914. * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
  1915. to make them available to logged-in users has been reverted.
  1916. * Much of the common code of the various systemd components is now
  1917. built into an internal shared library
  1918. (incorporating the systemd version number in the name, to be updated
  1919. with future releases) that the components link to. This should
  1920. decrease systemd footprint both in memory during runtime and on
  1921. disk. Note that the shared library is not for public use, and is
  1922. neither API nor ABI stable, but is likely to change with every new
  1923. released update. Packagers need to make sure that binaries
  1924. linking to are updated in step with the
  1925. library.
  1926. * Configuration for "mkosi" is now part of the systemd
  1927. repository. mkosi is a tool to easily build legacy-free OS images,
  1928. and is available on github: If
  1929. "mkosi" is invoked in the build tree a new raw OS image is generated
  1930. incorporating the systemd sources currently being worked on and a
  1931. clean, fresh distribution installation. The generated OS image may be
  1932. booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
  1933. UEFI PC. This functionality is particularly useful to easily test
  1934. local changes made to systemd in a pristine, defined environment. See
  1935. doc/HACKING for details.
  1936. * configure learned the --with-support-url= option to specify the
  1937. distribution's bugtracker.
  1938. Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
  1939. Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
  1940. Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
  1941. Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
  1942. Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
  1943. Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
  1944. Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
  1945. Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
  1946. Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
  1947. Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
  1948. Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
  1949. Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
  1950. Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
  1951. Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
  1952. Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
  1953. Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
  1954. Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
  1955. Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
  1956. WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
  1957. — Berlin, 2016-07-25
  1958. CHANGES WITH 230:
  1959. * DNSSEC is now turned on by default in systemd-resolved (in
  1960. "allow-downgrade" mode), but may be turned off during compile time by
  1961. passing "--with-default-dnssec=no" to "configure" (and of course,
  1962. during runtime with DNSSEC= in resolved.conf). We recommend
  1963. downstreams to leave this on at least during development cycles and
  1964. report any issues with the DNSSEC logic upstream. We are very
  1965. interested in collecting feedback about the DNSSEC validator and its
  1966. limitations in the wild. Note however, that DNSSEC support is
  1967. probably nothing downstreams should turn on in stable distros just
  1968. yet, as it might create incompatibilities with a few DNS servers and
  1969. networks. We tried hard to make sure we downgrade to non-DNSSEC mode
  1970. automatically whenever we detect such incompatible setups, but there
  1971. might be systems we do not cover yet. Hence: please help us testing
  1972. the DNSSEC code, leave this on where you can, report back, but then
  1973. again don't consider turning this on in your stable, LTS or
  1974. production release just yet. (Note that you have to enable
  1975. nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
  1976. and its DNSSEC mode for host name resolution from local
  1977. applications.)
  1978. * systemd-resolve conveniently resolves DANE records with the --tlsa
  1979. option and OPENPGPKEY records with the --openpgp option. It also
  1980. supports dumping raw DNS record data via the new --raw= switch.
  1981. * systemd-logind will now by default terminate user processes that are
  1982. part of the user session scope unit (session-XX.scope) when the user
  1983. logs out. This behavior is controlled by the KillUserProcesses=
  1984. setting in logind.conf, and the previous default of "no" is now
  1985. changed to "yes". This means that user sessions will be properly
  1986. cleaned up after, but additional steps are necessary to allow
  1987. intentionally long-running processes to survive logout.
  1988. While the user is logged in at least once, user@.service is running,
  1989. and any service that should survive the end of any individual login
  1990. session can be started at a user service or scope using systemd-run.
  1991. systemd-run(1) man page has been extended with an example which shows
  1992. how to run screen in a scope unit underneath user@.service. The same
  1993. command works for tmux.
  1994. After the user logs out of all sessions, user@.service will be
  1995. terminated too, by default, unless the user has "lingering" enabled.
  1996. To effectively allow users to run long-term tasks even if they are
  1997. logged out, lingering must be enabled for them. See loginctl(1) for
  1998. details. The default polkit policy was modified to allow users to
  1999. set lingering for themselves without authentication.
  2000. Previous defaults can be restored at compile time by the
  2001. --without-kill-user-processes option to "configure".
  2002. * systemd-logind gained new configuration settings SessionsMax= and
  2003. InhibitorsMax=, both with a default of 8192. It will not register new
  2004. user sessions or inhibitors above this limit.
  2005. * systemd-logind will now reload configuration on SIGHUP.
  2006. * The unified cgroup hierarchy added in Linux 4.5 is now supported.
  2007. Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
  2008. enable. Also, support for the "io" cgroup controller in the unified
  2009. hierarchy has been added, so that the "memory", "pids" and "io" are
  2010. now the controllers that are supported on the unified hierarchy.
  2011. WARNING: it is not possible to use previous systemd versions with
  2012. systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
  2013. is necessary to also update systemd in the initramfs if using the
  2014. unified hierarchy. An updated SELinux policy is also required.
  2015. * LLDP support has been extended, and both passive (receive-only) and
  2016. active (sender) modes are supported. Passive mode ("routers-only") is
  2017. enabled by default in systemd-networkd. Active LLDP mode is enabled
  2018. by default for containers on the internal network. The "networkctl
  2019. lldp" command may be used to list information gathered. "networkctl
  2020. status" will also show basic LLDP information on connected peers now.
  2021. * The IAID and DUID unique identifier sent in DHCP requests may now be
  2022. configured for the system and each .network file managed by
  2023. systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
  2024. * systemd-networkd gained support for configuring proxy ARP support for
  2025. each interface, via the ProxyArp= setting in .network files. It also
  2026. gained support for configuring the multicast querier feature of
  2027. bridge devices, via the new MulticastQuerier= setting in .netdev
  2028. files. Similarly, snooping on the IGMP traffic can be controlled
  2029. via the new setting MulticastSnooping=.
  2030. A new setting PreferredLifetime= has been added for addresses
  2031. configured in .network file to configure the lifetime intended for an
  2032. address.
  2033. The systemd-networkd DHCP server gained the option EmitRouter=, which
  2034. defaults to yes, to configure whether the DHCP Option 3 (Router)
  2035. should be emitted.
  2036. * The testing tool /usr/lib/systemd/systemd-activate is renamed to
  2037. systemd-socket-activate and installed into /usr/bin. It is now fully
  2038. supported.
  2039. * systemd-journald now uses separate threads to flush changes to disk
  2040. when closing journal files, thus reducing impact of slow disk I/O on
  2041. logging performance.
  2042. * The sd-journal API gained two new calls
  2043. sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
  2044. can be used to open journal files using file descriptors instead of
  2045. file or directory paths. sd_journal_open_container() has been
  2046. deprecated, sd_journal_open_directory_fd() should be used instead
  2047. with the flag SD_JOURNAL_OS_ROOT.
  2048. * journalctl learned a new output mode "-o short-unix" that outputs log
  2049. lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
  2050. UTC). It also gained support for a new --no-hostname setting to
  2051. suppress the hostname column in the family of "short" output modes.
  2052. * systemd-ask-password now optionally skips printing of the password to
  2053. stdout with --no-output which can be useful in scripts.
  2054. * Framebuffer devices (/dev/fb*) and 3D printers and scanners
  2055. (devices tagged with ID_MAKER_TOOL) are now tagged with
  2056. "uaccess" and are available to logged in users.
  2057. * The DeviceAllow= unit setting now supports specifiers (with "%").
  2058. * "systemctl show" gained a new --value switch, which allows print a
  2059. only the contents of a specific unit property, without also printing
  2060. the property's name. Similar support was added to "show*" verbs
  2061. of loginctl and machinectl that output "key=value" lists.
  2062. * A new unit type "generated" was added for files dynamically generated
  2063. by generator tools. Similarly, a new unit type "transient" is used
  2064. for unit files created using the runtime API. "systemctl enable" will
  2065. refuse to operate on such files.
  2066. * A new command "systemctl revert" has been added that may be used to
  2067. revert to the vendor version of a unit file, in case local changes
  2068. have been made by adding drop-ins or overriding the unit file.
  2069. * "machinectl clean" gained a new verb to automatically remove all or
  2070. just hidden container images.
  2071. * systemd-tmpfiles gained support for a new line type "e" for emptying
  2072. directories, if they exist, without creating them if they don't.
  2073. * systemd-nspawn gained support for automatically patching the UID/GIDs
  2074. of the owners and the ACLs of all files and directories in a
  2075. container tree to match the UID/GID user namespacing range selected
  2076. for the container invocation. This mode is enabled via the new
  2077. --private-users-chown switch. It also gained support for
  2078. automatically choosing a free, previously unused UID/GID range when
  2079. starting a container, via the new --private-users=pick setting (which
  2080. implies --private-users-chown). Together, these options for the first
  2081. time make user namespacing for nspawn containers fully automatic and
  2082. thus deployable. The systemd-nspawn@.service template unit file has
  2083. been changed to use this functionality by default.
  2084. * systemd-nspawn gained a new --network-zone= switch, that allows
  2085. creating ad-hoc virtual Ethernet links between multiple containers,
  2086. that only exist as long as at least one container referencing them is
  2087. running. This allows easy connecting of multiple containers with a
  2088. common link that implements an Ethernet broadcast domain. Each of
  2089. these network "zones" may be named relatively freely by the user, and
  2090. may be referenced by any number of containers, but each container may
  2091. only reference one of these "zones". On the lower level, this is
  2092. implemented by an automatically managed bridge network interface for
  2093. each zone, that is created when the first container referencing its
  2094. zone is created and removed when the last one referencing its zone
  2095. terminates.
  2096. * The default start timeout may now be configured on the kernel command
  2097. line via systemd.default_timeout_start_sec=. It was already
  2098. configurable via the DefaultTimeoutStartSec= option in
  2099. /etc/systemd/system.conf.
  2100. * Socket units gained a new TriggerLimitIntervalSec= and
  2101. TriggerLimitBurst= setting to configure a limit on the activation
  2102. rate of the socket unit.
  2103. * The LimitNICE= setting now optionally takes normal UNIX nice values
  2104. in addition to the raw integer limit value. If the specified
  2105. parameter is prefixed with "+" or "-" and is in the range -20..19 the
  2106. value is understood as UNIX nice value. If not prefixed like this it
  2107. is understood as raw RLIMIT_NICE limit.
  2108. * Note that the effect of the PrivateDevices= unit file setting changed
  2109. slightly with this release: the per-device /dev file system will be
  2110. mounted read-only from this version on, and will have "noexec"
  2111. set. This (minor) change of behavior might cause some (exceptional)
  2112. legacy software to break, when PrivateDevices=yes is set for its
  2113. service. Please leave PrivateDevices= off if you run into problems
  2114. with this.
  2115. * systemd-bootchart has been split out to a separate repository:
  2117. * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
  2118. merged into the kernel in its current form.
  2119. * The compatibility libraries,
  2120.,, and
  2121. which have been deprecated since systemd-209 have been removed along
  2122. with the corresponding pkg-config files. All symbols provided by
  2123. those libraries are provided by
  2124. * The Capabilities= unit file setting has been removed (it is ignored
  2125. for backwards compatibility). AmbientCapabilities= and
  2126. CapabilityBoundingSet= should be used instead.
  2127. * A new special target has been added,,
  2128. which creates a synchronization point for dependencies of the root
  2129. device in early userspace. Initramfs builders must ensure that this
  2130. target is now included in early userspace.
  2131. Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
  2132. Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
  2133. Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
  2134. Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
  2135. Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
  2136. R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
  2137. Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
  2138. Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
  2139. Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
  2140. Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
  2141. John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
  2142. Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
  2143. Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
  2144. Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
  2145. Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
  2146. mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
  2147. Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
  2148. Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
  2149. Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
  2150. Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
  2151. Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
  2152. Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
  2153. Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
  2154. Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
  2155. Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
  2156. Jędrzejewski-Szmek
  2157. — Fairfax, 2016-05-21
  2158. CHANGES WITH 229:
  2159. * The systemd-resolved DNS resolver service has gained a substantial
  2160. set of new features, most prominently it may now act as a DNSSEC
  2161. validating stub resolver. DNSSEC mode is currently turned off by
  2162. default, but is expected to be turned on by default in one of the
  2163. next releases. For now, we invite everybody to test the DNSSEC logic
  2164. by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
  2165. service also gained a full set of D-Bus interfaces, including calls
  2166. to configure DNS and DNSSEC settings per link (for use by external
  2167. network management software). systemd-resolved and systemd-networkd
  2168. now distinguish between "search" and "routing" domains. The former
  2169. are used to qualify single-label names, the latter are used purely
  2170. for routing lookups within certain domains to specific links.
  2171. resolved now also synthesizes RRs for all entries from /etc/hosts.
  2172. * The systemd-resolve tool (which is a client utility for
  2173. systemd-resolved) has been improved considerably and is now fully
  2174. supported and documented. Hence it has moved from /usr/lib/systemd to
  2175. /usr/bin.
  2176. * /dev/disk/by-path/ symlink support has been (re-)added for virtio
  2177. devices.
  2178. * The coredump collection logic has been reworked: when a coredump is
  2179. collected it is now written to disk, compressed and processed
  2180. (including stacktrace extraction) from a new instantiated service
  2181. systemd-coredump@.service, instead of directly from the
  2182. /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
  2183. processing large coredumps can take up a substantial amount of
  2184. resources and time, and this previously happened entirely outside of
  2185. systemd's service supervision. With the new logic the core_pattern
  2186. hook only does minimal metadata collection before passing off control
  2187. to the new instantiated service, which is configured with a time
  2188. limit, a nice level and other settings to minimize negative impact on
  2189. the rest of the system. Also note that the new logic will honour the
  2190. RLIMIT_CORE setting of the crashed process, which now allows users
  2191. and processes to turn off coredumping for their processes by setting
  2192. this limit.
  2193. * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
  2194. and all forked processes by default. Previously, PID 1 would leave
  2195. the setting at "0" for all processes, as set by the kernel. Note that
  2196. the resource limit traditionally has no effect on the generated
  2197. coredumps on the system if the /proc/sys/kernel/core_pattern hook
  2198. logic is used. Since the limit is now honoured (see above) its
  2199. default has been changed so that the coredumping logic is enabled by
  2200. default for all processes, while allowing specific opt-out.
  2201. * When the stacktrace is extracted from processes of system users, this
  2202. is now done as "systemd-coredump" user, in order to sandbox this
  2203. potentially security sensitive parsing operation. (Note that when
  2204. processing coredumps of normal users this is done under the user ID
  2205. of process that crashed, as before.) Packagers should take notice
  2206. that it is now necessary to create the "systemd-coredump" system user
  2207. and group at package installation time.
  2208. * The systemd-activate socket activation testing tool gained support
  2209. for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
  2210. and --seqpacket switches. It also has been extended to support both
  2211. new-style and inetd-style file descriptor passing. Use the new
  2212. --inetd switch to request inetd-style file descriptor passing.
  2213. * Most systemd tools now honor a new $SYSTEMD_COLORS environment
  2214. variable, which takes a boolean value. If set to false, ANSI color
  2215. output is disabled in the tools even when run on a terminal that
  2216. supports it.
  2217. * The VXLAN support in networkd now supports two new settings
  2218. DestinationPort= and PortRange=.
  2219. * A new systemd.machine_id= kernel command line switch has been added,
  2220. that may be used to set the machine ID in /etc/machine-id if it is
  2221. not initialized yet. This command line option has no effect if the
  2222. file is already initialized.
  2223. * systemd-nspawn gained a new --as-pid2 switch that invokes any
  2224. specified command line as PID 2 rather than PID 1 in the
  2225. container. In this mode PID 1 is a minimal stub init process that
  2226. implements the special POSIX and Linux semantics of PID 1 regarding
  2227. signal and child process management. Note that this stub init process
  2228. is implemented in nspawn itself and requires no support from the
  2229. container image. This new logic is useful to support running
  2230. arbitrary commands in the container, as normal processes are
  2231. generally not prepared to run as PID 1.
  2232. * systemd-nspawn gained a new --chdir= switch for setting the current
  2233. working directory for the process started in the container.
  2234. * "journalctl /dev/sda" will now output all kernel log messages for
  2235. specified device from the current boot, in addition to all devices
  2236. that are parents of it. This should make log output about devices
  2237. pretty useful, as long as kernel drivers attach enough metadata to
  2238. the log messages. (The usual SATA drivers do.)
  2239. * The sd-journal API gained two new calls
  2240. sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
  2241. that report whether log data from /run or /var has been found.
  2242. * journalctl gained a new switch "--fields" that prints all journal
  2243. record field names currently in use in the journal. This is backed
  2244. by two new sd-journal API calls sd_journal_enumerate_fields() and
  2245. sd_journal_restart_fields().
  2246. * Most configurable timeouts in systemd now expect an argument of
  2247. "infinity" to turn them off, instead of "0" as before. The semantics
  2248. from now on is that a timeout of "0" means "now", and "infinity"
  2249. means "never". To maintain backwards compatibility, "0" continues to
  2250. turn off previously existing timeout settings.
  2251. * "systemctl reload-or-try-restart" has been renamed to "systemctl
  2252. try-reload-or-restart" to clarify what it actually does: the "try"
  2253. logic applies to both reloading and restarting, not just restarting.
  2254. The old name continues to be accepted for compatibility.
  2255. * On boot-up, when PID 1 detects that the system clock is behind the
  2256. release date of the systemd version in use, the clock is now set
  2257. to the latter. Previously, this was already done in timesyncd, in order
  2258. to avoid running with clocks set to the various clock epochs such as
  2259. 1902, 1938 or 1970. With this change the logic is now done in PID 1
  2260. in addition to timesyncd during early boot-up, so that it is enforced
  2261. before the first process is spawned by systemd. Note that the logic
  2262. in timesyncd remains, as it is more comprehensive and ensures
  2263. clock monotonicity by maintaining a persistent timestamp file in
  2264. /var. Since /var is generally not available in earliest boot or the
  2265. initrd, this part of the logic remains in timesyncd, and is not done
  2266. by PID 1.
  2267. * Support for tweaking details in net_cls.class_id through the
  2268. NetClass= configuration directive has been removed, as the kernel
  2269. people have decided to deprecate that controller in cgroup v2.
  2270. Userspace tools such as nftables are moving over to setting rules
  2271. that are specific to the full cgroup path of a task, which obsoletes
  2272. these controllers anyway. The NetClass= directive is kept around for
  2273. legacy compatibility reasons. For a more in-depth description of the
  2274. kernel change, please refer to the respective upstream commit:
  2276. * A new service setting RuntimeMaxSec= has been added that may be used
  2277. to specify a maximum runtime for a service. If the timeout is hit, the
  2278. service is terminated and put into a failure state.
  2279. * A new service setting AmbientCapabilities= has been added. It allows
  2280. configuration of additional Linux process capabilities that are
  2281. passed to the activated processes. This is only available on very
  2282. recent kernels.
  2283. * The process resource limit settings in service units may now be used
  2284. to configure hard and soft limits individually.
  2285. * The various libsystemd APIs such as sd-bus or sd-event now publicly
  2286. expose support for gcc's __attribute__((cleanup())) C extension.
  2287. Specifically, for many object destructor functions alternative
  2288. versions have been added that have names suffixed with "p" and take a
  2289. pointer to a pointer to the object to destroy, instead of just a
  2290. pointer to the object itself. This is useful because these destructor
  2291. functions may be used directly as parameters to the cleanup
  2292. construct. Internally, systemd has been a heavy user of this GCC
  2293. extension for a long time, and with this change similar support is
  2294. now available to consumers of the library outside of systemd. Note
  2295. that by using this extension in your sources compatibility with old
  2296. and strictly ANSI compatible C compilers is lost. However, all gcc or
  2297. LLVM versions of recent years support this extension.
  2298. * Timer units gained support for a new setting RandomizedDelaySec= that
  2299. allows configuring some additional randomized delay to the configured
  2300. time. This is useful to spread out timer events to avoid load peaks in
  2301. clusters or larger setups.
  2302. * Calendar time specifications now support sub-second accuracy.
  2303. * Socket units now support listening on SCTP and UDP-lite protocol
  2304. sockets.
  2305. * The sd-event API now comes with a full set of man pages.
  2306. * Older versions of systemd contained experimental support for
  2307. compressing journal files and coredumps with the LZ4 compressor that
  2308. was not compatible with the lz4 binary (due to API limitations of the
  2309. lz4 library). This support has been removed; only support for files
  2310. compatible with the lz4 binary remains. This LZ4 logic is now
  2311. officially supported and no longer considered experimental.
  2312. * The dkr image import logic has been removed again from importd. dkr's
  2313. micro-services focus doesn't fit into the machine image focus of
  2314. importd, and quickly got out of date with the upstream dkr API.
  2315. * Creation of the /run/lock/lockdev/ directory was dropped from
  2316. tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
  2317. been available for many years. If you still need this, you need to
  2318. create your own tmpfiles.d config file with:
  2319. d /run/lock/lockdev 0775 root lock -
  2320. * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction=
  2321. and RebootArgument= have been moved from the [Service] section of
  2322. unit files to [Unit], and they are now supported on all unit types,
  2323. not just service units. Of course, elogind will continue to
  2324. understand these settings also at the old location, in order to
  2325. maintain compatibility.
  2326. Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
  2327. Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
  2328. Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
  2329. Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
  2330. Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
  2331. David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
  2332. Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
  2333. Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
  2334. Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
  2335. Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
  2336. Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
  2337. lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
  2338. Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
  2339. Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
  2340. Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
  2341. Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
  2342. Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
  2343. Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
  2344. Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
  2345. — Berlin, 2016-02-11
  2346. CHANGES WITH 228:
  2347. * A number of properties previously only settable in unit
  2348. files are now also available as properties to set when
  2349. creating transient units programmatically via the bus, as it
  2350. is exposed with systemd-run's --property=
  2351. setting. Specifically, these are: SyslogIdentifier=,
  2352. SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
  2353. EnvironmentFile=, ReadWriteDirectories=,
  2354. ReadOnlyDirectories=, InaccessibleDirectories=,
  2355. ProtectSystem=, ProtectHome=, RuntimeDirectory=.
  2356. * When creating transient services via the bus API it is now
  2357. possible to pass in a set of file descriptors to use as
  2358. STDIN/STDOUT/STDERR for the invoked process.
  2359. * Slice units may now be created transiently via the bus APIs,
  2360. similar to the way service and scope units may already be
  2361. created transiently.
  2362. * Wherever systemd expects a calendar timestamp specification
  2363. (like in journalctl's --since= and --until= switches) UTC
  2364. timestamps are now supported. Timestamps suffixed with "UTC"
  2365. are now considered to be in Universal Time Coordinated
  2366. instead of the local timezone. Also, timestamps may now
  2367. optionally be specified with sub-second accuracy. Both of
  2368. these additions also apply to recurring calendar event
  2369. specification, such as OnCalendar= in timer units.
  2370. * journalctl gained a new "--sync" switch that asks the
  2371. journal daemon to write all so far unwritten log messages to
  2372. disk and sync the files, before returning.
  2373. * systemd-tmpfiles learned two new line types "q" and "Q" that
  2374. operate like "v", but also set up a basic btrfs quota
  2375. hierarchy when used on a btrfs file system with quota
  2376. enabled.
  2377. * tmpfiles' "v", "q" and "Q" will now create a plain directory
  2378. instead of a subvolume (even on a btrfs file system) if the
  2379. root directory is a plain directory, and not a
  2380. subvolume. This should simplify things with certain chroot()
  2381. environments which are not aware of the concept of btrfs
  2382. subvolumes.
  2383. * systemd-detect-virt gained a new --chroot switch to detect
  2384. whether execution takes place in a chroot() environment.
  2385. * CPUAffinity= now takes CPU index ranges in addition to
  2386. individual indexes.
  2387. * The various memory-related resource limit settings (such as
  2388. LimitAS=) now understand the usual K, M, G, ... suffixes to
  2389. the base of 1024 (IEC). Similar, the time-related resource
  2390. limit settings understand the usual min, h, day, ...
  2391. suffixes now.
  2392. * There's a new system.conf setting DefaultTasksMax= to
  2393. control the default TasksMax= setting for services and
  2394. scopes running on the system. (TasksMax= is the primary
  2395. setting that exposes the "pids" cgroup controller on systemd
  2396. and was introduced in the previous systemd release.) The
  2397. setting now defaults to 512, which means services that are
  2398. not explicitly configured otherwise will only be able to
  2399. create 512 processes or threads at maximum, from this
  2400. version on. Note that this means that thread- or
  2401. process-heavy services might need to be reconfigured to set
  2402. TasksMax= to a higher value. It is sufficient to set
  2403. TasksMax= in these specific unit files to a higher value, or
  2404. even "infinity". Similar, there's now a logind.conf setting
  2405. UserTasksMax= that defaults to 4096 and limits the total
  2406. number of processes or tasks each user may own
  2407. concurrently. nspawn containers also have the TasksMax=
  2408. value set by default now, to 8192. Note that all of this
  2409. only has an effect if the "pids" cgroup controller is
  2410. enabled in the kernel. The general benefit of these changes
  2411. should be a more robust and safer system, that provides a
  2412. certain amount of per-service fork() bomb protection.
  2413. * systemd-nspawn gained the new --network-veth-extra= switch
  2414. to define additional and arbitrarily-named virtual Ethernet
  2415. links between the host and the container.
  2416. * A new service execution setting PassEnvironment= has been
  2417. added that allows importing select environment variables
  2418. from PID1's environment block into the environment block of
  2419. the service.
  2420. * Timer units gained support for a new RemainAfterElapse=
  2421. setting which takes a boolean argument. It defaults to on,
  2422. exposing behaviour unchanged to previous releases. If set to
  2423. off, timer units are unloaded after they elapsed if they
  2424. cannot elapse again. This is particularly useful for
  2425. transient timer units, which shall not stay around longer
  2426. than until they first elapse.
  2427. * systemd will now bump the net.unix.max_dgram_qlen to 512 by
  2428. default now (the kernel default is 16). This is beneficial
  2429. for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
  2430. allows substantially larger numbers of queued
  2431. datagrams. This should increase the capability of systemd to
  2432. parallelize boot-up, as logging and sd_notify() are unlikely
  2433. to stall execution anymore. If you need to change the value
  2434. from the new defaults, use the usual sysctl.d/ snippets.
  2435. * The compression framing format used by the journal or
  2436. coredump processing has changed to be in line with what the
  2437. official LZ4 tools generate. LZ4 compression support in
  2438. systemd was considered unsupported previously, as the format
  2439. was not compatible with the normal tools. With this release
  2440. this has changed now, and it is hence safe for downstream
  2441. distributions to turn it on. While not compressing as well
  2442. as the XZ, LZ4 is substantially faster, which makes
  2443. it a good default choice for the compression logic in the
  2444. journal and in coredump handling.
  2445. * Any reference to /etc/mtab has been dropped from
  2446. systemd. The file has been obsolete since a while, but
  2447. systemd refused to work on systems where it was incorrectly
  2448. set up (it should be a symlink or non-existent). Please make
  2449. sure to update to util-linux 2.27.1 or newer in conjunction
  2450. with this systemd release, which also drops any reference to
  2451. /etc/mtab. If you maintain a distribution make sure that no
  2452. software you package still references it, as this is a
  2453. likely source of bugs. There's also a glibc bug pending,
  2454. asking for removal of any reference to this obsolete file:
  2456. Note that only util-linux versions built with
  2457. --enable-libmount-force-mountinfo are supported.
  2458. * Support for the ".snapshot" unit type has been removed. This
  2459. feature turned out to be little useful and little used, and
  2460. has now been removed from the core and from systemctl.
  2461. * The dependency types RequiresOverridable= and
  2462. RequisiteOverridable= have been removed from systemd. They
  2463. have been used only very sparingly to our knowledge and
  2464. other options that provide a similar effect (such as
  2465. systemctl --mode=ignore-dependencies) are much more useful
  2466. and commonly used. Moreover, they were only half-way
  2467. implemented as the option to control behaviour regarding
  2468. these dependencies was never added to systemctl. By removing
  2469. these dependency types the execution engine becomes a bit
  2470. simpler. Unit files that use these dependencies should be
  2471. changed to use the non-Overridable dependency types
  2472. instead. In fact, when parsing unit files with these
  2473. options, that's what systemd will automatically convert them
  2474. too, but it will also warn, asking users to fix the unit
  2475. files accordingly. Removal of these dependency types should
  2476. only affect a negligible number of unit files in the wild.
  2477. * Behaviour of networkd's IPForward= option changed
  2478. (again). It will no longer maintain a per-interface setting,
  2479. but propagate one way from interfaces where this is enabled
  2480. to the global kernel setting. The global setting will be
  2481. enabled when requested by a network that is set up, but
  2482. never be disabled again. This change was made to make sure
  2483. IPv4 and IPv6 behaviour regarding packet forwarding is
  2484. similar (as the Linux IPv6 stack does not support
  2485. per-interface control of this setting) and to minimize
  2486. surprises.
  2487. * In unit files the behaviour of %u, %U, %h, %s has
  2488. changed. These specifiers will now unconditionally resolve
  2489. to the various user database fields of the user that the
  2490. systemd instance is running as, instead of the user
  2491. configured in the specific unit via User=. Note that this
  2492. effectively doesn't change much, as resolving of these
  2493. specifiers was already turned off in the --system instance
  2494. of systemd, as we cannot do NSS lookups from PID 1. In the
  2495. --user instance of systemd these specifiers where correctly
  2496. resolved, but hardly made any sense, since the user instance
  2497. lacks privileges to do user switches anyway, and User= is
  2498. hence useless. Moreover, even in the --user instance of
  2499. systemd behaviour was awkward as it would only take settings
  2500. from User= assignment placed before the specifier into
  2501. account. In order to unify and simplify the logic around
  2502. this the specifiers will now always resolve to the
  2503. credentials of the user invoking the manager (which in case
  2504. of PID 1 is the root user).
  2505. Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
  2506. Yang, Daniel Machon, Daniel Mack, David Herrmann, David
  2507. Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
  2508. Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
  2509. Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
  2510. Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
  2511. Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
  2512. Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
  2513. Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
  2514. Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
  2515. Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
  2516. Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
  2517. Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
  2518. Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
  2519. Jędrzejewski-Szmek
  2520. — Berlin, 2015-11-18
  2521. CHANGES WITH 227:
  2522. * systemd now depends on util-linux v2.27. More specifically,
  2523. the newly added mount monitor feature in libmount now
  2524. replaces systemd's former own implementation.
  2525. * libmount mandates /etc/mtab not to be regular file, and
  2526. systemd now enforces this condition at early boot.
  2527. /etc/mtab has been deprecated and warned about for a very
  2528. long time, so systems running systemd should already have
  2529. stopped having this file around as anything else than a
  2530. symlink to /proc/self/mounts.
  2531. * Support for the "pids" cgroup controller has been added. It
  2532. allows accounting the number of tasks in a cgroup and
  2533. enforcing limits on it. This adds two new setting
  2534. TasksAccounting= and TasksMax= to each unit, as well as a
  2535. global option DefaultTasksAccounting=.
  2536. * Support for the "net_cls" cgroup controller has been added.
  2537. It allows assigning a net class ID to each task in the
  2538. cgroup, which can then be used in firewall rules and traffic
  2539. shaping configurations. Note that the kernel netfilter net
  2540. class code does not currently work reliably for ingress
  2541. packets on unestablished sockets.
  2542. This adds a new config directive called NetClass= to CGroup
  2543. enabled units. Allowed values are positive numbers for fixed
  2544. assignments and "auto" for picking a free value
  2545. automatically.
  2546. * 'systemctl is-system-running' now returns 'offline' if the
  2547. system is not booted with systemd. This command can now be
  2548. used as a substitute for 'systemd-notify --booted'.
  2549. * Watchdog timeouts have been increased to 3 minutes for all
  2550. in-tree service files. Apparently, disk IO issues are more
  2551. frequent than we hoped, and user reported >1 minute waiting
  2552. for disk IO.
  2553. * 'machine-id-commit' functionality has been merged into
  2554. 'machine-id-setup --commit'. The separate binary has been
  2555. removed.
  2556. * The WorkingDirectory= directive in unit files may now be set
  2557. to the special value '~'. In this case, the working
  2558. directory is set to the home directory of the user
  2559. configured in User=.
  2560. * "machinectl shell" will now open the shell in the home
  2561. directory of the selected user by default.
  2562. * The CrashChVT= configuration file setting is renamed to
  2563. CrashChangeVT=, following our usual logic of not
  2564. abbreviating unnecessarily. The old directive is still
  2565. supported for compat reasons. Also, this directive now takes
  2566. an integer value between 1 and 63, or a boolean value. The
  2567. formerly supported '-1' value for disabling stays around for
  2568. compat reasons.
  2569. * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
  2570. NoNewPrivileges=, TTYPath=, WorkingDirectory= and
  2571. RootDirectory= properties can now be set for transient
  2572. units.
  2573. * The systemd-analyze tool gained a new "set-log-target" verb
  2574. to change the logging target the system manager logs to
  2575. dynamically during runtime. This is similar to how
  2576. "systemd-analyze set-log-level" already changes the log
  2577. level.
  2578. * In nspawn /sys is now mounted as tmpfs, with only a selected
  2579. set of subdirectories mounted in from the real sysfs. This
  2580. enhances security slightly, and is useful for ensuring user
  2581. namespaces work correctly.
  2582. * Support for USB FunctionFS activation has been added. This
  2583. allows implementation of USB gadget services that are
  2584. activated as soon as they are requested, so that they don't
  2585. have to run continuously, similar to classic socket
  2586. activation.
  2587. * The "systemctl exit" command now optionally takes an
  2588. additional parameter that sets the exit code to return from
  2589. the systemd manager when exiting. This is only relevant when
  2590. running the systemd user instance, or when running the
  2591. system instance in a container.
  2592. * sd-bus gained the new API calls sd_bus_path_encode_many()
  2593. and sd_bus_path_decode_many() that allow easy encoding and
  2594. decoding of multiple identifier strings inside a D-Bus
  2595. object path. Another new call sd_bus_default_flush_close()
  2596. has been added to flush and close per-thread default
  2597. connections.
  2598. * systemd-cgtop gained support for a -M/--machine= switch to
  2599. show the control groups within a certain container only.
  2600. * "systemctl kill" gained support for an optional --fail
  2601. switch. If specified the requested operation will fail of no
  2602. processes have been killed, because the unit had no
  2603. processes attached, or similar.
  2604. * A new systemd.crash_reboot=1 kernel command line option has
  2605. been added that triggers a reboot after crashing. This can
  2606. also be set through CrashReboot= in systemd.conf.
  2607. * The RuntimeDirectory= setting now understands unit
  2608. specifiers like %i or %f.
  2609. * A new (still internal) library API sd-ipv4acd has been added,
  2610. that implements address conflict detection for IPv4. It's
  2611. based on code from sd-ipv4ll, and will be useful for
  2612. detecting DHCP address conflicts.
  2613. * File descriptors passed during socket activation may now be
  2614. named. A new API sd_listen_fds_with_names() is added to
  2615. access the names. The default names may be overridden,
  2616. either in the .socket file using the FileDescriptorName=
  2617. parameter, or by passing FDNAME= when storing the file
  2618. descriptors using sd_notify().
  2619. * systemd-networkd gained support for:
  2620. - Setting the IPv6 Router Advertisement settings via
  2621. IPv6AcceptRouterAdvertisements= in .network files.
  2622. - Configuring the HelloTimeSec=, MaxAgeSec= and
  2623. ForwardDelaySec= bridge parameters in .netdev files.
  2624. - Configuring PreferredSource= for static routes in
  2625. .network files.
  2626. * The "ask-password" framework used to query for LUKS harddisk
  2627. passwords or SSL passwords during boot gained support for
  2628. caching passwords in the kernel keyring, if it is
  2629. available. This makes sure that the user only has to type in
  2630. a passphrase once if there are multiple objects to unlock
  2631. with the same one. Previously, such password caching was
  2632. available only when Plymouth was used; this moves the
  2633. caching logic into the systemd codebase itself. The
  2634. "systemd-ask-password" utility gained a new --keyname=
  2635. switch to control which kernel keyring key to use for
  2636. caching a password in. This functionality is also useful for
  2637. enabling display managers such as gdm to automatically
  2638. unlock the user's GNOME keyring if its passphrase, the
  2639. user's password and the harddisk password are the same, if
  2640. gdm-autologin is used.
  2641. * When downloading tar or raw images using "machinectl
  2642. pull-tar" or "machinectl pull-raw", a matching ".nspawn"
  2643. file is now also downloaded, if it is available and stored
  2644. next to the image file.
  2645. * Units of type ".socket" gained a new boolean setting
  2646. Writable= which is only useful in conjunction with
  2647. ListenSpecial=. If true, enables opening the specified
  2648. special file in O_RDWR mode rather than O_RDONLY mode.
  2649. * systemd-rfkill has been reworked to become a singleton
  2650. service that is activated through /dev/rfkill on each rfkill
  2651. state change and saves the settings to disk. This way,
  2652. systemd-rfkill is now compatible with devices that exist
  2653. only intermittendly, and even restores state if the previous
  2654. system shutdown was abrupt rather than clean.
  2655. * The journal daemon gained support for vacuuming old journal
  2656. files controlled by the number of files that shall remain,
  2657. in addition to the already existing control by size and by
  2658. date. This is useful as journal interleaving performance
  2659. degrades with too many separate journal files, and allows
  2660. putting an effective limit on them. The new setting defaults
  2661. to 100, but this may be changed by setting SystemMaxFiles=
  2662. and RuntimeMaxFiles= in journald.conf. Also, the
  2663. "journalctl" tool gained the new --vacuum-files= switch to
  2664. manually vacuum journal files to leave only the specified
  2665. number of files in place.
  2666. * udev will now create /dev/disk/by-path links for ATA devices
  2667. on kernels where that is supported.
  2668. * Galician, Serbian, Turkish and Korean translations were added.
  2669. Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
  2670. Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
  2671. (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
  2672. Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
  2673. Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
  2674. de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
  2675. Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
  2676. Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
  2677. Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
  2678. Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
  2679. Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
  2680. Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
  2681. Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
  2682. Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
  2683. Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
  2684. Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
  2685. Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
  2686. Zbigniew Jędrzejewski-Szmek, Марко М. Костић
  2687. — Berlin, 2015-10-07
  2688. CHANGES WITH 226:
  2689. * The DHCP implementation of systemd-networkd gained a set of
  2690. new features:
  2691. - The DHCP server now supports emitting DNS and NTP
  2692. information. It may be enabled and configured via
  2693. EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
  2694. and NTP information is enabled, but no servers are
  2695. configured, the corresponding uplink information (if there
  2696. is any) is propagated.
  2697. - Server and client now support transmission and reception
  2698. of timezone information. It can be configured via the
  2699. newly introduced network options UseTimezone=,
  2700. EmitTimezone=, and Timezone=. Transmission of timezone
  2701. information is enabled between host and containers by
  2702. default now: the container will change its local timezone
  2703. to what the host has set.
  2704. - Lease timeouts can now be configured via
  2705. MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
  2706. - The DHCP server improved on the stability of
  2707. leases. Clients are more likely to get the same lease
  2708. information back, even if the server loses state.
  2709. - The DHCP server supports two new configuration options to
  2710. control the lease address pool metrics, PoolOffset= and
  2711. PoolSize=.
  2712. * The encapsulation limit of tunnels in systemd-networkd may
  2713. now be configured via 'EncapsulationLimit='. It allows
  2714. modifying the maximum additional levels of encapsulation
  2715. that are permitted to be prepended to a packet.
  2716. * systemd now supports the concept of user buses replacing
  2717. session buses, if used with dbus-1.10 (and enabled via dbus
  2718. --enable-user-session). It previously only supported this on
  2719. kdbus-enabled systems, and this release expands this to
  2720. 'dbus-daemon' systems.
  2721. * systemd-networkd now supports predictable interface names
  2722. for virtio devices.
  2723. * systemd now optionally supports the new Linux kernel
  2724. "unified" control group hierarchy. If enabled via the kernel
  2725. command-line option 'systemd.unified_cgroup_hierarchy=1',
  2726. systemd will try to mount the unified cgroup hierarchy
  2727. directly on /sys/fs/cgroup. If not enabled, or not
  2728. available, systemd will fall back to the legacy cgroup
  2729. hierarchy setup, as before. Host system and containers can
  2730. mix and match legacy and unified hierarchies as they
  2731. wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
  2732. environment variable to individually select the hierarchy to
  2733. use for executed containers. By default, nspawn will use the
  2734. unified hierarchy for the containers if the host uses the
  2735. unified hierarchy, and the legacy hierarchy otherwise.
  2736. Please note that at this point the unified hierarchy is an
  2737. experimental kernel feature and is likely to change in one
  2738. of the next kernel releases. Therefore, it should not be
  2739. enabled by default in downstream distributions yet. The
  2740. minimum required kernel version for the unified hierarchy to
  2741. work is 4.2. Note that when the unified hierarchy is used
  2742. for the first time delegated access to controllers is
  2743. safe. Because of this systemd-nspawn containers will get
  2744. access to controllers now, as will systemd user
  2745. sessions. This means containers and user sessions may now
  2746. manage their own resources, partitioning up what the system
  2747. grants them.
  2748. * A new special scope unit "init.scope" has been introduced
  2749. that encapsulates PID 1 of the system. It may be used to
  2750. determine resource usage and enforce resource limits on PID
  2751. 1 itself. PID 1 hence moved out of the root of the control
  2752. group tree.
  2753. * The cgtop tool gained support for filtering out kernel
  2754. threads when counting tasks in a control group. Also, the
  2755. count of processes is now recursively summed up by
  2756. default. Two options -k and --recursive= have been added to
  2757. revert to old behaviour. The tool has also been updated to
  2758. work correctly in containers now.
  2759. * systemd-nspawn's --bind= and --bind-ro= options have been
  2760. extended to allow creation of non-recursive bind mounts.
  2761. * libsystemd gained two new calls sd_pid_get_cgroup() and
  2762. sd_peer_get_cgroup() which return the control group path of
  2763. a process or peer of a connected AF_UNIX socket. This
  2764. function call is particularly useful when implementing
  2765. delegated subtrees support in the control group hierarchy.
  2766. * The "sd-event" event loop API of libsystemd now supports
  2767. correct dequeuing of real-time signals, without losing
  2768. signal events.
  2769. * When systemd requests a PolicyKit decision when managing
  2770. units it will now add additional fields to the request,
  2771. including unit name and desired operation. This enables more
  2772. powerful PolicyKit policies, that make decisions depending
  2773. on these parameters.
  2774. * nspawn learnt support for .nspawn settings files, that may
  2775. accompany the image files or directories of containers, and
  2776. may contain additional settings for the container. This is
  2777. an alternative to configuring container parameters via the
  2778. nspawn command line.
  2779. Contributions from: Cristian Rodríguez, Daniel Mack, David
  2780. Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
  2781. Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
  2782. Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
  2783. Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
  2784. Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
  2785. Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
  2786. Andersen, Tom Gundersen, Torstein Husebø
  2787. — Berlin, 2015-09-08
  2788. CHANGES WITH 225:
  2789. * machinectl gained a new verb 'shell' which opens a fresh
  2790. shell on the target container or the host. It is similar to
  2791. the existing 'login' command of machinectl, but spawns the
  2792. shell directly without prompting for username or
  2793. password. The pseudo machine '.host' now refers to the local
  2794. host and is used by default. Hence, 'machinectl shell' can
  2795. be used as replacement for 'su -' which spawns a session as
  2796. a fresh systemd unit in a way that is fully isolated from
  2797. the originating session.
  2798. * systemd-networkd learned to cope with private-zone DHCP
  2799. options and allows other programs to query the values.
  2800. * SELinux access control when enabling/disabling units is no
  2801. longer enforced with this release. The previous
  2802. implementation was incorrect, and a new corrected
  2803. implementation is not yet available. As unit file operations
  2804. are still protected via PolicyKit and D-Bus policy this is
  2805. not a security problem. Yet, distributions which care about
  2806. optimal SELinux support should probably not stabilize on
  2807. this release.
  2808. * sd-bus gained support for matches of type "arg0has=", that
  2809. test for membership of strings in string arrays sent in bus
  2810. messages.
  2811. * systemd-resolved now dumps the contents of its DNS and LLMNR
  2812. caches to the logs on reception of the SIGUSR1 signal. This
  2813. is useful to debug DNS behaviour.
  2814. * The coredumpctl tool gained a new --directory= option to
  2815. operate on journal files in a specific directory.
  2816. * "systemctl reboot" and related commands gained a new
  2817. "--message=" option which may be used to set a free-text
  2818. wall message when shutting down or rebooting the
  2819. system. This message is also logged, which is useful for
  2820. figuring out the reason for a reboot or shutdown a
  2821. posteriori.
  2822. * The "systemd-resolve-host" tool's -i switch now takes
  2823. network interface numbers as alternative to interface names.
  2824. * A new unit file setting for services has been introduced:
  2825. UtmpMode= allows configuration of how precisely systemd
  2826. handles utmp and wtmp entries for the service if this is
  2827. enabled. This allows writing services that appear similar to
  2828. user sessions in the output of the "w", "who", "last" and
  2829. "lastlog" tools.
  2830. * systemd-resolved will now locally synthesize DNS resource
  2831. records for the "localhost" and "gateway" domains as well as
  2832. the local hostname. This should ensure that clients querying
  2833. RRs via resolved will get similar results as those going via
  2834. NSS, if nss-myhostname is enabled.
  2835. Contributions from: Alastair Hughes, Alex Crawford, Daniel
  2836. Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
  2837. Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
  2838. Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
  2839. Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
  2840. Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
  2841. Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
  2842. Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
  2843. reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
  2844. Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
  2845. Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
  2846. WaLyong Cho, Zbigniew Jędrzejewski-Szmek
  2847. — Berlin, 2015-08-27
  2848. CHANGES WITH 224:
  2849. * The systemd-efi-boot-generator functionality was merged into
  2850. systemd-gpt-auto-generator.
  2851. * systemd-networkd now supports Group Policy for vxlan
  2852. devices. It can be enabled via the new boolean configuration
  2853. option called 'GroupPolicyExtension='.
  2854. Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
  2855. Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
  2856. Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
  2857. — Berlin, 2015-07-31
  2858. CHANGES WITH 223:
  2859. * The python-systemd code has been removed from the systemd repository.
  2860. A new repository has been created which accommodates the code from
  2861. now on, and we kindly ask distributions to create a separate package
  2862. for this:
  2863. * The systemd daemon will now reload its main configuration
  2864. (/etc/systemd/system.conf) on daemon-reload.
  2865. * sd-dhcp now exposes vendor specific extensions via
  2866. sd_dhcp_lease_get_vendor_specific().
  2867. * systemd-networkd gained a number of new configuration options.
  2868. - A new boolean configuration option for TAP devices called
  2869. 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
  2870. device, thus allowing to send and receive GSO packets.
  2871. - A new tunnel configuration option called 'CopyDSCP='.
  2872. If enabled, the DSCP field of ip6 tunnels is copied into the
  2873. decapsulated packet.
  2874. - A set of boolean bridge configuration options were added.
  2875. 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
  2876. and 'UnicastFlood=' are now parsed by networkd and applied to the
  2877. respective bridge link device via the respective IFLA_BRPORT_*
  2878. netlink attribute.
  2879. - A new string configuration option to override the hostname sent
  2880. to a DHCP server, called 'Hostname='. If set and 'SendHostname='
  2881. is true, networkd will use the configured hostname instead of the
  2882. system hostname when sending DHCP requests.
  2883. - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
  2884. networkd will configure the IPv6 flow-label of the tunnel device
  2885. according to RFC2460.
  2886. - The 'macvtap' virtual network devices are now supported, similar to
  2887. the already supported 'macvlan' devices.
  2888. * systemd-resolved now implements RFC5452 to improve resilience against
  2889. cache poisoning. Additionally, source port randomization is enabled
  2890. by default to further protect against DNS spoofing attacks.
  2891. * nss-mymachines now supports translating UIDs and GIDs of running
  2892. containers with user-namespaces enabled. If a container 'foo'
  2893. translates a host uid 'UID' to the container uid 'TUID', then
  2894. nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
  2895. (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
  2896. mapped as 'vg-foo-TGID'.
  2897. Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
  2898. Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
  2899. HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
  2900. Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
  2901. Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
  2902. Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
  2903. Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
  2904. Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
  2905. Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
  2906. Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
  2907. — Berlin, 2015-07-29
  2908. CHANGES WITH 222:
  2909. * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
  2910. There are no known issues with current sysfs, and udev does not need
  2911. or should be used to work around such bugs.
  2912. * udev does no longer enable USB HID power management. Several reports
  2913. indicate, that some devices cannot handle that setting.
  2914. * The udev accelerometer helper was removed. The functionality
  2915. is now fully included in iio-sensor-proxy. But this means,
  2916. older iio-sensor-proxy versions will no longer provide
  2917. accelerometer/orientation data with this systemd version.
  2918. Please upgrade iio-sensor-proxy to version 1.0.
  2919. * networkd gained a new configuration option IPv6PrivacyExtensions=
  2920. which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
  2921. for Stateless Address") on selected networks.
  2922. * For the sake of fewer build-time dependencies and less code in the
  2923. main repository, the python bindings are about to be removed in the
  2924. next release. A new repository has been created which accommodates
  2925. the code from now on, and we kindly ask distributions to create a
  2926. separate package for this. The removal will take place in v223.
  2928. Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
  2929. Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
  2930. daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
  2931. Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
  2932. Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
  2933. (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
  2934. Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
  2935. Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
  2936. Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
  2937. Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
  2938. — Berlin, 2015-07-07
  2939. CHANGES WITH 221:
  2940. * The sd-bus.h and sd-event.h APIs have now been declared
  2941. stable and have been added to the official interface of
  2942. sd-bus implements an alternative D-Bus client
  2943. library, that is relatively easy to use, very efficient and
  2944. supports both classic D-Bus as well as kdbus as transport
  2945. backend. sd-event is a generic event loop abstraction that
  2946. is built around Linux epoll, but adds features such as event
  2947. prioritization or efficient timer handling. Both APIs are good
  2948. choices for C programs looking for a bus and/or event loop
  2949. implementation that is minimal and does not have to be
  2950. portable to other kernels.
  2951. * kdbus support is no longer compile-time optional. It is now
  2952. always built-in. However, it can still be disabled at
  2953. runtime using the kdbus=0 kernel command line setting, and
  2954. that setting may be changed to default to off, by specifying
  2955. --disable-kdbus at build-time. Note though that the kernel
  2956. command line setting has no effect if the kdbus.ko kernel
  2957. module is not installed, in which case kdbus is (obviously)
  2958. also disabled. We encourage all downstream distributions to
  2959. begin testing kdbus by adding it to the kernel images in the
  2960. development distributions, and leaving kdbus support in
  2961. systemd enabled.
  2962. * The minimal required util-linux version has been bumped to
  2963. 2.26.
  2964. * Support for chkconfig (--enable-chkconfig) was removed in
  2965. favor of calling an abstraction tool
  2966. /lib/systemd/systemd-sysv-install. This needs to be
  2967. implemented for your distribution. See "SYSV INIT.D SCRIPTS"
  2968. in README for details.
  2969. * If there's a systemd unit and a SysV init script for the
  2970. same service name, and the user executes "systemctl enable"
  2971. for it (or a related call), then this will now enable both
  2972. (or execute the related operation on both), not just the
  2973. unit.
  2974. * The libudev API documentation has been converted from gtkdoc
  2975. into man pages.
  2976. * gudev has been removed from the systemd tree, it is now an
  2977. external project.
  2978. * The systemd-cgtop tool learnt a new --raw switch to generate
  2979. "raw" (machine parsable) output.
  2980. * networkd's IPForwarding= .network file setting learnt the
  2981. new setting "kernel", which ensures that networkd does not
  2982. change the IP forwarding sysctl from the default kernel
  2983. state.
  2984. * The systemd-logind bus API now exposes a new boolean
  2985. property "Docked" that reports whether logind considers the
  2986. system "docked", i.e. connected to a docking station or not.
  2987. Contributions from: Alex Crawford, Andreas Pokorny, Andrei
  2988. Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
  2989. Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
  2990. David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
  2991. Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
  2992. Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
  2993. Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
  2994. Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
  2995. Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
  2996. Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
  2997. Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
  2998. Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
  2999. Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
  3000. Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
  3001. Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
  3002. Fink, Zbigniew Jędrzejewski-Szmek
  3003. — Berlin, 2015-06-19
  3004. CHANGES WITH 220:
  3005. * The gudev library has been extracted into a separate repository
  3006. available at:
  3007. It is now managed as part of the Gnome project. Distributions
  3008. are recommended to pass --disable-gudev to systemd and use
  3009. gudev from the Gnome project instead. gudev is still included
  3010. in systemd, for now. It will be removed soon, though. Please
  3011. also see the announcement-thread on systemd-devel:
  3013. * systemd now exposes a CPUUsageNSec= property for each
  3014. service unit on the bus, that contains the overall consumed
  3015. CPU time of a service (the sum of what each process of the
  3016. service consumed). This value is only available if
  3017. CPUAccounting= is turned on for a service, and is then shown
  3018. in the "systemctl status" output.
  3019. * Support for configuring alternative mappings of the old SysV
  3020. runlevels to systemd targets has been removed. They are now
  3021. hardcoded in a way that runlevels 2, 3, 4 all map to
  3022. and 5 to (which
  3023. previously was already the default behaviour).
  3024. * The auto-mounter logic gained support for mount point
  3025. expiry, using a new TimeoutIdleSec= setting in .automount
  3026. units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
  3027. * The EFI System Partition (ESP) as mounted to /boot by
  3028. systemd-efi-boot-generator will now be unmounted
  3029. automatically after 2 minutes of not being used. This should
  3030. minimize the risk of ESP corruptions.
  3031. * New /etc/fstab options x-systemd.requires= and
  3032. x-systemd.requires-mounts-for= are now supported to express
  3033. additional dependencies for mounts. This is useful for
  3034. journalling file systems that support external journal
  3035. devices or overlay file systems that require underlying file
  3036. systems to be mounted.
  3037. * systemd does not support direct live-upgrades (via systemctl
  3038. daemon-reexec) from versions older than v44 anymore. As no
  3039. distribution we are aware of shipped such old versions in a
  3040. stable release this should not be problematic.
  3041. * When systemd forks off a new per-connection service instance
  3042. it will now set the $REMOTE_ADDR environment variable to the
  3043. remote IP address, and $REMOTE_PORT environment variable to
  3044. the remote IP port. This behaviour is similar to the
  3045. corresponding environment variables defined by CGI.
  3046. * systemd-networkd gained support for uplink failure
  3047. detection. The BindCarrier= option allows binding interface
  3048. configuration dynamically to the link sense of other
  3049. interfaces. This is useful to achieve behaviour like in
  3050. network switches.
  3051. * systemd-networkd gained support for configuring the DHCP
  3052. client identifier to use when requesting leases.
  3053. * systemd-networkd now has a per-network UseNTP= option to
  3054. configure whether NTP server information acquired via DHCP
  3055. is passed on to services like systemd-timesyncd.
  3056. * systemd-networkd gained support for vti6 tunnels.
  3057. * Note that systemd-networkd manages the sysctl variable
  3058. /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
  3059. it is configured for since v219. The variable controls IP
  3060. forwarding, and is a per-interface alternative to the global
  3061. /proc/sys/net/ipv[46]/ip_forward. This setting is
  3062. configurable in the IPForward= option, which defaults to
  3063. "no". This means if networkd is used for an interface it is
  3064. no longer sufficient to set the global sysctl option to turn
  3065. on IP forwarding! Instead, the .network file option
  3066. IPForward= needs to be turned on! Note that the
  3067. implementation of this behaviour was broken in v219 and has
  3068. been fixed in v220.
  3069. * Many bonding and vxlan options are now configurable in
  3070. systemd-networkd.
  3071. * systemd-nspawn gained a new --property= setting to set unit
  3072. properties for the container scope. This is useful for
  3073. setting resource parameters (e.g. "CPUShares=500") on
  3074. containers started from the command line.
  3075. * systemd-nspawn gained a new --private-users= switch to make
  3076. use of user namespacing available on recent Linux kernels.
  3077. * systemd-nspawn may now be called as part of a shell pipeline
  3078. in which case the pipes used for stdin and stdout are passed
  3079. directly to the process invoked in the container, without
  3080. indirection via a pseudo tty.
  3081. * systemd-nspawn gained a new switch to control the UNIX
  3082. signal to use when killing the init process of the container
  3083. when shutting down.
  3084. * systemd-nspawn gained a new --overlay= switch for mounting
  3085. overlay file systems into the container using the new kernel
  3086. overlayfs support.
  3087. * When a container image is imported via systemd-importd and
  3088. the host file system is not btrfs, a loopback block device
  3089. file is created in /var/lib/machines.raw with a btrfs file
  3090. system inside. It is then mounted to /var/lib/machines to
  3091. enable btrfs features for container management. The loopback
  3092. file and btrfs file system is grown as needed when container
  3093. images are imported via systemd-importd.
  3094. * systemd-machined/systemd-importd gained support for btrfs
  3095. quota, to enforce container disk space limits on disk. This
  3096. is exposed in "machinectl set-limit".
  3097. * systemd-importd now can import containers from local .tar,
  3098. .raw and .qcow2 images, and export them to .tar and .raw. It
  3099. can also import dkr v2 images now from the network (on top
  3100. of v1 as before).
  3101. * systemd-importd gained support for verifying downloaded
  3102. images with gpg2 (previously only gpg1 was supported).
  3103. * systemd-machined, systemd-logind, systemd: most bus calls
  3104. are now accessible to unprivileged processes via
  3105. PolicyKit. Also, systemd-logind will now allow users to kill
  3106. their own sessions without further privileges or
  3107. authorization.
  3108. * systemd-shutdownd has been removed. This service was
  3109. previously responsible for implementing scheduled shutdowns
  3110. as exposed in /usr/bin/shutdown's time parameter. This
  3111. functionality has now been moved into systemd-logind and is
  3112. accessible via a bus interface.
  3113. * "systemctl reboot" gained a new switch --firmware-setup that
  3114. can be used to reboot into the EFI firmware setup, if that
  3115. is available. systemd-logind now exposes an API on the bus
  3116. to trigger such reboots, in case graphical desktop UIs want
  3117. to cover this functionality.
  3118. * "systemctl enable", "systemctl disable" and "systemctl mask"
  3119. now support a new "--now" switch. If specified the units
  3120. that are enabled will also be started, and the ones
  3121. disabled/masked also stopped.
  3122. * The Gummiboot EFI boot loader tool has been merged into
  3123. systemd, and renamed to "systemd-boot". The bootctl tool has been
  3124. updated to support systemd-boot.
  3125. * An EFI kernel stub has been added that may be used to create
  3126. kernel EFI binaries that contain not only the actual kernel,
  3127. but also an initrd, boot splash, command line and OS release
  3128. information. This combined binary can then be signed as a
  3129. single image, so that the firmware can verify it all in one
  3130. step. systemd-boot has special support for EFI binaries created
  3131. like this and can extract OS release information from them
  3132. and show them in the boot menu. This functionality is useful
  3133. to implement cryptographically verified boot schemes.
  3134. * Optional support has been added to systemd-fsck to pass
  3135. fsck's progress report to an AF_UNIX socket in the file
  3136. system.
  3137. * udev will no longer create device symlinks for all block
  3138. devices by default. A blacklist for excluding special block
  3139. devices from this logic has been turned into a whitelist
  3140. that requires picking block devices explicitly that require
  3141. device symlinks.
  3142. * A new (currently still internal) API sd-device.h has been
  3143. added to libsystemd. This modernized API is supposed to
  3144. replace libudev eventually. In fact, already much of libudev
  3145. is now just a wrapper around sd-device.h.
  3146. * A new hwdb database for storing metadata about pointing
  3147. stick devices has been added.
  3148. * systemd-tmpfiles gained support for setting file attributes
  3149. similar to the "chattr" tool with new 'h' and 'H' lines.
  3150. * systemd-journald will no longer unconditionally set the
  3151. btrfs NOCOW flag on new journal files. This is instead done
  3152. with tmpfiles snippet using the new 'h' line type. This
  3153. allows easy disabling of this logic, by masking the
  3154. journal-nocow.conf tmpfiles file.
  3155. * systemd-journald will now translate audit message types to
  3156. human readable identifiers when writing them to the
  3157. journal. This should improve readability of audit messages.
  3158. * The LUKS logic gained support for the offset= and skip=
  3159. options in /etc/crypttab, as previously implemented by
  3160. Debian.
  3161. * /usr/lib/os-release gained a new optional field VARIANT= for
  3162. distributions that support multiple variants (such as a
  3163. desktop edition, a server edition, ...)
  3164. Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
  3165. Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
  3166. Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
  3167. Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
  3168. Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
  3169. Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
  3170. Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
  3171. Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
  3172. Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
  3173. Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
  3174. Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
  3175. Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
  3176. Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
  3177. Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
  3178. De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
  3179. Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
  3180. Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
  3181. Michael Biebl, Michael Marineau, Michael Olbrich, Michal
  3182. Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
  3183. Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
  3184. Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
  3185. Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
  3186. Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
  3187. Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
  3188. Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
  3189. Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
  3190. Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
  3191. — Berlin, 2015-05-22
  3192. CHANGES WITH 219:
  3193. * Introduce a new API "sd-hwdb.h" for querying the hardware
  3194. metadata database. With this minimal interface one can query
  3195. and enumerate the udev hwdb, decoupled from the old libudev
  3196. library. libudev's interface for this is now only a wrapper
  3197. around sd-hwdb. A new tool systemd-hwdb has been added to
  3198. interface with and update the database.
  3199. * When any of systemd's tools copies files (for example due to
  3200. tmpfiles' C lines) a btrfs reflink will attempted first,
  3201. before bytewise copying is done.
  3202. * systemd-nspawn gained a new --ephemeral switch. When
  3203. specified a btrfs snapshot is taken of the container's root
  3204. directory, and immediately removed when the container
  3205. terminates again. Thus, a container can be started whose
  3206. changes never alter the container's root directory, and are
  3207. lost on container termination. This switch can also be used
  3208. for starting a container off the root file system of the
  3209. host without affecting the host OS. This switch is only
  3210. available on btrfs file systems.
  3211. * systemd-nspawn gained a new --template= switch. It takes the
  3212. path to a container tree to use as template for the tree
  3213. specified via --directory=, should that directory be
  3214. missing. This allows instantiating containers dynamically,
  3215. on first run. This switch is only available on btrfs file
  3216. systems.
  3217. * When a .mount unit refers to a mount point on which multiple
  3218. mounts are stacked, and the .mount unit is stopped all of
  3219. the stacked mount points will now be unmounted until no
  3220. mount point remains.
  3221. * systemd now has an explicit notion of supported and
  3222. unsupported unit types. Jobs enqueued for unsupported unit
  3223. types will now fail with an "unsupported" error code. More
  3224. specifically .swap, .automount and .device units are not
  3225. supported in containers, .busname units are not supported on
  3226. non-kdbus systems. .swap and .automount are also not
  3227. supported if their respective kernel compile time options
  3228. are disabled.
  3229. * machinectl gained support for two new "copy-from" and
  3230. "copy-to" commands for copying files from a running
  3231. container to the host or vice versa.
  3232. * machinectl gained support for a new "bind" command to bind
  3233. mount host directories into local containers. This is
  3234. currently only supported for nspawn containers.
  3235. * networkd gained support for configuring bridge forwarding
  3236. database entries (fdb) from .network files.
  3237. * A new tiny daemon "systemd-importd" has been added that can
  3238. download container images in tar, raw, qcow2 or dkr formats,
  3239. and make them available locally in /var/lib/machines, so
  3240. that they can run as nspawn containers. The daemon can GPG
  3241. verify the downloads (not supported for dkr, since it has no
  3242. provisions for verifying downloads). It will transparently
  3243. decompress bz2, xz, gzip compressed downloads if necessary,
  3244. and restore sparse files on disk. The daemon uses privilege
  3245. separation to ensure the actual download logic runs with
  3246. fewer privileges than the daemon itself. machinectl has
  3247. gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
  3248. make the functionality of importd available to the
  3249. user. With this in place the Fedora and Ubuntu "Cloud"
  3250. images can be downloaded and booted as containers unmodified
  3251. (the Fedora images lack the appropriate GPG signature files
  3252. currently, so they cannot be verified, but this will change
  3253. soon, hopefully). Note that downloading images is currently
  3254. only fully supported on btrfs.
  3255. * machinectl is now able to list container images found in
  3256. /var/lib/machines, along with some metadata about sizes of
  3257. disk and similar. If the directory is located on btrfs and
  3258. quota is enabled, this includes quota display. A new command
  3259. "image-status" has been added that shows additional
  3260. information about images.
  3261. * machinectl is now able to clone container images
  3262. efficiently, if the underlying file system (btrfs) supports
  3263. it, with the new "machinectl clone" command. It also
  3264. gained commands for renaming and removing images, as well as
  3265. marking them read-only or read-write (supported also on
  3266. legacy file systems).
  3267. * networkd gained support for collecting LLDP network
  3268. announcements, from hardware that supports this. This is
  3269. shown in networkctl output.
  3270. * systemd-run gained support for a new -t (--pty) switch for
  3271. invoking a binary on a pty whose input and output is
  3272. connected to the invoking terminal. This allows executing
  3273. processes as system services while interactively
  3274. communicating with them via the terminal. Most interestingly
  3275. this is supported across container boundaries. Invoking
  3276. "systemd-run -t /bin/bash" is an alternative to running a
  3277. full login session, the difference being that the former
  3278. will not register a session, nor go through the PAM session
  3279. setup.
  3280. * tmpfiles gained support for a new "v" line type for creating
  3281. btrfs subvolumes. If the underlying file system is a legacy
  3282. file system, this automatically degrades to creating a
  3283. normal directory. Among others /var/lib/machines is now
  3284. created like this at boot, should it be missing.
  3285. * The directory /var/lib/containers/ has been deprecated and
  3286. been replaced by /var/lib/machines. The term "machines" has
  3287. been used in the systemd context as generic term for both
  3288. VMs and containers, and hence appears more appropriate for
  3289. this, as the directory can also contain raw images bootable
  3290. via qemu/kvm.
  3291. * systemd-nspawn when invoked with -M but without --directory=
  3292. or --image= is now capable of searching for the container
  3293. root directory, subvolume or disk image automatically, in
  3294. /var/lib/machines. systemd-nspawn@.service has been updated
  3295. to make use of this, thus allowing it to be used for raw
  3296. disk images, too.
  3297. * A new unit has been introduced that is
  3298. supposed to group all containers/VMs invoked as services on
  3299. the system. systemd-nspawn@.service has been updated to
  3300. integrate with that.
  3301. * machinectl gained a new "start" command, for invoking a
  3302. container as a service. "machinectl start foo" is mostly
  3303. equivalent to "systemctl start systemd-nspawn@foo.service",
  3304. but handles escaping in a nicer way.
  3305. * systemd-nspawn will now mount most of the cgroupfs tree
  3306. read-only into each container, with the exception of the
  3307. container's own subtree in the name=systemd hierarchy.
  3308. * journald now sets the special FS_NOCOW file flag for its
  3309. journal files. This should improve performance on btrfs, by
  3310. avoiding heavy fragmentation when journald's write-pattern
  3311. is used on COW file systems. It degrades btrfs' data
  3312. integrity guarantees for the files to the same levels as for
  3313. ext3/ext4 however. This should be OK though as journald does
  3314. its own data integrity checks and all its objects are
  3315. checksummed on disk. Also, journald should handle btrfs disk
  3316. full events a lot more gracefully now, by processing SIGBUS
  3317. errors, and not relying on fallocate() anymore.
  3318. * When journald detects that journal files it is writing to
  3319. have been deleted it will immediately start new journal
  3320. files.
  3321. * systemd now provides a way to store file descriptors
  3322. per-service in PID 1. This is useful for daemons to ensure
  3323. that fds they require are not lost during a daemon
  3324. restart. The fds are passed to the daemon on the next
  3325. invocation in the same way socket activation fds are
  3326. passed. This is now used by journald to ensure that the
  3327. various sockets connected to all the system's stdout/stderr
  3328. are not lost when journald is restarted. File descriptors
  3329. may be stored in PID 1 via the sd_pid_notify_with_fds() API,
  3330. an extension to sd_notify(). Note that a limit is enforced
  3331. on the number of fds a service can store in PID 1, and it
  3332. defaults to 0, so that no fds may be stored, unless this is
  3333. explicitly turned on.
  3334. * The default TERM variable to use for units connected to a
  3335. terminal, when no other value is explicitly is set is now
  3336. vt220 rather than vt102. This should be fairly safe still,
  3337. but allows PgUp/PgDn work.
  3338. * The /etc/crypttab option header= as known from Debian is now
  3339. supported.
  3340. * "loginctl user-status" and "loginctl session-status" will
  3341. now show the last 10 lines of log messages of the
  3342. user/session following the status output. Similar,
  3343. "machinectl status" will show the last 10 log lines
  3344. associated with a virtual machine or container
  3345. service. (Note that this is usually not the log messages
  3346. done in the VM/container itself, but simply what the
  3347. container manager logs. For nspawn this includes all console
  3348. output however.)
  3349. * "loginctl session-status" without further argument will now
  3350. show the status of the session of the caller. Similar,
  3351. "lock-session", "unlock-session", "activate",
  3352. "enable-linger", "disable-linger" may now be called without
  3353. session/user parameter in which case they apply to the
  3354. caller's session/user.
  3355. * An X11 session scriptlet is now shipped that uploads
  3356. $DISPLAY and $XAUTHORITY into the environment of the systemd
  3357. --user daemon if a session begins. This should improve
  3358. compatibility with X11 enabled applications run as systemd
  3359. user services.
  3360. * Generators are now subject to masking via /etc and /run, the
  3361. same way as unit files.
  3362. * networkd .network files gained support for configuring
  3363. per-link IPv4/IPv6 packet forwarding as well as IPv4
  3364. masquerading. This is by default turned on for veth links to
  3365. containers, as registered by systemd-nspawn. This means that
  3366. nspawn containers run with --network-veth will now get
  3367. automatic routed access to the host's networks without any
  3368. further configuration or setup, as long as networkd runs on
  3369. the host.
  3370. * systemd-nspawn gained the --port= (-p) switch to expose TCP
  3371. or UDP posts of a container on the host. With this in place
  3372. it is possible to run containers with private veth links
  3373. (--network-veth), and have their functionality exposed on
  3374. the host as if their services were running directly on the
  3375. host.
  3376. * systemd-nspawn's --network-veth switch now gained a short
  3377. version "-n", since with the changes above it is now truly
  3378. useful out-of-the-box. The systemd-nspawn@.service has been
  3379. updated to make use of it too by default.
  3380. * systemd-nspawn will now maintain a per-image R/W lock, to
  3381. ensure that the same image is not started more than once
  3382. writable. (It's OK to run an image multiple times
  3383. simultaneously in read-only mode.)
  3384. * systemd-nspawn's --image= option is now capable of
  3385. dissecting and booting MBR and GPT disk images that contain
  3386. only a single active Linux partition. Previously it
  3387. supported only GPT disk images with proper GPT type
  3388. IDs. This allows running cloud images from major
  3389. distributions directly with systemd-nspawn, without
  3390. modification.
  3391. * In addition to collecting mouse dpi data in the udev
  3392. hardware database, there's now support for collecting angle
  3393. information for mouse scroll wheels. The database is
  3394. supposed to guarantee similar scrolling behavior on mice
  3395. that it knows about. There's also support for collecting
  3396. information about Touchpad types.
  3397. * udev's input_id built-in will now also collect touch screen
  3398. dimension data and attach it to probed devices.
  3399. * /etc/os-release gained support for a Distribution Privacy
  3400. Policy link field.
  3401. * networkd gained support for creating "ipvlan", "gretap",
  3402. "ip6gre", "ip6gretap" and "ip6tnl" network devices.
  3403. * systemd-tmpfiles gained support for "a" lines for setting
  3404. ACLs on files.
  3405. * systemd-nspawn will now mount /tmp in the container to
  3406. tmpfs, automatically.
  3407. * systemd now exposes the memory.usage_in_bytes cgroup
  3408. attribute and shows it for each service in the "systemctl
  3409. status" output, if available.
  3410. * When the user presses Ctrl-Alt-Del more than 7x within 2s an
  3411. immediate reboot is triggered. This useful if shutdown is
  3412. hung and is unable to complete, to expedite the
  3413. operation. Note that this kind of reboot will still unmount
  3414. all file systems, and hence should not result in fsck being
  3415. run on next reboot.
  3416. * A .device unit for an optical block device will now be
  3417. considered active only when a medium is in the drive. Also,
  3418. mount units are now bound to their backing devices thus
  3419. triggering automatic unmounting when devices become
  3420. unavailable. With this in place systemd will now
  3421. automatically unmount left-over mounts when a CD-ROM is
  3422. ejected or an USB stick is yanked from the system.
  3423. * networkd-wait-online now has support for waiting for
  3424. specific interfaces only (with globbing), and for giving up
  3425. after a configurable timeout.
  3426. * networkd now exits when idle. It will be automatically
  3427. restarted as soon as interfaces show up, are removed or
  3428. change state. networkd will stay around as long as there is
  3429. at least one DHCP state machine or similar around, that keep
  3430. it non-idle.
  3431. * networkd may now configure IPv6 link-local addressing in
  3432. addition to IPv4 link-local addressing.
  3433. * The IPv6 "token" for use in SLAAC may now be configured for
  3434. each .network interface in networkd.
  3435. * Routes configured with networkd may now be assigned a scope
  3436. in .network files.
  3437. * networkd's [Match] sections now support globbing and lists
  3438. of multiple space-separated matches per item.
  3439. Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
  3440. Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
  3441. Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
  3442. Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
  3443. Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
  3444. Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
  3445. Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
  3446. Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
  3447. Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
  3448. Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
  3449. Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
  3450. Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
  3451. Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
  3452. Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
  3453. Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
  3454. Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
  3455. Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
  3456. Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
  3457. Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
  3458. Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
  3459. Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
  3460. Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
  3461. Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
  3462. Hoffmann, Zbigniew Jędrzejewski-Szmek
  3463. — Berlin, 2015-02-16
  3464. CHANGES WITH 218:
  3465. * When querying unit file enablement status (for example via
  3466. "systemctl is-enabled"), a new state "indirect" is now known
  3467. which indicates that a unit might not be enabled itself, but
  3468. another unit listed in its Also= setting might be.
  3469. * Similar to the various existing ConditionXYZ= settings for
  3470. units, there are now matching AssertXYZ= settings. While
  3471. failing conditions cause a unit to be skipped, but its job
  3472. to succeed, failing assertions declared like this will cause
  3473. a unit start operation and its job to fail.
  3474. * hostnamed now knows a new chassis type "embedded".
  3475. * systemctl gained a new "edit" command. When used on a unit
  3476. file, this allows extending unit files with .d/ drop-in
  3477. configuration snippets or editing the full file (after
  3478. copying it from /usr/lib to /etc). This will invoke the
  3479. user's editor (as configured with $EDITOR), and reload the
  3480. modified configuration after editing.
  3481. * "systemctl status" now shows the suggested enablement state
  3482. for a unit, as declared in the (usually vendor-supplied)
  3483. system preset files.
  3484. * nss-myhostname will now resolve the single-label host name
  3485. "gateway" to the locally configured default IP routing
  3486. gateways, ordered by their metrics. This assigns a stable
  3487. name to the used gateways, regardless which ones are
  3488. currently configured. Note that the name will only be
  3489. resolved after all other name sources (if nss-myhostname is
  3490. configured properly) and should hence not negatively impact
  3491. systems that use the single-label host name "gateway" in
  3492. other contexts.
  3493. * systemd-inhibit now allows filtering by mode when listing
  3494. inhibitors.
  3495. * Scope and service units gained a new "Delegate" boolean
  3496. property, which, when set, allows processes running inside the
  3497. unit to further partition resources. This is primarily
  3498. useful for systemd user instances as well as container
  3499. managers.
  3500. * journald will now pick up audit messages directly from
  3501. the kernel, and log them like any other log message. The
  3502. audit fields are split up and fully indexed. This means that
  3503. journalctl in many ways is now a (nicer!) alternative to
  3504. ausearch, the traditional audit client. Note that this
  3505. implements only a minimal audit client. If you want the
  3506. special audit modes like reboot-on-log-overflow, please use
  3507. the traditional auditd instead, which can be used in
  3508. parallel to journald.
  3509. * The ConditionSecurity= unit file option now understands the
  3510. special string "audit" to check whether auditing is
  3511. available.
  3512. * journalctl gained two new commands --vacuum-size= and
  3513. --vacuum-time= to delete old journal files until the
  3514. remaining ones take up no more than the specified size on disk,
  3515. or are not older than the specified time.
  3516. * A new, native PPPoE library has been added to sd-network,
  3517. systemd's library of light-weight networking protocols. This
  3518. library will be used in a future version of networkd to
  3519. enable PPPoE communication without an external pppd daemon.
  3520. * The busctl tool now understands a new "capture" verb that
  3521. works similar to "monitor", but writes a packet capture
  3522. trace to STDOUT that can be redirected to a file which is
  3523. compatible with libcap's capture file format. This can then
  3524. be loaded in Wireshark and similar tools to inspect bus
  3525. communication.
  3526. * The busctl tool now understands a new "tree" verb that shows
  3527. the object trees of a specific service on the bus, or of all
  3528. services.
  3529. * The busctl tool now understands a new "introspect" verb that
  3530. shows all interfaces and members of objects on the bus,
  3531. including their signature and values. This is particularly
  3532. useful to get more information about bus objects shown by
  3533. the new "busctl tree" command.
  3534. * The busctl tool now understands new verbs "call",
  3535. "set-property" and "get-property" for invoking bus method
  3536. calls, setting and getting bus object properties in a
  3537. friendly way.
  3538. * busctl gained a new --augment-creds= argument that controls
  3539. whether the tool shall augment credential information it
  3540. gets from the bus with data from /proc, in a possibly
  3541. race-ful way.
  3542. * nspawn's --link-journal= switch gained two new values
  3543. "try-guest" and "try-host" that work like "guest" and
  3544. "host", but do not fail if the host has no persistent
  3545. journalling enabled. -j is now equivalent to
  3546. --link-journal=try-guest.
  3547. * macvlan network devices created by nspawn will now have
  3548. stable MAC addresses.
  3549. * A new SmackProcessLabel= unit setting has been added, which
  3550. controls the SMACK security label processes forked off by
  3551. the respective unit shall use.
  3552. * If compiled with --enable-xkbcommon, systemd-localed will
  3553. verify x11 keymap settings by compiling the given keymap. It
  3554. will spew out warnings if the compilation fails. This
  3555. requires libxkbcommon to be installed.
  3556. * When a coredump is collected, a larger number of metadata
  3557. fields is now collected and included in the journal records
  3558. created for it. More specifically, control group membership,
  3559. environment variables, memory maps, working directory,
  3560. chroot directory, /proc/$PID/status, and a list of open file
  3561. descriptors is now stored in the log entry.
  3562. * The udev hwdb now contains DPI information for mice. For
  3563. details see:
  3565. * All systemd programs that read standalone configuration
  3566. files in /etc now also support a corresponding series of
  3567. .conf.d configuration directories in /etc/, /run/,
  3568. /usr/local/lib/, /usr/lib/, and (if configured with
  3569. --enable-split-usr) /lib/. In particular, the following
  3570. configuration files now have corresponding configuration
  3571. directories: system.conf user.conf, logind.conf,
  3572. journald.conf, sleep.conf, bootchart.conf, coredump.conf,
  3573. resolved.conf, timesyncd.conf, journal-remote.conf, and
  3574. journal-upload.conf. Note that distributions should use the
  3575. configuration directories in /usr/lib/; the directories in
  3576. /etc/ are reserved for the system administrator.
  3577. * systemd-rfkill will no longer take the rfkill device name
  3578. into account when storing rfkill state on disk, as the name
  3579. might be dynamically assigned and not stable. Instead, the
  3580. ID_PATH udev variable combined with the rfkill type (wlan,
  3581. bluetooth, ...) is used.
  3582. * A new service systemd-machine-id-commit.service has been
  3583. added. When used on systems where /etc is read-only during
  3584. boot, and /etc/machine-id is not initialized (but an empty
  3585. file), this service will copy the temporary machine ID
  3586. created as replacement into /etc after the system is fully
  3587. booted up. This is useful for systems that are freshly
  3588. installed with a non-initialized machine ID, but should get
  3589. a fixed machine ID for subsequent boots.
  3590. * networkd's .netdev files now provide a large set of
  3591. configuration parameters for VXLAN devices. Similarly, the
  3592. bridge port cost parameter is now configurable in .network
  3593. files. There's also new support for configuring IP source
  3594. routing. networkd .link files gained support for a new
  3595. OriginalName= match that is useful to match against the
  3596. original interface name the kernel assigned. .network files
  3597. may include MTU= and MACAddress= fields for altering the MTU
  3598. and MAC address while being connected to a specific network
  3599. interface.
  3600. * The LUKS logic gained supported for configuring
  3601. UUID-specific key files. There's also new support for naming
  3602. LUKS device from the kernel command line, using the new
  3603. argument.
  3604. * Timer units may now be transiently created via the bus API
  3605. (this was previously already available for scope and service
  3606. units). In addition it is now possible to create multiple
  3607. transient units at the same time with a single bus call. The
  3608. "systemd-run" tool has been updated to make use of this for
  3609. running commands on a specified time, in at(1)-style.
  3610. * tmpfiles gained support for "t" lines, for assigning
  3611. extended attributes to files. Among other uses this may be
  3612. used to assign SMACK labels to files.
  3613. Contributions from: Alin Rauta, Alison Chaiken, Andrej
  3614. Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
  3615. Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
  3616. Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
  3617. Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
  3618. Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
  3619. Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
  3620. Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
  3621. Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
  3622. Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
  3623. Michael Biebl, Michael Chapman, Michael Marineau, Michal
  3624. Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
  3625. Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
  3626. Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
  3627. Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
  3628. Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
  3629. Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
  3630. Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
  3631. — Berlin, 2014-12-10
  3632. CHANGES WITH 217:
  3633. * journalctl gained the new options -t/--identifier= to match
  3634. on the syslog identifier (aka "tag"), as well as --utc to
  3635. show log timestamps in the UTC timezone. journalctl now also
  3636. accepts -n/--lines=all to disable line capping in a pager.
  3637. * journalctl gained a new switch, --flush, that synchronously
  3638. flushes logs from /run/log/journal to /var/log/journal if
  3639. persistent storage is enabled. systemd-journal-flush.service
  3640. now waits until the operation is complete.
  3641. * Services can notify the manager before they start a reload
  3642. (by sending RELOADING=1) or shutdown (by sending
  3643. STOPPING=1). This allows the manager to track and show the
  3644. internal state of daemons and closes a race condition when
  3645. the process is still running but has closed its D-Bus
  3646. connection.
  3647. * Services with Type=oneshot do not have to have any ExecStart
  3648. commands anymore.
  3649. * User units are now loaded also from
  3650. $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
  3651. /run/systemd/user directory that was already previously
  3652. supported, but is under the control of the user.
  3653. * Job timeouts (i.e. time-outs on the time a job that is
  3654. queued stays in the run queue) can now optionally result in
  3655. immediate reboot or power-off actions (JobTimeoutAction= and
  3656. JobTimeoutRebootArgument=). This is useful on ".target"
  3657. units, to limit the maximum time a target remains
  3658. undispatched in the run queue, and to trigger an emergency
  3659. operation in such a case. This is now used by default to
  3660. turn off the system if boot-up (as defined by everything in
  3661. hangs and does not complete for at least
  3662. 15min. Also, if power-off or reboot hang for at least 30min
  3663. an immediate power-off/reboot operation is triggered. This
  3664. functionality is particularly useful to increase reliability
  3665. on embedded devices, but also on laptops which might
  3666. accidentally get powered on when carried in a backpack and
  3667. whose boot stays stuck in a hard disk encryption passphrase
  3668. question.
  3669. * systemd-logind can be configured to also handle lid switch
  3670. events even when the machine is docked or multiple displays
  3671. are attached (HandleLidSwitchDocked= option).
  3672. * A helper binary and a service have been added which can be
  3673. used to resume from hibernation in the initramfs. A
  3674. generator will parse the resume= option on the kernel
  3675. command line to trigger resume.
  3676. * A user console daemon systemd-consoled has been
  3677. added. Currently, it is a preview, and will so far open a
  3678. single terminal on each session of the user marked as
  3679. Desktop=systemd-console.
  3680. * Route metrics can be specified for DHCP routes added by
  3681. systemd-networkd.
  3682. * The SELinux context of socket-activated services can be set
  3683. from the information provided by the networking stack
  3684. (SELinuxContextFromNet= option).
  3685. * Userspace firmware loading support has been removed and
  3686. the minimum supported kernel version is thus bumped to 3.7.
  3687. * Timeout for udev workers has been increased from 1 to 3
  3688. minutes, but a warning will be printed after 1 minute to
  3689. help diagnose kernel modules that take a long time to load.
  3690. * Udev rules can now remove tags on devices with TAG-="foobar".
  3691. * systemd's readahead implementation has been removed. In many
  3692. circumstances it didn't give expected benefits even for
  3693. rotational disk drives and was becoming less relevant in the
  3694. age of SSDs. As none of the developers has been using
  3695. rotating media anymore, and nobody stepped up to actively
  3696. maintain this component of systemd it has now been removed.
  3697. * Swap units can use Options= to specify discard options.
  3698. Discard options specified for swaps in /etc/fstab are now
  3699. respected.
  3700. * Docker containers are now detected as a separate type of
  3701. virtualization.
  3702. * The Password Agent protocol gained support for queries where
  3703. the user input is shown, useful e.g. for user names.
  3704. systemd-ask-password gained a new --echo option to turn that
  3705. on.
  3706. * The default sysctl.d/ snippets will now set:
  3707. net.core.default_qdisc = fq_codel
  3708. This selects Fair Queuing Controlled Delay as the default
  3709. queuing discipline for network interfaces. fq_codel helps
  3710. fight the network bufferbloat problem. It is believed to be
  3711. a good default with no tuning required for most workloads.
  3712. Downstream distributions may override this choice. On 10Gbit
  3713. servers that do not do forwarding, "fq" may perform better.
  3714. Systems without a good clocksource should use "pfifo_fast".
  3715. * If kdbus is enabled during build a new option BusPolicy= is
  3716. available for service units, that allows locking all service
  3717. processes into a stricter bus policy, in order to limit
  3718. access to various bus services, or even hide most of them
  3719. from the service's view entirely.
  3720. * networkctl will now show the .network and .link file
  3721. networkd has applied to a specific interface.
  3722. * sd-login gained a new API call sd_session_get_desktop() to
  3723. query which desktop environment has been selected for a
  3724. session.
  3725. * UNIX utmp support is now compile-time optional to support
  3726. legacy-free systems.
  3727. * systemctl gained two new commands "add-wants" and
  3728. "add-requires" for pulling in units from specific targets
  3729. easily.
  3730. * If the word "rescue" is specified on the kernel command line
  3731. the system will now boot into rescue mode (aka
  3732., which was previously available only by
  3733. specifying "1" or "" on the kernel
  3734. command line. This new kernel command line option nicely
  3735. mirrors the already existing "emergency" kernel command line
  3736. option.
  3737. * New kernel command line options mount.usr=, mount.usrflags=,
  3738. mount.usrfstype= have been added that match root=, rootflags=,
  3739. rootfstype= but allow mounting a specific file system to
  3740. /usr.
  3741. * The $NOTIFY_SOCKET is now also passed to control processes of
  3742. services, not only the main process.
  3743. * This version reenables support for fsck's -l switch. This
  3744. means at least version v2.25 of util-linux is required for
  3745. operation, otherwise dead-locks on device nodes may
  3746. occur. Again: you need to update util-linux to at least
  3747. v2.25 when updating systemd to v217.
  3748. * The "multi-seat-x" tool has been removed from systemd, as
  3749. its functionality has been integrated into X servers 1.16,
  3750. and the tool is hence redundant. It is recommended to update
  3751. display managers invoking this tool to simply invoke X
  3752. directly from now on, again.
  3753. * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
  3754. message flag has been added for all of systemd's PolicyKit
  3755. authenticated method calls has been added. In particular
  3756. this now allows optional interactive authorization via
  3757. PolicyKit for many of PID1's privileged operations such as
  3758. unit file enabling and disabling.
  3759. * "udevadm hwdb --update" learnt a new switch "--usr" for
  3760. placing the rebuilt hardware database in /usr instead of
  3761. /etc. When used only hardware database entries stored in
  3762. /usr will be used, and any user database entries in /etc are
  3763. ignored. This functionality is useful for vendors to ship a
  3764. pre-built database on systems where local configuration is
  3765. unnecessary or unlikely.
  3766. * Calendar time specifications in .timer units now also
  3767. understand the strings "semi-annually", "quarterly" and
  3768. "minutely" as shortcuts (in addition to the preexisting
  3769. "anually", "hourly", ...).
  3770. * systemd-tmpfiles will now correctly create files in /dev
  3771. at boot which are marked for creation only at boot. It is
  3772. recommended to always create static device nodes with 'c!'
  3773. and 'b!', so that they are created only at boot and not
  3774. overwritten at runtime.
  3775. * When the watchdog logic is used for a service (WatchdogSec=)
  3776. and the watchdog timeout is hit the service will now be
  3777. terminated with SIGABRT (instead of just SIGTERM), in order
  3778. to make sure a proper coredump and backtrace is
  3779. generated. This ensures that hanging services will result in
  3780. similar coredump/backtrace behaviour as services that hit a
  3781. segmentation fault.
  3782. Contributions from: Andreas Henriksson, Andrei Borzenkov,
  3783. Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
  3784. Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
  3785. Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
  3786. Herrmann, David Sommerseth, David Strauss, Emil Renner
  3787. Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
  3788. Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
  3789. Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
  3790. Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
  3791. Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
  3792. Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
  3793. Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
  3794. Michael Marineau, Michael Olbrich, Michael Scherer, Michal
  3795. Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
  3796. Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
  3797. Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
  3798. Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
  3799. Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
  3800. Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
  3801. Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
  3802. Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
  3803. Jędrzejewski-Szmek
  3804. — Berlin, 2014-10-28
  3805. CHANGES WITH 216:
  3806. * timedated no longer reads NTP implementation unit names from
  3807. /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
  3808. implementations should add a
  3809. Conflicts=systemd-timesyncd.service
  3810. to their unit files to take over and replace systemd's NTP
  3811. default functionality.
  3812. * systemd-sysusers gained a new line type "r" for configuring
  3813. which UID/GID ranges to allocate system users/groups
  3814. from. Lines of type "u" may now add an additional column
  3815. that specifies the home directory for the system user to be
  3816. created. Also, systemd-sysusers may now optionally read user
  3817. information from STDIN instead of a file. This is useful for
  3818. invoking it from RPM preinst scriptlets that need to create
  3819. users before the first RPM file is installed since these
  3820. files might need to be owned by them. A new
  3821. %sysusers_create_inline RPM macro has been introduced to do
  3822. just that. systemd-sysusers now updates the shadow files as
  3823. well as the user/group databases, which should enhance
  3824. compatibility with certain tools like grpck.
  3825. * A number of bus APIs of PID 1 now optionally consult
  3826. PolicyKit to permit access for otherwise unprivileged
  3827. clients under certain conditions. Note that this currently
  3828. doesn't support interactive authentication yet, but this is
  3829. expected to be added eventually, too.
  3830. * /etc/machine-info now has new fields for configuring the
  3831. deployment environment of the machine, as well as the
  3832. location of the machine. hostnamectl has been updated with
  3833. new command to update these fields.
  3834. * systemd-timesyncd has been updated to automatically acquire
  3835. NTP server information from systemd-networkd, which might
  3836. have been discovered via DHCP.
  3837. * systemd-resolved now includes a caching DNS stub resolver
  3838. and a complete LLMNR name resolution implementation. A new
  3839. NSS module "nss-resolve" has been added which can be used
  3840. instead of glibc's own "nss-dns" to resolve hostnames via
  3841. systemd-resolved. Hostnames, addresses and arbitrary RRs may
  3842. be resolved via systemd-resolved D-Bus APIs. In contrast to
  3843. the glibc internal resolver systemd-resolved is aware of
  3844. multi-homed system, and keeps DNS server and caches separate
  3845. and per-interface. Queries are sent simultaneously on all
  3846. interfaces that have DNS servers configured, in order to
  3847. properly handle VPNs and local LANs which might resolve
  3848. separate sets of domain names. systemd-resolved may acquire
  3849. DNS server information from systemd-networkd automatically,
  3850. which in turn might have discovered them via DHCP. A tool
  3851. "systemd-resolve-host" has been added that may be used to
  3852. query the DNS logic in resolved. systemd-resolved implements
  3853. IDNA and automatically uses IDNA or UTF-8 encoding depending
  3854. on whether classic DNS or LLMNR is used as transport. In the
  3855. next releases we intend to add a DNSSEC and mDNS/DNS-SD
  3856. implementation to systemd-resolved.
  3857. * A new NSS module nss-mymachines has been added, that
  3858. automatically resolves the names of all local registered
  3859. containers to their respective IP addresses.
  3860. * A new client tool "networkctl" for systemd-networkd has been
  3861. added. It currently is entirely passive and will query
  3862. networking configuration from udev, rtnetlink and networkd,
  3863. and present it to the user in a very friendly
  3864. way. Eventually, we hope to extend it to become a full
  3865. control utility for networkd.
  3866. * .socket units gained a new DeferAcceptSec= setting that
  3867. controls the kernels' TCP_DEFER_ACCEPT sockopt for
  3868. TCP. Similarly, support for controlling TCP keep-alive
  3869. settings has been added (KeepAliveTimeSec=,
  3870. KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
  3871. turning off Nagle's algorithm on TCP has been added
  3872. (NoDelay=).
  3873. * logind learned a new session type "web", for use in projects
  3874. like Cockpit which register web clients as PAM sessions.
  3875. * timer units with at least one OnCalendar= setting will now
  3876. be started only after has been
  3877. reached. This way they will not elapse before the system
  3878. clock has been corrected by a local NTP client or
  3879. similar. This is particular useful on RTC-less embedded
  3880. machines, that come up with an invalid system clock.
  3881. * systemd-nspawn's --network-veth= switch should now result in
  3882. stable MAC addresses for both the outer and the inner side
  3883. of the link.
  3884. * systemd-nspawn gained a new --volatile= switch for running
  3885. container instances with /etc or /var unpopulated.
  3886. * The kdbus client code has been updated to use the new Linux
  3887. 3.17 memfd subsystem instead of the old kdbus-specific one.
  3888. * systemd-networkd's DHCP client and server now support
  3889. FORCERENEW. There are also new configuration options to
  3890. configure the vendor client identifier and broadcast mode
  3891. for DHCP.
  3892. * systemd will no longer inform the kernel about the current
  3893. timezone, as this is necessarily incorrect and racy as the
  3894. kernel has no understanding of DST and similar
  3895. concepts. This hence means FAT timestamps will be always
  3896. considered UTC, similar to what Android is already
  3897. doing. Also, when the RTC is configured to the local time
  3898. (rather than UTC) systemd will never synchronize back to it,
  3899. as this might confuse Windows at a later boot.
  3900. * systemd-analyze gained a new command "verify" for offline
  3901. validation of unit files.
  3902. * systemd-networkd gained support for a couple of additional
  3903. settings for bonding networking setups. Also, the metric for
  3904. statically configured routes may now be configured. For
  3905. network interfaces where this is appropriate the peer IP
  3906. address may now be configured.
  3907. * systemd-networkd's DHCP client will no longer request
  3908. broadcasting by default, as this tripped up some networks.
  3909. For hardware where broadcast is required the feature should
  3910. be switched back on using RequestBroadcast=yes.
  3911. * systemd-networkd will now set up IPv4LL addresses (when
  3912. enabled) even if DHCP is configured successfully.
  3913. * udev will now default to respect network device names given
  3914. by the kernel when the kernel indicates that these are
  3915. predictable. This behavior can be tweaked by changing
  3916. NamePolicy= in the relevant .link file.
  3917. * A new library systemd-terminal has been added that
  3918. implements full TTY stream parsing and rendering. This
  3919. library is supposed to be used later on for implementing a
  3920. full userspace VT subsystem, replacing the current kernel
  3921. implementation.
  3922. * A new tool systemd-journal-upload has been added to push
  3923. journal data to a remote system running
  3924. systemd-journal-remote.
  3925. * journald will no longer forward all local data to another
  3926. running syslog daemon. This change has been made because
  3927. rsyslog (which appears to be the most commonly used syslog
  3928. implementation these days) no longer makes use of this, and
  3929. instead pulls the data out of the journal on its own. Since
  3930. forwarding the messages to a non-existent syslog server is
  3931. more expensive than we assumed we have now turned this
  3932. off. If you run a syslog server that is not a recent rsyslog
  3933. version, you have to turn this option on again
  3934. (ForwardToSyslog= in journald.conf).
  3935. * journald now optionally supports the LZ4 compressor for
  3936. larger journal fields. This compressor should perform much
  3937. better than XZ which was the previous default.
  3938. * machinectl now shows the IP addresses of local containers,
  3939. if it knows them, plus the interface name of the container.
  3940. * A new tool "systemd-escape" has been added that makes it
  3941. easy to escape strings to build unit names and similar.
  3942. * sd_notify() messages may now include a new ERRNO= field
  3943. which is parsed and collected by systemd and shown among the
  3944. "systemctl status" output for a service.
  3945. * A new component "systemd-firstboot" has been added that
  3946. queries the most basic systemd information (timezone,
  3947. hostname, root password) interactively on first
  3948. boot. Alternatively it may also be used to provision these
  3949. things offline on OS images installed into directories.
  3950. * The default sysctl.d/ snippets will now set
  3951. net.ipv4.conf.default.promote_secondaries=1
  3952. This has the benefit of no flushing secondary IP addresses
  3953. when primary addresses are removed.
  3954. Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
  3955. Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
  3956. Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
  3957. Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
  3958. Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
  3959. B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
  3960. Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
  3961. Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
  3962. Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
  3963. Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
  3964. Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
  3965. Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
  3966. Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
  3967. Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
  3968. Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
  3969. — Berlin, 2014-08-19
  3970. CHANGES WITH 215:
  3971. * A new tool systemd-sysusers has been added. This tool
  3972. creates system users and groups in /etc/passwd and
  3973. /etc/group, based on static declarative system user/group
  3974. definitions in /usr/lib/sysusers.d/. This is useful to
  3975. enable factory resets and volatile systems that boot up with
  3976. an empty /etc directory, and thus need system users and
  3977. groups created during early boot. systemd now also ships
  3978. with two default sysusers.d/ files for the most basic
  3979. users and groups systemd and the core operating system
  3980. require.
  3981. * A new tmpfiles snippet has been added that rebuilds the
  3982. essential files in /etc on boot, should they be missing.
  3983. * A directive for ensuring automatic clean-up of
  3984. /var/cache/man/ has been removed from the default
  3985. configuration. This line should now be shipped by the man
  3986. implementation. The necessary change has been made to the
  3987. man-db implementation. Note that you need to update your man
  3988. implementation to one that ships this line, otherwise no
  3989. automatic clean-up of /var/cache/man will take place.
  3990. * A new condition ConditionNeedsUpdate= has been added that
  3991. may conditionalize services to only run when /etc or /var
  3992. are "older" than the vendor operating system resources in
  3993. /usr. This is useful for reconstructing or updating /etc
  3994. after an offline update of /usr or a factory reset, on the
  3995. next reboot. Services that want to run once after such an
  3996. update or reset should use this condition and order
  3997. themselves before the new systemd-update-done.service, which
  3998. will mark the two directories as fully updated. A number of
  3999. service files have been added making use of this, to rebuild
  4000. the udev hardware database, the journald message catalog and
  4001. dynamic loader cache (ldconfig). The systemd-sysusers tool
  4002. described above also makes use of this now. With this in
  4003. place it is now possible to start up a minimal operating
  4004. system with /etc empty cleanly. For more information on the
  4005. concepts involved see this recent blog story:
  4007. * A new system group "input" has been introduced, and all
  4008. input device nodes get this group assigned. This is useful
  4009. for system-level software to get access to input devices. It
  4010. complements what is already done for "audio" and "video".
  4011. * systemd-networkd learnt minimal DHCPv4 server support in
  4012. addition to the existing DHCPv4 client support. It also
  4013. learnt DHCPv6 client and IPv6 Router Solicitation client
  4014. support. The DHCPv4 client gained support for static routes
  4015. passed in from the server. Note that the [DHCPv4] section
  4016. known in older systemd-networkd versions has been renamed to
  4017. [DHCP] and is now also used by the DHCPv6 client. Existing
  4018. .network files using settings of this section should be
  4019. updated, though compatibility is maintained. Optionally, the
  4020. client hostname may now be sent to the DHCP server.
  4021. * networkd gained support for vxlan virtual networks as well
  4022. as tun/tap and dummy devices.
  4023. * networkd gained support for automatic allocation of address
  4024. ranges for interfaces from a system-wide pool of
  4025. addresses. This is useful for dynamically managing a large
  4026. number of interfaces with a single network configuration
  4027. file. In particular this is useful to easily assign
  4028. appropriate IP addresses to the veth links of a large number
  4029. of nspawn instances.
  4030. * RPM macros for processing sysusers, sysctl and binfmt
  4031. drop-in snippets at package installation time have been
  4032. added.
  4033. * The /etc/os-release file should now be placed in
  4034. /usr/lib/os-release. The old location is automatically
  4035. created as symlink. /usr/lib is the more appropriate
  4036. location of this file, since it shall actually describe the
  4037. vendor operating system shipped in /usr, and not the
  4038. configuration stored in /etc.
  4039. * .mount units gained a new boolean SloppyOptions= setting
  4040. that maps to mount(8)'s -s option which enables permissive
  4041. parsing of unknown mount options.
  4042. * tmpfiles learnt a new "L+" directive which creates a symlink
  4043. but (unlike "L") deletes a pre-existing file first, should
  4044. it already exist and not already be the correct
  4045. symlink. Similarly, "b+", "c+" and "p+" directives have been
  4046. added as well, which create block and character devices, as
  4047. well as fifos in the filesystem, possibly removing any
  4048. pre-existing files of different types.
  4049. * For tmpfiles' "L", "L+", "C" and "C+" directives the final
  4050. 'argument' field (which so far specified the source to
  4051. symlink/copy the files from) is now optional. If omitted the
  4052. same file os copied from /usr/share/factory/ suffixed by the
  4053. full destination path. This is useful for populating /etc
  4054. with essential files, by copying them from vendor defaults
  4055. shipped in /usr/share/factory/etc.
  4056. * A new command "systemctl preset-all" has been added that
  4057. applies the service preset settings to all installed unit
  4058. files. A new switch --preset-mode= has been added that
  4059. controls whether only enable or only disable operations
  4060. shall be executed.
  4061. * A new command "systemctl is-system-running" has been added
  4062. that allows checking the overall state of the system, for
  4063. example whether it is fully up and running.
  4064. * When the system boots up with an empty /etc, the equivalent
  4065. to "systemctl preset-all" is executed during early boot, to
  4066. make sure all default services are enabled after a factory
  4067. reset.
  4068. * systemd now contains a minimal preset file that enables the
  4069. most basic services systemd ships by default.
  4070. * Unit files' [Install] section gained a new DefaultInstance=
  4071. field for defining the default instance to create if a
  4072. template unit is enabled with no instance specified.
  4073. * A new passive target has been added
  4074. that may be used by services that need to make they run and
  4075. finish before the first LUKS cryptographic device is set up.
  4076. * The /dev/loop-control and /dev/btrfs-control device nodes
  4077. are now owned by the "disk" group by default, opening up
  4078. access to this group.
  4079. * systemd-coredump will now automatically generate a
  4080. stack trace of all core dumps taking place on the system,
  4081. based on elfutils' libdw library. This stack trace is logged
  4082. to the journal.
  4083. * systemd-coredump may now optionally store coredumps directly
  4084. on disk (in /var/lib/systemd/coredump, possibly compressed),
  4085. instead of storing them unconditionally in the journal. This
  4086. mode is the new default. A new configuration file
  4087. /etc/systemd/coredump.conf has been added to configure this
  4088. and other parameters of systemd-coredump.
  4089. * coredumpctl gained a new "info" verb to show details about a
  4090. specific coredump. A new switch "-1" has also been added
  4091. that makes sure to only show information about the most
  4092. recent entry instead of all entries. Also, as the tool is
  4093. generally useful now the "systemd-" prefix of the binary
  4094. name has been removed. Distributions that want to maintain
  4095. compatibility with the old name should add a symlink from
  4096. the old name to the new name.
  4097. * journald's SplitMode= now defaults to "uid". This makes sure
  4098. that unprivileged users can access their own coredumps with
  4099. coredumpctl without restrictions.
  4100. * New kernel command line options "systemd.wants=" (for
  4101. pulling an additional unit during boot), "systemd.mask="
  4102. (for masking a specific unit for the boot), and
  4103. "systemd.debug-shell" (for enabling the debug shell on tty9)
  4104. have been added. This is implemented in the new generator
  4105. "systemd-debug-generator".
  4106. * systemd-nspawn will now by default filter a couple of
  4107. syscalls for containers, among them those required for
  4108. kernel module loading, direct x86 IO port access, swap
  4109. management, and kexec. Most importantly though
  4110. open_by_handle_at() is now prohibited for containers,
  4111. closing a hole similar to a recently discussed vulnerability
  4112. in docker regarding access to files on file hierarchies the
  4113. container should normally not have access to. Note that, for
  4114. nspawn, we generally make no security claims anyway (and
  4115. this is explicitly documented in the man page), so this is
  4116. just a fix for one of the most obvious problems.
  4117. * A new man page file-hierarchy(7) has been added that
  4118. contains a minimized, modernized version of the file system
  4119. layout systemd expects, similar in style to the FHS
  4120. specification or hier(5). A new tool systemd-path(1) has
  4121. been added to query many of these paths for the local
  4122. machine and user.
  4123. * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
  4124. longer done. Since the directory now has a per-user size
  4125. limit, and is cleaned on logout this appears unnecessary,
  4126. in particular since this now brings the lifecycle of this
  4127. directory closer in line with how IPC objects are handled.
  4128. * systemd.pc now exports a number of additional directories,
  4129. including $libdir (which is useful to identify the library
  4130. path for the primary architecture of the system), and a
  4131. couple of drop-in directories.
  4132. * udev's predictable network interface names now use the dev_port
  4133. sysfs attribute, introduced in linux 3.15 instead of dev_id to
  4134. distinguish between ports of the same PCI function. dev_id should
  4135. only be used for ports using the same HW address, hence the need
  4136. for dev_port.
  4137. * machined has been updated to export the OS version of a
  4138. container (read from /etc/os-release and
  4139. /usr/lib/os-release) on the bus. This is now shown in
  4140. "machinectl status" for a machine.
  4141. * A new service setting RestartForceExitStatus= has been
  4142. added. If configured to a set of exit signals or process
  4143. return values, the service will be restarted when the main
  4144. daemon process exits with any of them, regardless of the
  4145. Restart= setting.
  4146. * systemctl's -H switch for connecting to remote systemd
  4147. machines has been extended so that it may be used to
  4148. directly connect to a specific container on the
  4149. host. "systemctl -H root@foobar:waldi" will now connect as
  4150. user "root" to host "foobar", and then proceed directly to
  4151. the container named "waldi". Note that currently you have to
  4152. authenticate as user "root" for this to work, as entering
  4153. containers is a privileged operation.
  4154. Contributions from: Andreas Henriksson, Benjamin Steinwender,
  4155. Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
  4156. Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
  4157. Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
  4158. Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
  4159. Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
  4160. Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
  4161. Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
  4162. Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
  4163. Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
  4164. Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
  4165. Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
  4166. — Berlin, 2014-07-03
  4167. CHANGES WITH 214:
  4168. * As an experimental feature, udev now tries to lock the
  4169. disk device node (flock(LOCK_SH|LOCK_NB)) while it
  4170. executes events for the disk or any of its partitions.
  4171. Applications like partitioning programs can lock the
  4172. disk device node (flock(LOCK_EX)) and claim temporary
  4173. device ownership that way; udev will entirely skip all event
  4174. handling for this disk and its partitions. If the disk
  4175. was opened for writing, the close will trigger a partition
  4176. table rescan in udev's "watch" facility, and if needed
  4177. synthesize "change" events for the disk and all its partitions.
  4178. This is now unconditionally enabled, and if it turns out to
  4179. cause major problems, we might turn it on only for specific
  4180. devices, or might need to disable it entirely. Device Mapper
  4181. devices are excluded from this logic.
  4182. * We temporarily dropped the "-l" switch for fsck invocations,
  4183. since they collide with the flock() logic above. util-linux
  4184. upstream has been changed already to avoid this conflict,
  4185. and we will readd "-l" as soon as util-linux with this
  4186. change has been released.
  4187. * The dependency on libattr has been removed. Since a long
  4188. time, the extended attribute calls have moved to glibc, and
  4189. libattr is thus unnecessary.
  4190. * Virtualization detection works without privileges now. This
  4191. means the systemd-detect-virt binary no longer requires
  4192. CAP_SYS_PTRACE file capabilities, and our daemons can run
  4193. with fewer privileges.
  4194. * systemd-networkd now runs under its own "systemd-network"
  4195. user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
  4196. CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
  4197. loses the ability to write to files owned by root this way.
  4198. * Similarly, systemd-resolved now runs under its own
  4199. "systemd-resolve" user with no capabilities remaining.
  4200. * Similarly, systemd-bus-proxyd now runs under its own
  4201. "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
  4202. * systemd-networkd gained support for setting up "veth"
  4203. virtual Ethernet devices for container connectivity, as well
  4204. as GRE and VTI tunnels.
  4205. * systemd-networkd will no longer automatically attempt to
  4206. manually load kernel modules necessary for certain tunnel
  4207. transports. Instead, it is assumed the kernel loads them
  4208. automatically when required. This only works correctly on
  4209. very new kernels. On older kernels, please consider adding
  4210. the kernel modules to /etc/modules-load.d/ as a work-around.
  4211. * The resolv.conf file systemd-resolved generates has been
  4212. moved to /run/systemd/resolve/. If you have a symlink from
  4213. /etc/resolv.conf, it might be necessary to correct it.
  4214. * Two new service settings, ProtectHome= and ProtectSystem=,
  4215. have been added. When enabled, they will make the user data
  4216. (such as /home) inaccessible or read-only and the system
  4217. (such as /usr) read-only, for specific services. This allows
  4218. very light-weight per-service sandboxing to avoid
  4219. modifications of user data or system files from
  4220. services. These two new switches have been enabled for all
  4221. of systemd's long-running services, where appropriate.
  4222. * Socket units gained new SocketUser= and SocketGroup=
  4223. settings to set the owner user and group of AF_UNIX sockets
  4224. and FIFOs in the file system.
  4225. * Socket units gained a new RemoveOnStop= setting. If enabled,
  4226. all FIFOS and sockets in the file system will be removed
  4227. when the specific socket unit is stopped.
  4228. * Socket units gained a new Symlinks= setting. It takes a list
  4229. of symlinks to create to file system sockets or FIFOs
  4230. created by the specific Unix sockets. This is useful to
  4231. manage symlinks to socket nodes with the same life-cycle as
  4232. the socket itself.
  4233. * The /dev/log socket and /dev/initctl FIFO have been moved to
  4234. /run, and have been replaced by symlinks. This allows
  4235. connecting to these facilities even if PrivateDevices=yes is
  4236. used for a service (which makes /dev/log itself unavailable,
  4237. but /run is left). This also has the benefit of ensuring
  4238. that /dev only contains device nodes, directories and
  4239. symlinks, and nothing else.
  4240. * sd-daemon gained two new calls sd_pid_notify() and
  4241. sd_pid_notifyf(). They are similar to sd_notify() and
  4242. sd_notifyf(), but allow overriding of the source PID of
  4243. notification messages if permissions permit this. This is
  4244. useful to send notify messages on behalf of a different
  4245. process (for example, the parent process). The
  4246. systemd-notify tool has been updated to make use of this
  4247. when sending messages (so that notification messages now
  4248. originate from the shell script invoking systemd-notify and
  4249. not the systemd-notify process itself. This should minimize
  4250. a race where systemd fails to associate notification
  4251. messages to services when the originating process already
  4252. vanished.
  4253. * A new "on-abnormal" setting for Restart= has been added. If
  4254. set, it will result in automatic restarts on all "abnormal"
  4255. reasons for a process to exit, which includes unclean
  4256. signals, core dumps, timeouts and watchdog timeouts, but
  4257. does not include clean and unclean exit codes or clean
  4258. signals. Restart=on-abnormal is an alternative for
  4259. Restart=on-failure for services that shall be able to
  4260. terminate and avoid restarts on certain errors, by
  4261. indicating so with an unclean exit code. Restart=on-failure
  4262. or Restart=on-abnormal is now the recommended setting for
  4263. all long-running services.
  4264. * If the InaccessibleDirectories= service setting points to a
  4265. mount point (or if there are any submounts contained within
  4266. it), it is now attempted to completely unmount it, to make
  4267. the file systems truly unavailable for the respective
  4268. service.
  4269. * The ReadOnlyDirectories= service setting and
  4270. systemd-nspawn's --read-only parameter are now recursively
  4271. applied to all submounts, too.
  4272. * Mount units may now be created transiently via the bus APIs.
  4273. * The support for SysV and LSB init scripts has been removed
  4274. from the systemd daemon itself. Instead, it is now
  4275. implemented as a generator that creates native systemd units
  4276. from these scripts when needed. This enables us to remove a
  4277. substantial amount of legacy code from PID 1, following the
  4278. fact that many distributions only ship a very small number
  4279. of LSB/SysV init scripts nowadays.
  4280. * Privileged Xen (dom0) domains are not considered
  4281. virtualization anymore by the virtualization detection
  4282. logic. After all, they generally have unrestricted access to
  4283. the hardware and usually are used to manage the unprivileged
  4284. (domU) domains.
  4285. * systemd-tmpfiles gained a new "C" line type, for copying
  4286. files or entire directories.
  4287. * systemd-tmpfiles "m" lines are now fully equivalent to "z"
  4288. lines. So far, they have been non-globbing versions of the
  4289. latter, and have thus been redundant. In future, it is
  4290. recommended to only use "z". "m" has hence been removed
  4291. from the documentation, even though it stays supported.
  4292. * A tmpfiles snippet to recreate the most basic structure in
  4293. /var has been added. This is enough to create the /var/run →
  4294. /run symlink and create a couple of structural
  4295. directories. This allows systems to boot up with an empty or
  4296. volatile /var. Of course, while with this change, the core OS
  4297. now is capable with dealing with a volatile /var, not all
  4298. user services are ready for it. However, we hope that sooner
  4299. or later, many service daemons will be changed upstream so
  4300. that they are able to automatically create their necessary
  4301. directories in /var at boot, should they be missing. This is
  4302. the first step to allow state-less systems that only require
  4303. the vendor image for /usr to boot.
  4304. * systemd-nspawn has gained a new --tmpfs= switch to mount an
  4305. empty tmpfs instance to a specific directory. This is
  4306. particularly useful for making use of the automatic
  4307. reconstruction of /var (see above), by passing --tmpfs=/var.
  4308. * Access modes specified in tmpfiles snippets may now be
  4309. prefixed with "~", which indicates that they shall be masked
  4310. by whether the existing file or directory is currently
  4311. writable, readable or executable at all. Also, if specified,
  4312. the sgid/suid/sticky bits will be masked for all
  4313. non-directories.
  4314. * A new passive target unit "" has been
  4315. added which is useful for services that shall run before any
  4316. network is configured, for example firewall scripts.
  4317. * The "floppy" group that previously owned the /dev/fd*
  4318. devices is no longer used. The "disk" group is now used
  4319. instead. Distributions should probably deprecate usage of
  4320. this group.
  4321. Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
  4322. King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
  4323. Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
  4324. Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
  4325. Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
  4326. Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
  4327. Jędrzejewski-Szmek
  4328. — Berlin, 2014-06-11
  4329. CHANGES WITH 213:
  4330. * A new "systemd-timesyncd" daemon has been added for
  4331. synchronizing the system clock across the network. It
  4332. implements an SNTP client. In contrast to NTP
  4333. implementations such as chrony or the NTP reference server,
  4334. this only implements a client side, and does not bother with
  4335. the full NTP complexity, focusing only on querying time from
  4336. one remote server and synchronizing the local clock to
  4337. it. Unless you intend to serve NTP to networked clients or
  4338. want to connect to local hardware clocks, this simple NTP
  4339. client should be more than appropriate for most
  4340. installations. The daemon runs with minimal privileges, and
  4341. has been hooked up with networkd to only operate when
  4342. network connectivity is available. The daemon saves the
  4343. current clock to disk every time a new NTP sync has been
  4344. acquired, and uses this to possibly correct the system clock
  4345. early at bootup, in order to accommodate for systems that
  4346. lack an RTC such as the Raspberry Pi and embedded devices,
  4347. and to make sure that time monotonically progresses on these
  4348. systems, even if it is not always correct. To make use of
  4349. this daemon, a new system user and group "systemd-timesync"
  4350. needs to be created on installation of systemd.
  4351. * The queue "seqnum" interface of libudev has been disabled, as
  4352. it was generally incompatible with device namespacing as
  4353. sequence numbers of devices go "missing" if the devices are
  4354. part of a different namespace.
  4355. * "systemctl list-timers" and "systemctl list-sockets" gained
  4356. a --recursive switch for showing units of these types also
  4357. for all local containers, similar in style to the already
  4358. supported --recursive switch for "systemctl list-units".
  4359. * A new RebootArgument= setting has been added for service
  4360. units, which may be used to specify a kernel reboot argument
  4361. to use when triggering reboots with StartLimitAction=.
  4362. * A new FailureAction= setting has been added for service
  4363. units which may be used to specify an operation to trigger
  4364. when a service fails. This works similarly to
  4365. StartLimitAction=, but unlike it, controls what is done
  4366. immediately rather than only after several attempts to
  4367. restart the service in question.
  4368. * hostnamed got updated to also expose the kernel name,
  4369. release, and version on the bus. This is useful for
  4370. executing commands like hostnamectl with the -H switch.
  4371. systemd-analyze makes use of this to properly display
  4372. details when running non-locally.
  4373. * The bootchart tool can now show cgroup information in the
  4374. graphs it generates.
  4375. * The CFS CPU quota cgroup attribute is now exposed for
  4376. services. The new CPUQuota= switch has been added for this
  4377. which takes a percentage value. Setting this will have the
  4378. result that a service may never get more CPU time than the
  4379. specified percentage, even if the machine is otherwise idle.
  4380. * systemd-networkd learned IPIP and SIT tunnel support.
  4381. * LSB init scripts exposing a dependency on $network will now
  4382. get a dependency on rather than simply
  4383. This should bring LSB handling closer to
  4384. what it was on SysV systems.
  4385. * A new kernel option has been added to control
  4386. how fsck shall deal with unclean file systems at boot.
  4387. * The (.ini) configuration file parser will now silently
  4388. ignore sections whose name begins with "X-". This may be
  4389. used to maintain application-specific extension sections in unit
  4390. files.
  4391. * machined gained a new API to query the IP addresses of
  4392. registered containers. "machinectl status" has been updated
  4393. to show these addresses in its output.
  4394. * A new call sd_uid_get_display() has been added to the
  4395. sd-login APIs for querying the "primary" session of a
  4396. user. The "primary" session of the user is elected from the
  4397. user's sessions and generally a graphical session is
  4398. preferred over a text one.
  4399. * A minimal systemd-resolved daemon has been added. It
  4400. currently simply acts as a companion to systemd-networkd and
  4401. manages resolv.conf based on per-interface DNS
  4402. configuration, possibly supplied via DHCP. In the long run
  4403. we hope to extend this into a local DNSSEC enabled DNS and
  4404. mDNS cache.
  4405. * The systemd-networkd-wait-online tool is now enabled by
  4406. default. It will delay until a network
  4407. connection has been configured. The tool primarily integrates
  4408. with networkd, but will also make a best effort to make sense
  4409. of network configuration performed in some other way.
  4410. * Two new service options StartupCPUShares= and
  4411. StartupBlockIOWeight= have been added that work similarly to
  4412. CPUShares= and BlockIOWeight= however only apply during
  4413. system startup. This is useful to prioritize certain services
  4414. differently during bootup than during normal runtime.
  4415. * hostnamed has been changed to prefer the statically
  4416. configured hostname in /etc/hostname (unless set to
  4417. 'localhost' or empty) over any dynamic one supplied by
  4418. dhcp. With this change, the rules for picking the hostname
  4419. match more closely the rules of other configuration settings
  4420. where the local administrator's configuration in /etc always
  4421. overrides any other settings.
  4422. Contributions fron: Ali H. Caliskan, Alison Chaiken, Bas van
  4423. den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
  4424. Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
  4425. David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
  4426. Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
  4427. Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
  4428. Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
  4429. Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
  4430. Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
  4431. Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
  4432. Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
  4433. Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
  4434. Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
  4435. Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
  4436. Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
  4437. Lindskog, WaLyong Cho, Will Woods, Zbigniew
  4438. Jędrzejewski-Szmek
  4439. — Beijing, 2014-05-28
  4440. CHANGES WITH 212:
  4441. * When restoring the screen brightness at boot, stay away from
  4442. the darkest setting or from the lowest 5% of the available
  4443. range, depending on which is the larger value of both. This
  4444. should effectively protect the user from rebooting into a
  4445. black screen, should the brightness have been set to minimum
  4446. by accident.
  4447. * sd-login gained a new sd_machine_get_class() call to
  4448. determine the class ("vm" or "container") of a machine
  4449. registered with machined.
  4450. * sd-login gained new calls
  4451. sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
  4452. to query the identity of the peer of a local AF_UNIX
  4453. connection. They operate similarly to their sd_pid_get_xyz()
  4454. counterparts.
  4455. * PID 1 will now maintain a system-wide system state engine
  4456. with the states "starting", "running", "degraded",
  4457. "maintenance", "stopping". These states are bound to system
  4458. startup, normal runtime, runtime with at least one failed
  4459. service, rescue/emergency mode and system shutdown. This
  4460. state is shown in the "systemctl status" output when no unit
  4461. name is passed. It is useful to determine system state, in
  4462. particularly when doing so for many systems or containers at
  4463. once.
  4464. * A new command "list-machines" has been added to "systemctl"
  4465. that lists all local OS containers and shows their system
  4466. state (see above), if systemd runs inside of them.
  4467. * systemctl gained a new "-r" switch to recursively enumerate
  4468. units on all local containers, when used with the
  4469. "list-unit" command (which is the default one that is
  4470. executed when no parameters are specified).
  4471. * The GPT automatic partition discovery logic will now honour
  4472. two GPT partition flags: one may be set on a partition to
  4473. cause it to be mounted read-only, and the other may be set
  4474. on a partition to ignore it during automatic discovery.
  4475. * Two new GPT type UUIDs have been added for automatic root
  4476. partition discovery, for 32-bit and 64-bit ARM. This is not
  4477. particularly useful for discovering the root directory on
  4478. these architectures during bare-metal boots (since UEFI is
  4479. not common there), but still very useful to allow booting of
  4480. ARM disk images in nspawn with the -i option.
  4481. * MAC addresses of interfaces created with nspawn's
  4482. --network-interface= switch will now be generated from the
  4483. machine name, and thus be stable between multiple invocations
  4484. of the container.
  4485. * logind will now automatically remove all IPC objects owned
  4486. by a user if she or he fully logs out. This makes sure that
  4487. users who are logged out cannot continue to consume IPC
  4488. resources. This covers SysV memory, semaphores and message
  4489. queues as well as POSIX shared memory and message
  4490. queues. Traditionally, SysV and POSIX IPC had no life-cycle
  4491. limits. With this functionality, that is corrected. This may
  4492. be turned off by using the RemoveIPC= switch of logind.conf.
  4493. * The systemd-machine-id-setup and tmpfiles tools gained a
  4494. --root= switch to operate on a specific root directory,
  4495. instead of /.
  4496. * journald can now forward logged messages to the TTYs of all
  4497. logged in users ("wall"). This is the default for all
  4498. emergency messages now.
  4499. * A new tool systemd-journal-remote has been added to stream
  4500. journal log messages across the network.
  4501. * /sys/fs/cgroup/ is now mounted read-only after all cgroup
  4502. controller trees are mounted into it. Note that the
  4503. directories mounted beneath it are not read-only. This is a
  4504. security measure and is particularly useful because glibc
  4505. actually includes a search logic to pick any tmpfs it can
  4506. find to implement shm_open() if /dev/shm is not available
  4507. (which it might very well be in namespaced setups).
  4508. * machinectl gained a new "poweroff" command to cleanly power
  4509. down a local OS container.
  4510. * The PrivateDevices= unit file setting will now also drop the
  4511. CAP_MKNOD capability from the capability bound set, and
  4512. imply DevicePolicy=closed.
  4513. * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
  4514. comprehensively on all long-running systemd services where
  4515. this is appropriate.
  4516. * systemd-udevd will now run in a disassociated mount
  4517. namespace. To mount directories from udev rules, make sure to
  4518. pull in mount units via SYSTEMD_WANTS properties.
  4519. * The kdbus support gained support for uploading policy into
  4520. the kernel. sd-bus gained support for creating "monitoring"
  4521. connections that can eavesdrop into all bus communication
  4522. for debugging purposes.
  4523. * Timestamps may now be specified in seconds since the UNIX
  4524. epoch Jan 1st, 1970 by specifying "@" followed by the value
  4525. in seconds.
  4526. * Native tcpwrap support in systemd has been removed. tcpwrap
  4527. is old code, not really maintained anymore and has serious
  4528. shortcomings, and better options such as firewalls
  4529. exist. For setups that require tcpwrap usage, please
  4530. consider invoking your socket-activated service via tcpd,
  4531. like on traditional inetd.
  4532. * A new system.conf configuration option
  4533. DefaultTimerAccuracySec= has been added that controls the
  4534. default AccuracySec= setting of .timer units.
  4535. * Timer units gained a new WakeSystem= switch. If enabled,
  4536. timers configured this way will cause the system to resume
  4537. from system suspend (if the system supports that, which most
  4538. do these days).
  4539. * Timer units gained a new Persistent= switch. If enabled,
  4540. timers configured this way will save to disk when they have
  4541. been last triggered. This information is then used on next
  4542. reboot to possible execute overdue timer events, that
  4543. could not take place because the system was powered off.
  4544. This enables simple anacron-like behaviour for timer units.
  4545. * systemctl's "list-timers" will now also list the time a
  4546. timer unit was last triggered in addition to the next time
  4547. it will be triggered.
  4548. * systemd-networkd will now assign predictable IPv4LL
  4549. addresses to its local interfaces.
  4550. Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
  4551. Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
  4552. Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
  4553. Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
  4554. Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
  4555. Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
  4556. Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
  4557. Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
  4558. Jędrzejewski-Szmek
  4559. — Berlin, 2014-03-25
  4560. CHANGES WITH 211:
  4561. * A new unit file setting RestrictAddressFamilies= has been
  4562. added to restrict which socket address families unit
  4563. processes gain access to. This takes address family names
  4564. like "AF_INET" or "AF_UNIX", and is useful to minimize the
  4565. attack surface of services via exotic protocol stacks. This
  4566. is built on seccomp system call filters.
  4567. * Two new unit file settings RuntimeDirectory= and
  4568. RuntimeDirectoryMode= have been added that may be used to
  4569. manage a per-daemon runtime directories below /run. This is
  4570. an alternative for setting up directory permissions with
  4571. tmpfiles snippets, and has the advantage that the runtime
  4572. directory's lifetime is bound to the daemon runtime and that
  4573. the daemon starts up with an empty directory each time. This
  4574. is particularly useful when writing services that drop
  4575. privileges using the User= or Group= setting.
  4576. * The DeviceAllow= unit setting now supports globbing for
  4577. matching against device group names.
  4578. * The systemd configuration file system.conf gained new
  4579. settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
  4580. DefaultMemoryAccounting= to globally turn on/off accounting
  4581. for specific resources (cgroups) for all units. These
  4582. settings may still be overridden individually in each unit
  4583. though.
  4584. * systemd-gpt-auto-generator is now able to discover /srv and
  4585. root partitions in addition to /home and swap partitions. It
  4586. also supports LUKS-encrypted partitions now. With this in
  4587. place, automatic discovery of partitions to mount following
  4588. the Discoverable Partitions Specification
  4589. (
  4590. is now a lot more complete. This allows booting without
  4591. /etc/fstab and without root= on the kernel command line on
  4592. systems prepared appropriately.
  4593. * systemd-nspawn gained a new --image= switch which allows
  4594. booting up disk images and Linux installations on any block
  4595. device that follow the Discoverable Partitions Specification
  4596. (see above). This means that installations made with
  4597. appropriately updated installers may now be started and
  4598. deployed using container managers, completely
  4599. unmodified. (We hope that libvirt-lxc will add support for
  4600. this feature soon, too.)
  4601. * systemd-nspawn gained a new --network-macvlan= setting to
  4602. set up a private macvlan interface for the
  4603. container. Similarly, systemd-networkd gained a new
  4604. Kind=macvlan setting in .netdev files.
  4605. * systemd-networkd now supports configuring local addresses
  4606. using IPv4LL.
  4607. * A new tool systemd-network-wait-online has been added to
  4608. synchronously wait for network connectivity using
  4609. systemd-networkd.
  4610. * The sd-bus.h bus API gained a new sd_bus_track object for
  4611. tracking the life-cycle of bus peers. Note that sd-bus.h is
  4612. still not a public API though (unless you specify
  4613. --enable-kdbus on the configure command line, which however
  4614. voids your warranty and you get no API stability guarantee).
  4615. * The $XDG_RUNTIME_DIR runtime directories for each user are
  4616. now individual tmpfs instances, which has the benefit of
  4617. introducing separate pools for each user, with individual
  4618. size limits, and thus making sure that unprivileged clients
  4619. can no longer negatively impact the system or other users by
  4620. filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
  4621. RuntimeDirectorySize= has been introduced that allows
  4622. controlling the default size limit for all users. It
  4623. defaults to 10% of the available physical memory. This is no
  4624. replacement for quotas on tmpfs though (which the kernel
  4625. still does not support), as /dev/shm and /tmp are still
  4626. shared resources used by both the system and unprivileged
  4627. users.
  4628. * logind will now automatically turn off automatic suspending
  4629. on laptop lid close when more than one display is
  4630. connected. This was previously expected to be implemented
  4631. individually in desktop environments (such as GNOME),
  4632. however has been added to logind now, in order to fix a
  4633. boot-time race where a desktop environment might not have
  4634. been started yet and thus not been able to take an inhibitor
  4635. lock at the time where logind already suspends the system
  4636. due to a closed lid.
  4637. * logind will now wait at least 30s after each system
  4638. suspend/resume cycle, and 3min after system boot before
  4639. suspending the system due to a closed laptop lid. This
  4640. should give USB docking stations and similar enough time to
  4641. be probed and configured after system resume and boot in
  4642. order to then act as suspend blocker.
  4643. * systemd-run gained a new --property= setting which allows
  4644. initialization of resource control properties (and others)
  4645. for the created scope or service unit. Example: "systemd-run
  4646. --property=BlockIOWeight=10 updatedb" may be used to run
  4647. updatedb at a low block IO scheduling weight.
  4648. * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
  4649. now also work in --scope mode.
  4650. * When systemd is compiled with kdbus support, basic support
  4651. for enforced policies is now in place. (Note that enabling
  4652. kdbus still voids your warranty and no API compatibility
  4653. promises are made.)
  4654. Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
  4655. K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
  4656. Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
  4657. Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
  4658. Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
  4659. Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
  4660. Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
  4661. Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
  4662. Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
  4663. Zbigniew Jędrzejewski-Szmek
  4664. — Berlin, 2014-03-12
  4665. CHANGES WITH 210:
  4666. * systemd will now relabel /dev after loading the SMACK policy
  4667. according to SMACK rules.
  4668. * A new unit file option AppArmorProfile= has been added to
  4669. set the AppArmor profile for the processes of a unit.
  4670. * A new condition check ConditionArchitecture= has been added
  4671. to conditionalize units based on the system architecture, as
  4672. reported by uname()'s "machine" field.
  4673. * systemd-networkd now supports matching on the system
  4674. virtualization, architecture, kernel command line, host name
  4675. and machine ID.
  4676. * logind is now a lot more aggressive when suspending the
  4677. machine due to a closed laptop lid. Instead of acting only
  4678. on the lid close action, it will continuously watch the lid
  4679. status and act on it. This is useful for laptops where the
  4680. power button is on the outside of the chassis so that it can
  4681. be reached without opening the lid (such as the Lenovo
  4682. Yoga). On those machines, logind will now immediately
  4683. re-suspend the machine if the power button has been
  4684. accidentally pressed while the laptop was suspended and in a
  4685. backpack or similar.
  4686. * logind will now watch SW_DOCK switches and inhibit reaction
  4687. to the lid switch if it is pressed. This means that logind
  4688. will not suspend the machine anymore if the lid is closed
  4689. and the system is docked, if the laptop supports SW_DOCK
  4690. notifications via the input layer. Note that ACPI docking
  4691. stations do not generate this currently. Also note that this
  4692. logic is usually not fully sufficient and Desktop
  4693. Environments should take a lid switch inhibitor lock when an
  4694. external display is connected, as systemd will not watch
  4695. this on its own.
  4696. * nspawn will now make use of the devices cgroup controller by
  4697. default, and only permit creation of and access to the usual
  4698. API device nodes like /dev/null or /dev/random, as well as
  4699. access to (but not creation of) the pty devices.
  4700. * We will now ship a default .network file for
  4701. systemd-networkd that automatically configures DHCP for
  4702. network interfaces created by nspawn's --network-veth or
  4703. --network-bridge= switches.
  4704. * systemd will now understand the usual M, K, G, T suffixes
  4705. according to SI conventions (i.e. to the base 1000) when
  4706. referring to throughput and hardware metrics. It will stay
  4707. with IEC conventions (i.e. to the base 1024) for software
  4708. metrics, according to what is customary according to
  4709. Wikipedia. We explicitly document which base applies for
  4710. each configuration option.
  4711. * The DeviceAllow= setting in unit files now supports a syntax
  4712. to whitelist an entire group of devices node majors at once,
  4713. based on the /proc/devices listing. For example, with the
  4714. string "char-pts", it is now possible to whitelist all
  4715. current and future pseudo-TTYs at once.
  4716. * sd-event learned a new "post" event source. Event sources of
  4717. this type are triggered by the dispatching of any event
  4718. source of a type that is not "post". This is useful for
  4719. implementing clean-up and check event sources that are
  4720. triggered by other work being done in the program.
  4721. * systemd-networkd is no longer statically enabled, but uses
  4722. the usual [Install] sections so that it can be
  4723. enabled/disabled using systemctl. It still is enabled by
  4724. default however.
  4725. * When creating a veth interface pair with systemd-nspawn, the
  4726. host side will now be prefixed with "vb-" if
  4727. --network-bridge= is used, and with "ve-" if --network-veth
  4728. is used. This way, it is easy to distinguish these cases on
  4729. the host, for example to apply different configuration to
  4730. them with systemd-networkd.
  4731. * The compatibility libraries for,
  4732., and
  4733. do not make use of IFUNC
  4734. anymore. Instead, we now build multiple times
  4735. under these alternative names. This means that the footprint
  4736. is drastically increased, but given that these are
  4737. transitional compatibility libraries, this should not matter
  4738. much. This change has been made necessary to support the ARM
  4739. platform for these compatibility libraries, as the ARM
  4740. toolchain is not really at the same level as the toolchain
  4741. for other architectures like x86 and does not support
  4742. IFUNC. Please make sure to use --enable-compat-libs only
  4743. during a transitional period!
  4744. Contributions from: Andreas Fuchs, Armin K., Colin Walters,
  4745. Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
  4746. Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
  4747. St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
  4748. Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
  4749. Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
  4750. Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
  4751. Zbigniew Jędrzejewski-Szmek
  4752. — Berlin, 2014-02-24
  4753. CHANGES WITH 209:
  4754. * A new component "systemd-networkd" has been added that can
  4755. be used to configure local network interfaces statically or
  4756. via DHCP. It is capable of bringing up bridges, VLANs, and
  4757. bonding. Currently, no hook-ups for interactive network
  4758. configuration are provided. Use this for your initrd,
  4759. container, embedded, or server setup if you need a simple,
  4760. yet powerful, network configuration solution. This
  4761. configuration subsystem is quite nifty, as it allows wildcard
  4762. hotplug matching in interfaces. For example, with a single
  4763. configuration snippet, you can configure that all Ethernet
  4764. interfaces showing up are automatically added to a bridge,
  4765. or similar. It supports link-sensing and more.
  4766. * A new tool "systemd-socket-proxyd" has been added which can
  4767. act as a bidirectional proxy for TCP sockets. This is
  4768. useful for adding socket activation support to services that
  4769. do not actually support socket activation, including virtual
  4770. machines and the like.
  4771. * Add a new tool to save/restore rfkill state on
  4772. shutdown/boot.
  4773. * Save/restore state of keyboard backlights in addition to
  4774. display backlights on shutdown/boot.
  4775. * udev learned a new SECLABEL{} construct to label device
  4776. nodes with a specific security label when they appear. For
  4777. now, only SECLABEL{selinux} is supported, but the syntax is
  4778. prepared for additional security frameworks.
  4779. * udev gained a new scheme to configure link-level attributes
  4780. from files in /etc/systemd/network/*.link. These files can
  4781. match against MAC address, device path, driver name and type,
  4782. and will apply attributes like the naming policy, link speed,
  4783. MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
  4784. address assignment policy (randomized, ...).
  4785. * The configuration of network interface naming rules for
  4786. "permanent interface names" has changed: a new NamePolicy=
  4787. setting in the [Link] section of .link files determines the
  4788. priority of possible naming schemes (onboard, slot, MAC,
  4789. path). The default value of this setting is determined by
  4790. /usr/lib/net/links/ Old
  4791. 80-net-name-slot.rules udev configuration file has been
  4792. removed, so local configuration overriding this file should
  4793. be adapted to override instead.
  4794. * When the User= switch is used in a unit file, also
  4795. initialize $SHELL= based on the user database entry.
  4796. * systemd no longer depends on libdbus. All communication is
  4797. now done with sd-bus, systemd's low-level bus library
  4798. implementation.
  4799. * kdbus support has been added to PID 1 itself. When kdbus is
  4800. enabled, this causes PID 1 to set up the system bus and
  4801. enable support for a new ".busname" unit type that
  4802. encapsulates bus name activation on kdbus. It works a little
  4803. bit like ".socket" units, except for bus names. A new
  4804. generator has been added that converts classic dbus1 service
  4805. activation files automatically into native systemd .busname
  4806. and .service units.
  4807. * sd-bus: add a light-weight vtable implementation that allows
  4808. defining objects on the bus with a simple static const
  4809. vtable array of its methods, signals and properties.
  4810. * systemd will not generate or install static dbus
  4811. introspection data anymore to /usr/share/dbus-1/interfaces,
  4812. as the precise format of these files is unclear, and
  4813. nothing makes use of it.
  4814. * A proxy daemon is now provided to proxy clients connecting
  4815. via classic D-Bus AF_UNIX sockets to kdbus, to provide full
  4816. compatibility with classic D-Bus.
  4817. * A bus driver implementation has been added that supports the
  4818. classic D-Bus bus driver calls on kdbus, also for
  4819. compatibility purposes.
  4820. * A new API "sd-event.h" has been added that implements a
  4821. minimal event loop API built around epoll. It provides a
  4822. couple of features that direct epoll usage is lacking:
  4823. prioritization of events, scales to large numbers of timer
  4824. events, per-event timer slack (accuracy), system-wide
  4825. coalescing of timer events, exit handlers, watchdog
  4826. supervision support using systemd's sd_notify() API, child
  4827. process handling.
  4828. * A new API "sd-rntl.h" has been added that provides an API
  4829. around the route netlink interface of the kernel, similar in
  4830. style to "sd-bus.h".
  4831. * A new API "sd-dhcp-client.h" has been added that provides a
  4832. small DHCPv4 client-side implementation. This is used by
  4833. "systemd-networkd".
  4834. * There is a new kernel command line option
  4835. "systemd.restore_state=0|1". When set to "0", none of the
  4836. systemd tools will restore saved runtime state to hardware
  4837. devices. More specifically, the rfkill and backlight states
  4838. are not restored.
  4839. * The FsckPassNo= compatibility option in mount/service units
  4840. has been removed. The fstab generator will now add the
  4841. necessary dependencies automatically, and does not require
  4842. PID1's support for that anymore.
  4843. * journalctl gained a new switch, --list-boots, that lists
  4844. recent boots with their times and boot IDs.
  4845. * The various tools like systemctl, loginctl, timedatectl,
  4846. busctl, systemd-run, ... have gained a new switch "-M" to
  4847. connect to a specific, local OS container (as direct
  4848. connection, without requiring SSH). This works on any
  4849. container that is registered with machined, such as those
  4850. created by libvirt-lxc or nspawn.
  4851. * systemd-run and systemd-analyze also gained support for "-H"
  4852. to connect to remote hosts via SSH. This is particularly
  4853. useful for systemd-run because it enables queuing of jobs
  4854. onto remote systems.
  4855. * machinectl gained a new command "login" to open a getty
  4856. login in any local container. This works with any container
  4857. that is registered with machined (such as those created by
  4858. libvirt-lxc or nspawn), and which runs systemd inside.
  4859. * machinectl gained a new "reboot" command that may be used to
  4860. trigger a reboot on a specific container that is registered
  4861. with machined. This works on any container that runs an init
  4862. system of some kind.
  4863. * systemctl gained a new "list-timers" command to print a nice
  4864. listing of installed timer units with the times they elapse
  4865. next.
  4866. * Alternative reboot() parameters may now be specified on the
  4867. "systemctl reboot" command line and are passed to the
  4868. reboot() system call.
  4869. * systemctl gained a new --job-mode= switch to configure the
  4870. mode to queue a job with. This is a more generic version of
  4871. --fail, --irreversible, and --ignore-dependencies, which are
  4872. still available but not advertised anymore.
  4873. * /etc/systemd/system.conf gained new settings to configure
  4874. various default timeouts of units, as well as the default
  4875. start limit interval and burst. These may still be overridden
  4876. within each Unit.
  4877. * PID1 will now export on the bus profile data of the security
  4878. policy upload process (such as the SELinux policy upload to
  4879. the kernel).
  4880. * journald: when forwarding logs to the console, include
  4881. timestamps (following the setting in
  4882. /sys/module/printk/parameters/time).
  4883. * OnCalendar= in timer units now understands the special
  4884. strings "yearly" and "annually". (Both are equivalent)
  4885. * The accuracy of timer units is now configurable with the new
  4886. AccuracySec= setting. It defaults to 1min.
  4887. * A new dependency type JoinsNamespaceOf= has been added that
  4888. allows running two services within the same /tmp and network
  4889. namespace, if PrivateNetwork= or PrivateTmp= are used.
  4890. * A new command "cat" has been added to systemctl. It outputs
  4891. the original unit file of a unit, and concatenates the
  4892. contents of additional "drop-in" unit file snippets, so that
  4893. the full configuration is shown.
  4894. * systemctl now supports globbing on the various "list-xyz"
  4895. commands, like "list-units" or "list-sockets", as well as on
  4896. those commands which take multiple unit names.
  4897. * journalctl's --unit= switch gained support for globbing.
  4898. * All systemd daemons now make use of the watchdog logic so
  4899. that systemd automatically notices when they hang.
  4900. * If the $container_ttys environment variable is set,
  4901. getty-generator will automatically spawn a getty for each
  4902. listed tty. This is useful for container managers to request
  4903. login gettys to be spawned on as many ttys as needed.
  4904. * %h, %s, %U specifier support is not available anymore when
  4905. used in unit files for PID 1. This is because NSS calls are
  4906. not safe from PID 1. They stay available for --user
  4907. instances of systemd, and as special case for the root user.
  4908. * loginctl gained a new "--no-legend" switch to turn off output
  4909. of the legend text.
  4910. * The "sd-login.h" API gained three new calls:
  4911. sd_session_is_remote(), sd_session_get_remote_user(),
  4912. sd_session_get_remote_host() to query information about
  4913. remote sessions.
  4914. * The udev hardware database now also carries vendor/product
  4915. information of SDIO devices.
  4916. * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
  4917. determine whether watchdog notifications are requested by
  4918. the system manager.
  4919. * Socket-activated per-connection services now include a
  4920. short description of the connection parameters in the
  4921. description.
  4922. * tmpfiles gained a new "--boot" option. When this is not used,
  4923. only lines where the command character is not suffixed with
  4924. "!" are executed. When this option is specified, those
  4925. options are executed too. This partitions tmpfiles
  4926. directives into those that can be safely executed at any
  4927. time, and those which should be run only at boot (for
  4928. example, a line that creates /run/nologin).
  4929. * A new API "sd-resolve.h" has been added which provides a simple
  4930. asynchronous wrapper around glibc NSS host name resolution
  4931. calls, such as getaddrinfo(). In contrast to glibc's
  4932. getaddrinfo_a(), it does not use signals. In contrast to most
  4933. other asynchronous name resolution libraries, this one does
  4934. not reimplement DNS, but reuses NSS, so that alternate
  4935. host name resolution systems continue to work, such as mDNS,
  4936. LDAP, etc. This API is based on libasyncns, but it has been
  4937. cleaned up for inclusion in systemd.
  4938. * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
  4939. "sd-daemon.h" are no longer found in individual libraries
  4941., Instead, we have
  4942. merged them into a single library,, which
  4943. provides all symbols. The reason for this is cyclic
  4944. dependencies, as these libraries tend to use each other's
  4945. symbols. So far, we have managed to workaround that by linking
  4946. a copy of a good part of our code into each of these
  4947. libraries again and again, which, however, makes certain
  4948. things hard to do, like sharing static variables. Also, it
  4949. substantially increases footprint. With this change, there
  4950. is only one library for the basic APIs systemd
  4951. provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
  4952. "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
  4953. library as well, however are subject to the --enable-kdbus
  4954. switch (see below). Note that "sd-dhcp-client.h" is not part
  4955. of this library (this is because it only consumes, never
  4956. provides, services of/to other APIs). To make the transition
  4957. easy from the separate libraries to the unified one, we
  4958. provide the --enable-compat-libs compile-time switch which
  4959. will generate stub libraries that are compatible with the
  4960. old ones but redirect all calls to the new one.
  4961. * All of the kdbus logic and the new APIs "sd-bus.h",
  4962. "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
  4963. and "sd-utf8.h" are compile-time optional via the
  4964. "--enable-kdbus" switch, and they are not compiled in by
  4965. default. To make use of kdbus, you have to explicitly enable
  4966. the switch. Note however, that neither the kernel nor the
  4967. userspace API for all of this is considered stable yet. We
  4968. want to maintain the freedom to still change the APIs for
  4969. now. By specifying this build-time switch, you acknowledge
  4970. that you are aware of the instability of the current
  4971. APIs.
  4972. * Also, note that while kdbus is pretty much complete,
  4973. it lacks one thing: proper policy support. This means you
  4974. can build a fully working system with all features; however,
  4975. it will be highly insecure. Policy support will be added in
  4976. one of the next releases, at the same time that we will
  4977. declare the APIs stable.
  4978. * When the kernel command line argument "kdbus" is specified,
  4979. systemd will automatically load the kdbus.ko kernel module. At
  4980. this stage of development, it is only useful for testing kdbus
  4981. and should not be used in production. Note: if "--enable-kdbus"
  4982. is specified, and the kdbus.ko kernel module is available, and
  4983. "kdbus" is added to the kernel command line, the entire system
  4984. runs with kdbus instead of dbus-daemon, with the above mentioned
  4985. problem of missing the system policy enforcement. Also a future
  4986. version of kdbus.ko or a newer systemd will not be compatible with
  4987. each other, and will unlikely be able to boot the machine if only
  4988. one of them is updated.
  4989. * systemctl gained a new "import-environment" command which
  4990. uploads the caller's environment (or parts thereof) into the
  4991. service manager so that it is inherited by services started
  4992. by the manager. This is useful to upload variables like
  4993. $DISPLAY into the user service manager.
  4994. * A new PrivateDevices= switch has been added to service units
  4995. which allows running a service with a namespaced /dev
  4996. directory that does not contain any device nodes for
  4997. physical devices. More specifically, it only includes devices
  4998. such as /dev/null, /dev/urandom, and /dev/zero which are API
  4999. entry points.
  5000. * logind has been extended to support behaviour like VT
  5001. switching on seats that do not support a VT. This makes
  5002. multi-session available on seats that are not the first seat
  5003. (seat0), and on systems where kernel support for VTs has
  5004. been disabled at compile-time.
  5005. * If a process holds a delay lock for system sleep or shutdown
  5006. and fails to release it in time, we will now log its
  5007. identity. This makes it easier to identify processes that
  5008. cause slow suspends or power-offs.
  5009. * When parsing /etc/crypttab, support for a new key-slot=
  5010. option as supported by Debian is added. It allows indicating
  5011. which LUKS slot to use on disk, speeding up key loading.
  5012. * The sd_journal_sendv() API call has been checked and
  5013. officially declared to be async-signal-safe so that it may
  5014. be invoked from signal handlers for logging purposes.
  5015. * Boot-time status output is now enabled automatically after a
  5016. short timeout if boot does not progress, in order to give
  5017. the user an indication what she or he is waiting for.
  5018. * The boot-time output has been improved to show how much time
  5019. remains until jobs expire.
  5020. * The KillMode= switch in service units gained a new possible
  5021. value "mixed". If set, and the unit is shut down, then the
  5022. initial SIGTERM signal is sent only to the main daemon
  5023. process, while the following SIGKILL signal is sent to
  5024. all remaining processes of the service.
  5025. * When a scope unit is registered, a new property "Controller"
  5026. may be set. If set to a valid bus name, systemd will send a
  5027. RequestStop() signal to this name when it would like to shut
  5028. down the scope. This may be used to hook manager logic into
  5029. the shutdown logic of scope units. Also, scope units may now
  5030. be put in a special "abandoned" state, in which case the
  5031. manager process which created them takes no further
  5032. responsibilities for it.
  5033. * When reading unit files, systemd will now verify
  5034. the access mode of these files, and warn about certain
  5035. suspicious combinations. This has been added to make it
  5036. easier to track down packaging bugs where unit files are
  5037. marked executable or world-writable.
  5038. * systemd-nspawn gained a new "--setenv=" switch to set
  5039. container-wide environment variables. The similar option in
  5040. systemd-activate was renamed from "--environment=" to
  5041. "--setenv=" for consistency.
  5042. * systemd-nspawn has been updated to create a new kdbus domain
  5043. for each container that is invoked, thus allowing each
  5044. container to have its own set of system and user buses,
  5045. independent of the host.
  5046. * systemd-nspawn gained a new --drop-capability= switch to run
  5047. the container with less capabilities than the default. Both
  5048. --drop-capability= and --capability= now take the special
  5049. string "all" for dropping or keeping all capabilities.
  5050. * systemd-nspawn gained new switches for executing containers
  5051. with specific SELinux labels set.
  5052. * systemd-nspawn gained a new --quiet switch to not generate
  5053. any additional output but the container's own console
  5054. output.
  5055. * systemd-nspawn gained a new --share-system switch to run a
  5056. container without PID namespacing enabled.
  5057. * systemd-nspawn gained a new --register= switch to control
  5058. whether the container is registered with systemd-machined or
  5059. not. This is useful for containers that do not run full
  5060. OS images, but only specific apps.
  5061. * systemd-nspawn gained a new --keep-unit which may be used
  5062. when invoked as the only program from a service unit, and
  5063. results in registration of the unit service itself in
  5064. systemd-machined, instead of a newly opened scope unit.
  5065. * systemd-nspawn gained a new --network-interface= switch for
  5066. moving arbitrary interfaces to the container. The new
  5067. --network-veth switch creates a virtual Ethernet connection
  5068. between host and container. The new --network-bridge=
  5069. switch then allows assigning the host side of this virtual
  5070. Ethernet connection to a bridge device.
  5071. * systemd-nspawn gained a new --personality= switch for
  5072. setting the kernel personality for the container. This is
  5073. useful when running a 32-bit container on a 64-bit host. A
  5074. similar option Personality= is now also available for service
  5075. units to use.
  5076. * logind will now also track a "Desktop" identifier for each
  5077. session which encodes the desktop environment of it. This is
  5078. useful for desktop environments that want to identify
  5079. multiple running sessions of itself easily.
  5080. * A new SELinuxContext= setting for service units has been
  5081. added that allows setting a specific SELinux execution
  5082. context for a service.
  5083. * Most systemd client tools will now honour $SYSTEMD_LESS for
  5084. settings of the "less" pager. By default, these tools will
  5085. override $LESS to allow certain operations to work, such as
  5086. jump-to-the-end. With $SYSTEMD_LESS, it is possible to
  5087. influence this logic.
  5088. * systemd's "seccomp" hook-up has been changed to make use of
  5089. the libseccomp library instead of using its own
  5090. implementation. This has benefits for portability among
  5091. other things.
  5092. * For usage together with SystemCallFilter=, a new
  5093. SystemCallErrorNumber= setting has been introduced that
  5094. allows configuration of a system error number to be returned
  5095. on filtered system calls, instead of immediately killing the
  5096. process. Also, SystemCallArchitectures= has been added to
  5097. limit access to system calls of a particular architecture
  5098. (in order to turn off support for unused secondary
  5099. architectures). There is also a global
  5100. SystemCallArchitectures= setting in system.conf now to turn
  5101. off support for non-native system calls system-wide.
  5102. * systemd requires a kernel with a working name_to_handle_at(),
  5103. please see the kernel config requirements in the README file.
  5104. Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
  5105. Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
  5106. Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
  5107. Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
  5108. Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
  5109. David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
  5110. Elia Pinto, Florian Weimer, George McCollister, Goffredo
  5111. Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
  5112. Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
  5113. Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
  5114. Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
  5115. Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
  5116. Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
  5117. Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
  5118. Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
  5119. Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
  5120. Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
  5121. Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
  5122. Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
  5123. Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
  5124. Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
  5125. Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
  5126. Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
  5127. Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
  5128. — Berlin, 2014-02-20
  5129. CHANGES WITH 208:
  5130. * logind has gained support for facilitating privileged input
  5131. and drm device access for unprivileged clients. This work is
  5132. useful to allow Wayland display servers (and similar
  5133. programs, such as kmscon) to run under the user's ID and
  5134. access input and drm devices which are normally
  5135. protected. When this is used (and the kernel is new enough)
  5136. logind will "mute" IO on the file descriptors passed to
  5137. Wayland as long as it is in the background and "unmute" it
  5138. if it returns into the foreground. This allows secure
  5139. session switching without allowing background sessions to
  5140. eavesdrop on input and display data. This also introduces
  5141. session switching support if VT support is turned off in the
  5142. kernel, and on seats that are not seat0.
  5143. * A new kernel command line option luks.options= is understood
  5144. now which allows specifying LUKS options for usage for LUKS
  5145. encrypted partitions specified with luks.uuid=.
  5146. * tmpfiles.d(5) snippets may now use specifier expansion in
  5147. path names. More specifically %m, %b, %H, %v, are now
  5148. replaced by the local machine id, boot id, hostname, and
  5149. kernel version number.
  5150. * A new tmpfiles.d(5) command "m" has been introduced which
  5151. may be used to change the owner/group/access mode of a file
  5152. or directory if it exists, but do nothing if it does not.
  5153. * This release removes high-level support for the
  5154. MemorySoftLimit= cgroup setting. The underlying kernel
  5155. cgroup attribute memory.soft_limit= is currently badly
  5156. designed and likely to be removed from the kernel API in its
  5157. current form, hence we should not expose it for now.
  5158. * The memory.use_hierarchy cgroup attribute is now enabled for
  5159. all cgroups systemd creates in the memory cgroup
  5160. hierarchy. This option is likely to be come the built-in
  5161. default in the kernel anyway, and the non-hierarchical mode
  5162. never made much sense in the intrinsically hierarchical
  5163. cgroup system.
  5164. * A new field _SYSTEMD_SLICE= is logged along with all journal
  5165. messages containing the slice a message was generated
  5166. from. This is useful to allow easy per-customer filtering of
  5167. logs among other things.
  5168. * systemd-journald will no longer adjust the group of journal
  5169. files it creates to the "systemd-journal" group. Instead we
  5170. rely on the journal directory to be owned by the
  5171. "systemd-journal" group, and its setgid bit set, so that the
  5172. kernel file system layer will automatically enforce that
  5173. journal files inherit this group assignment. The reason for
  5174. this change is that we cannot allow NSS look-ups from
  5175. journald which would be necessary to resolve
  5176. "systemd-journal" to a numeric GID, because this might
  5177. create deadlocks if NSS involves synchronous queries to
  5178. other daemons (such as nscd, or sssd) which in turn are
  5179. logging clients of journald and might block on it, which
  5180. would then dead lock. A tmpfiles.d(5) snippet included in
  5181. systemd will make sure the setgid bit and group are
  5182. properly set on the journal directory if it exists on every
  5183. boot. However, we recommend adjusting it manually after
  5184. upgrades too (or from RPM scriptlets), so that the change is
  5185. not delayed until next reboot.
  5186. * Backlight and random seed files in /var/lib/ have moved into
  5187. the /var/lib/systemd/ directory, in order to centralize all
  5188. systemd generated files in one directory.
  5189. * Boot time performance measurements (as displayed by
  5190. "systemd-analyze" for example) will now read ACPI 5.0 FPDT
  5191. performance information if that's available to determine how
  5192. much time BIOS and boot loader initialization required. With
  5193. a sufficiently new BIOS you hence no longer need to boot
  5194. with Gummiboot to get access to such information.
  5195. Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
  5196. Cristian Rodríguez, Dave Reisner, David Herrmann, David
  5197. Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
  5198. feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
  5199. Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
  5200. Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
  5201. Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
  5202. — Berlin, 2013-10-02
  5203. CHANGES WITH 207:
  5204. * The Restart= option for services now understands a new
  5205. on-watchdog setting, which will restart the service
  5206. automatically if the service stops sending out watchdog keep
  5207. alive messages (as configured with WatchdogSec=).
  5208. * The getty generator (which is responsible for bringing up a
  5209. getty on configured serial consoles) will no longer only
  5210. start a getty on the primary kernel console but on all
  5211. others, too. This makes the order in which console= is
  5212. specified on the kernel command line less important.
  5213. * libsystemd-logind gained a new sd_session_get_vt() call to
  5214. retrieve the VT number of a session.
  5215. * If the option "tries=0" is set for an entry of /etc/crypttab
  5216. its passphrase is queried indefinitely instead of any
  5217. maximum number of tries.
  5218. * If a service with a configure PID file terminates its PID
  5219. file will now be removed automatically if it still exists
  5220. afterwards. This should put an end to stale PID files.
  5221. * systemd-run will now also take relative binary path names
  5222. for execution and no longer insists on absolute paths.
  5223. * InaccessibleDirectories= and ReadOnlyDirectories= now take
  5224. paths that are optionally prefixed with "-" to indicate that
  5225. it should not be considered a failure if they do not exist.
  5226. * journalctl -o (and similar commands) now understands a new
  5227. output mode "short-precise", it is similar to "short" but
  5228. shows timestamps with usec accuracy.
  5229. * The option "discard" (as known from Debian) is now
  5230. synonymous to "allow-discards" in /etc/crypttab. In fact,
  5231. "discard" is preferred now (since it is easier to remember
  5232. and type).
  5233. * Some licensing clean-ups were made, so that more code is now
  5234. LGPL-2.1 licensed than before.
  5235. * A minimal tool to save/restore the display backlight
  5236. brightness across reboots has been added. It will store the
  5237. backlight setting as late as possible at shutdown, and
  5238. restore it as early as possible during reboot.
  5239. * A logic to automatically discover and enable home and swap
  5240. partitions on GPT disks has been added. With this in place
  5241. /etc/fstab becomes optional for many setups as systemd can
  5242. discover certain partitions located on the root disk
  5243. automatically. Home partitions are recognized under their
  5244. GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
  5245. partitions are recognized under their GPT type ID
  5246. 0657fd6da4ab43c484e50933c84b4f4f.
  5247. * systemd will no longer pass any environment from the kernel
  5248. or initrd to system services. If you want to set an
  5249. environment for all services, do so via the kernel command
  5250. line systemd.setenv= assignment.
  5251. * The systemd-sysctl tool no longer natively reads the file
  5252. /etc/sysctl.conf. If desired, the file should be symlinked
  5253. from /etc/sysctl.d/99-sysctl.conf. Apart from providing
  5254. legacy support by a symlink rather than built-in code, it
  5255. also makes the otherwise hidden order of application of the
  5256. different files visible. (Note that this partly reverts to a
  5257. pre-198 application order of sysctl knobs!)
  5258. * The "systemctl set-log-level" and "systemctl dump" commands
  5259. have been moved to systemd-analyze.
  5260. * systemd-run learned the new --remain-after-exit switch,
  5261. which causes the scope unit not to be cleaned up
  5262. automatically after the process terminated.
  5263. * tmpfiles learned a new --exclude-prefix= switch to exclude
  5264. certain paths from operation.
  5265. * journald will now automatically flush all messages to disk
  5266. as soon as a message at the log level CRIT, ALERT or EMERG
  5267. is received.
  5268. Contributions from: Andrew Cook, Brandon Philips, Christian
  5269. Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
  5270. Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
  5271. McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
  5272. Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
  5273. Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
  5274. Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
  5275. Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
  5276. Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
  5277. Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
  5278. Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
  5279. Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
  5280. William Giokas, Zbigniew Jędrzejewski-Szmek
  5281. — Berlin, 2013-09-13
  5282. CHANGES WITH 206:
  5283. * The documentation has been updated to cover the various new
  5284. concepts introduced with 205.
  5285. * Unit files now understand the new %v specifier which
  5286. resolves to the kernel version string as returned by "uname
  5287. -r".
  5288. * systemctl now supports filtering the unit list output by
  5289. load state, active state and sub state, using the new
  5290. --state= parameter.
  5291. * "systemctl status" will now show the results of the
  5292. condition checks (like ConditionPathExists= and similar) of
  5293. the last start attempts of the unit. They are also logged to
  5294. the journal.
  5295. * "journalctl -b" may now be used to look for boot output of a
  5296. specific boot. Try "journalctl -b -1" for the previous boot,
  5297. but the syntax is substantially more powerful.
  5298. * "journalctl --show-cursor" has been added which prints the
  5299. cursor string the last shown log line. This may then be used
  5300. with the new "journalctl --after-cursor=" switch to continue
  5301. browsing logs from that point on.
  5302. * "journalctl --force" may now be used to force regeneration
  5303. of an FSS key.
  5304. * Creation of "dead" device nodes has been moved from udev
  5305. into kmod and tmpfiles. Previously, udev would read the kmod
  5306. databases to pre-generate dead device nodes based on meta
  5307. information contained in kernel modules, so that these would
  5308. be auto-loaded on access rather then at boot. As this
  5309. does not really have much to do with the exposing actual
  5310. kernel devices to userspace this has always been slightly
  5311. alien in the udev codebase. Following the new scheme kmod
  5312. will now generate a runtime snippet for tmpfiles from the
  5313. module meta information and it now is tmpfiles' job to the
  5314. create the nodes. This also allows overriding access and
  5315. other parameters for the nodes using the usual tmpfiles
  5316. facilities. As side effect this allows us to remove the
  5317. CAP_SYS_MKNOD capability bit from udevd entirely.
  5318. * logind's device ACLs may now be applied to these "dead"
  5319. devices nodes too, thus finally allowing managed access to
  5320. devices such as /dev/snd/sequencer without loading the
  5321. backing module right-away.
  5322. * A new RPM macro has been added that may be used to apply
  5323. tmpfiles configuration during package installation.
  5324. * systemd-detect-virt and ConditionVirtualization= now can
  5325. detect User-Mode-Linux machines (UML).
  5326. * journald will now implicitly log the effective capabilities
  5327. set of processes in the message metadata.
  5328. * systemd-cryptsetup has gained support for TrueCrypt volumes.
  5329. * The initrd interface has been simplified (more specifically,
  5330. support for passing performance data via environment
  5331. variables and fsck results via files in /run has been
  5332. removed). These features were non-essential, and are
  5333. nowadays available in a much nicer way by having systemd in
  5334. the initrd serialize its state and have the hosts systemd
  5335. deserialize it again.
  5336. * The udev "keymap" data files and tools to apply keyboard
  5337. specific mappings of scan to key codes, and force-release
  5338. scan code lists have been entirely replaced by a udev
  5339. "keyboard" builtin and a hwdb data file.
  5340. * systemd will now honour the kernel's "quiet" command line
  5341. argument also during late shutdown, resulting in a
  5342. completely silent shutdown when used.
  5343. * There's now an option to control the SO_REUSEPORT socket
  5344. option in .socket units.
  5345. * Instance units will now automatically get a per-template
  5346. subslice of system.slice unless something else is explicitly
  5347. configured. For example, instances of sshd@.service will now
  5348. implicitly be placed in system-sshd.slice rather than
  5349. system.slice as before.
  5350. * Test coverage support may now be enabled at build time.
  5351. Contributions from: Dave Reisner, Frederic Crozat, Harald
  5352. Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
  5353. Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
  5354. Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
  5355. Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
  5356. Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
  5357. Giokas, Zbigniew Jędrzejewski-Szmek
  5358. — Berlin, 2013-07-23
  5359. CHANGES WITH 205:
  5360. * Two new unit types have been introduced:
  5361. Scope units are very similar to service units, however, are
  5362. created out of pre-existing processes — instead of PID 1
  5363. forking off the processes. By using scope units it is
  5364. possible for system services and applications to group their
  5365. own child processes (worker processes) in a powerful way
  5366. which then maybe used to organize them, or kill them
  5367. together, or apply resource limits on them.
  5368. Slice units may be used to partition system resources in an
  5369. hierarchical fashion and then assign other units to them. By
  5370. default there are now three slices: system.slice (for all
  5371. system services), user.slice (for all user sessions),
  5372. machine.slice (for VMs and containers).
  5373. Slices and scopes have been introduced primarily in
  5374. context of the work to move cgroup handling to a
  5375. single-writer scheme, where only PID 1
  5376. creates/removes/manages cgroups.
  5377. * There's a new concept of "transient" units. In contrast to
  5378. normal units these units are created via an API at runtime,
  5379. not from configuration from disk. More specifically this
  5380. means it is now possible to run arbitrary programs as
  5381. independent services, with all execution parameters passed
  5382. in via bus APIs rather than read from disk. Transient units
  5383. make systemd substantially more dynamic then it ever was,
  5384. and useful as a general batch manager.
  5385. * logind has been updated to make use of scope and slice units
  5386. for managing user sessions. As a user logs in he will get
  5387. his own private slice unit, to which all sessions are added
  5388. as scope units. We also added support for automatically
  5389. adding an instance of user@.service for the user into the
  5390. slice. Effectively logind will no longer create cgroup
  5391. hierarchies on its own now, it will defer entirely to PID 1
  5392. for this by means of scope, service and slice units. Since
  5393. user sessions this way become entities managed by PID 1
  5394. the output of "systemctl" is now a lot more comprehensive.
  5395. * A new mini-daemon "systemd-machined" has been added which
  5396. may be used by virtualization managers to register local
  5397. VMs/containers. nspawn has been updated accordingly, and
  5398. libvirt will be updated shortly. machined will collect a bit
  5399. of meta information about the VMs/containers, and assign
  5400. them their own scope unit (see above). The collected
  5401. meta-data is then made available via the "machinectl" tool,
  5402. and exposed in "ps" and similar tools. machined/machinectl
  5403. is compile-time optional.
  5404. * As discussed earlier, the low-level cgroup configuration
  5405. options ControlGroup=, ControlGroupModify=,
  5406. ControlGroupPersistent=, ControlGroupAttribute= have been
  5407. removed. Please use high-level attribute settings instead as
  5408. well as slice units.
  5409. * A new bus call SetUnitProperties() has been added to alter
  5410. various runtime parameters of a unit. This is primarily
  5411. useful to alter cgroup parameters dynamically in a nice way,
  5412. but will be extended later on to make more properties
  5413. modifiable at runtime. systemctl gained a new set-properties
  5414. command that wraps this call.
  5415. * A new tool "systemd-run" has been added which can be used to
  5416. run arbitrary command lines as transient services or scopes,
  5417. while configuring a number of settings via the command
  5418. line. This tool is currently very basic, however already
  5419. very useful. We plan to extend this tool to even allow
  5420. queuing of execution jobs with time triggers from the
  5421. command line, similar in fashion to "at".
  5422. * nspawn will now inform the user explicitly that kernels with
  5423. audit enabled break containers, and suggest the user to turn
  5424. off audit.
  5425. * Support for detecting the IMA and AppArmor security
  5426. frameworks with ConditionSecurity= has been added.
  5427. * journalctl gained a new "-k" switch for showing only kernel
  5428. messages, mimicking dmesg output; in addition to "--user"
  5429. and "--system" switches for showing only user's own logs
  5430. and system logs.
  5431. * systemd-delta can now show information about drop-in
  5432. snippets extending unit files.
  5433. * libsystemd-bus has been substantially updated but is still
  5434. not available as public API.
  5435. * systemd will now look for the "debug" argument on the kernel
  5436. command line and enable debug logging, similar to what
  5437. "systemd.log_level=debug" already did before.
  5438. * "systemctl set-default", "systemctl get-default" has been
  5439. added to configure the symlink, which
  5440. controls what to boot into by default.
  5441. * "systemctl set-log-level" has been added as a convenient
  5442. way to raise and lower systemd logging threshold.
  5443. * "systemd-analyze plot" will now show the time the various
  5444. generators needed for execution, as well as information
  5445. about the unit file loading.
  5446. * libsystemd-journal gained a new sd_journal_open_files() call
  5447. for opening specific journal files. journactl also gained a
  5448. new switch to