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.

7364 lines
370 KiB

  1. systemd System and Service Manager
  2. CHANGES WITH 234:
  3. * Meson is now supported as build system in addition to Automake. It is
  4. our plan to remove Automake in one of our next releases, so that
  5. Meson becomes our exclusive build system. Hence, please start using
  6. the Meson build system in your downstream packaging. There's plenty
  7. of documentation around how to use Meson, the extremely brief
  8. summary:
  9. ./ && ./configure && make && sudo make install
  10. becomes:
  11. meson build && ninja -C build && sudo ninja -C build install
  12. * Unit files gained support for a new JobRunningTimeoutUSec= setting,
  13. which permits configuring a timeout on the time a job is
  14. running. This is particularly useful for setting timeouts on jobs for
  15. .device units.
  16. * Unit files gained two new options ConditionUser= and ConditionGroup=
  17. for conditionalizing units based on the identity of the user/group
  18. running a systemd user instance.
  19. * systemd-networkd now understands a new FlowLabel= setting in the
  20. [VXLAN] section of .network files, as well as a Priority= in
  21. [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
  22. and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
  23. gained support for configuration of GENEVE links, and IPv6 address
  24. labels. The [Network] section gained the new IPv6ProxyNDP= setting.
  25. * .link files now understand a new Port= setting.
  26. * systemd-networkd's DHCP support gained support for DHCP option 119
  27. (domain search list).
  28. * systemd-networkd gained support for serving IPv6 address ranges using
  29. the Router Advertisment protocol. The new .network configuration
  30. section [IPv6Prefix] may be used to configure the ranges to
  31. serve. This is implemented based on a new, minimal, native server
  32. implementation of RA.
  33. * journalctl's --output= switch gained support for a new parameter
  34. "short-iso-precise" for a mode where timestamps are shown as precise
  35. ISO date values.
  36. * systemd-udevd's "net_id" builtin may now generate stable network
  37. interface names from IBM PowerVM VIO devices as well as ACPI platform
  38. devices.
  39. * MulticastDNS support in systemd-resolved may now be explicitly
  40. enabled/disabled using the new MulticastDNS= configuration file
  41. option.
  42. * systemd-resolved may now optionally use libidn2 instead of the libidn
  43. for processing internationalized domain names. Support for libidn2
  44. should be considered experimental and should not be enabled by
  45. default yet.
  46. * "machinectl pull-tar" and related call may now do verification of
  47. downloaded images using SUSE-style .sha256 checksum files in addition
  48. to the already existing support for validating using Ubuntu-style
  49. SHA256SUMS files.
  50. * sd-bus gained support for a new sd_bus_message_appendv() call which
  51. is va_list equivalent of sd_bus_message_append().
  52. * sd-boot gained support for validating images using SHIM/MOK.
  53. * The SMACK code learnt support for "onlycap".
  54. * systemd-mount --umount is now much smarter in figuring out how to
  55. properly unmount a device given its mount or device path.
  56. * The code to call libnss_dns as a fallback from libnss_resolve when
  57. the communication with systemd-resolved fails was removed. This
  58. fallback was redundant and interfered with the [!UNAVAIL=return]
  59. suffix. See nss-resolve(8) for the recommended configuration.
  60. * systemd-logind may now be restarted without losing state. It stores
  61. the file descriptors for devices it manages in the system manager
  62. using the FDSTORE= mechanism. Please note that further changes in
  63. other components may be required to make use of this (for example
  64. Xorg has code to listen for stops of systemd-logind and terminate
  65. itself when logind is stopped or restarted, in order to avoid using
  66. stale file descriptors for graphical devices, which is now
  67. counterproductive and must be reverted in order for restarts of
  68. systemd-logind to be safe. See
  70. * All kernel install plugins are called with the environment variable
  71. KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
  72. /etc/machine-id. If the file is missing or empty, the variable is
  73. empty and BOOT_DIR_ABS is the path of a temporary directory which is
  74. removed after all the plugins exit. So, if KERNEL_INSTALL_MACHINE_ID
  75. is empty, all plugins should not put anything in BOOT_DIR_ABS.
  76. Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
  77. Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
  78. Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
  79. Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
  80. Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
  81. Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
  82. Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
  83. Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
  84. Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
  85. Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
  86. hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
  87. Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
  88. Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
  89. Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
  90. Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
  91. Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
  92. Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
  93. Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
  94. Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
  95. Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
  96. Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
  97. Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
  98. Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
  99. Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
  100. Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
  101. H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
  102. Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
  103. userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
  104. Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
  105. Георгиевски
  106. — Berlin, 2017-07-12
  107. CHANGES WITH 233:
  108. * This version requires at least gperf 3.1 for building, 3.0 is not
  109. sufficient.
  110. * The "hybrid" control group mode has been modified to improve
  111. compatibility with "legacy" cgroups-v1 setups. Specifically, the
  112. "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
  113. "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
  114. cgroups-v1 hierarchy), the only externally visible change being that
  115. the cgroups-v2 hierarchy is also mounted, to
  116. /sys/fs/cgroup/unified. This should provide a large degree of
  117. compatibility with "legacy" cgroups-v1, while taking benefit of the
  118. better management capabilities of cgroups-v2.
  119. * The default control group setup mode may be selected both a boot-time
  120. via a set of kernel command line parameters (specifically:
  121. systemd.unified_cgroup_hierarchy= and
  122. systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
  123. default selected on the configure command line
  124. (--with-default-hierarchy=). The upstream default is "hybrid"
  125. (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
  126. this will change in a future systemd version to be "unified" (pure
  127. cgroups-v2 mode). The third option for the compile time option is
  128. "legacy", to enter pure cgroups-v1 mode. We recommend downstream
  129. distributions to default to "hybrid" mode for release distributions,
  130. starting with v233. We recommend "unified" for development
  131. distributions (specifically: distributions such as Fedora's rawhide)
  132. as that's where things are headed in the long run. Use "legacy" for
  133. greatest stability and compatibility only.
  134. * Note one current limitation of "unified" and "hybrid" control group
  135. setup modes: the kernel currently does not permit the systemd --user
  136. instance (i.e. unprivileged code) to migrate processes between two
  137. disconnected cgroup subtrees, even if both are managed and owned by
  138. the user. This effectively means "systemd-run --user --scope" doesn't
  139. work when invoked from outside of any "systemd --user" service or
  140. scope. Specifically, it is not supported from session scopes. We are
  141. working on fixing this in a future systemd version. (See #3388 for
  142. further details about this.)
  143. * DBus policy files are now installed into /usr rather than /etc. Make
  144. sure your system has dbus >= 1.9.18 running before upgrading to this
  145. version, or override the install path with --with-dbuspolicydir= .
  146. * All python scripts shipped with systemd (specifically: the various
  147. tests written in Python) now require Python 3.
  148. * systemd unit tests can now run standalone (without the source or
  149. build directories), and can be installed into /usr/lib/systemd/tests/
  150. with 'make install-tests'.
  151. * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
  152. CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
  153. kernel.
  154. * Support for the %c, %r, %R specifiers in unit files has been
  155. removed. Specifiers are not supposed to be dependent on configuration
  156. in the unit file itself (so that they resolve the same regardless
  157. where used in the unit files), but these specifiers were influenced
  158. by the Slice= option.
  159. * The shell invoked by debug-shell.service now defaults to /bin/sh in
  160. all cases. If distributions want to use a different shell for this
  161. purpose (for example Fedora's /sbin/sushell) they need to specify
  162. this explicitly at configure time using --with-debug-shell=.
  163. * The confirmation spawn prompt has been reworked to offer the
  164. following choices:
  165. (c)ontinue, proceed without asking anymore
  166. (D)ump, show the state of the unit
  167. (f)ail, don't execute the command and pretend it failed
  168. (h)elp
  169. (i)nfo, show a short summary of the unit
  170. (j)obs, show jobs that are in progress
  171. (s)kip, don't execute the command and pretend it succeeded
  172. (y)es, execute the command
  173. The 'n' choice for the confirmation spawn prompt has been removed,
  174. because its meaning was confusing.
  175. The prompt may now also be redirected to an alternative console by
  176. specifying the console as parameter to systemd.confirm_spawn=.
  177. * Services of Type=notify require a READY=1 notification to be sent
  178. during startup. If no such message is sent, the service now fails,
  179. even if the main process exited with a successful exit code.
  180. * Services that fail to start up correctly now always have their
  181. ExecStopPost= commands executed. Previously, they'd enter "failed"
  182. state directly, without executing these commands.
  183. * The option MulticastDNS= of network configuration files has acquired
  184. an actual implementation. With MulticastDNS=yes a host can resolve
  185. names of remote hosts and reply to mDNS A and AAAA requests.
  186. * When units are about to be started an additional check is now done to
  187. ensure that all dependencies of type BindsTo= (when used in
  188. combination with After=) have been started.
  189. * systemd-analyze gained a new verb "syscall-filter" which shows which
  190. system call groups are defined for the SystemCallFilter= unit file
  191. setting, and which system calls they contain.
  192. * A new system call filter group "@filesystem" has been added,
  193. consisting of various file system related system calls. Group
  194. "@reboot" has been added, covering reboot, kexec and shutdown related
  195. calls. Finally, group "@swap" has been added covering swap
  196. configuration related calls.
  197. * A new unit file option RestrictNamespaces= has been added that may be
  198. used to restrict access to the various process namespace types the
  199. Linux kernel provides. Specifically, it may be used to take away the
  200. right for a service unit to create additional file system, network,
  201. user, and other namespaces. This sandboxing option is particularly
  202. relevant due to the high amount of recently discovered namespacing
  203. related vulnerabilities in the kernel.
  204. * systemd-udev's .link files gained support for a new AutoNegotiation=
  205. setting for configuring Ethernet auto-negotiation.
  206. * systemd-networkd's .network files gained support for a new
  207. ListenPort= setting in the [DHCP] section to explicitly configure the
  208. UDP client port the DHCP client shall listen on.
  209. * .network files gained a new Unmanaged= boolean setting for explicitly
  210. excluding one or more interfaces from management by systemd-networkd.
  211. * The systemd-networkd ProxyARP= option has been renamed to
  212. IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
  213. renamed to ReduceARPProxy=. The old names continue to be available
  214. for compatibility.
  215. * systemd-networkd gained support for configuring IPv6 Proxy NDP
  216. addresses via the new IPv6ProxyNDPAddress= .network file setting.
  217. * systemd-networkd's bonding device support gained support for two new
  218. configuration options ActiveSlave= and PrimarySlave=.
  219. * The various options in the [Match] section of .network files gained
  220. support for negative matching.
  221. * New systemd-specific mount options are now understood in /etc/fstab:
  222. x-systemd.mount-timeout= may be used to configure the maximum
  223. permitted runtime of the mount command.
  224. x-systemd.device-bound may be set to bind a mount point to its
  225. backing device unit, in order to automatically remove a mount point
  226. if its backing device is unplugged. This option may also be
  227. configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
  228. on the block device, which is now automatically set for all CDROM
  229. drives, so that mounted CDs are automatically unmounted when they are
  230. removed from the drive.
  231. x-systemd.after= and x-systemd.before= may be used to explicitly
  232. order a mount after or before another unit or mount point.
  233. * Enqueued start jobs for device units are now automatically garbage
  234. collected if there are no jobs waiting for them anymore.
  235. * systemctl list-jobs gained two new switches: with --after, for every
  236. queued job the jobs it's waiting for are shown; with --before the
  237. jobs which it's blocking are shown.
  238. * systemd-nspawn gained support for ephemeral boots from disk images
  239. (or in other words: --ephemeral and --image= may now be
  240. combined). Moreover, ephemeral boots are now supported for normal
  241. directories, even if the backing file system is not btrfs. Of course,
  242. if the file system does not support file system snapshots or
  243. reflinks, the initial copy operation will be relatively expensive, but
  244. this should still be suitable for many use cases.
  245. * Calendar time specifications in .timer units now support
  246. specifications relative to the end of a month by using "~" instead of
  247. "-" as separator between month and day. For example, "*-02~03" means
  248. "the third last day in February". In addition a new syntax for
  249. repeated events has been added using the "/" character. For example,
  250. "9..17/2:00" means "every two hours from 9am to 5pm".
  251. * systemd-socket-proxyd gained a new parameter --connections-max= for
  252. configuring the maximum number of concurrent connections.
  253. * sd-id128 gained a new API for generating unique IDs for the host in a
  254. way that does not leak the machine ID. Specifically,
  255. sd_id128_get_machine_app_specific() derives an ID based on the
  256. machine ID a in well-defined, non-reversible, stable way. This is
  257. useful whenever an identifier for the host is needed but where the
  258. identifier shall not be useful to identify the system beyond the
  259. scope of the application itself. (Internally this uses HMAC-SHA256 as
  260. keyed hash function using the machine ID as input.)
  261. * NotifyAccess= gained a new supported value "exec". When set
  262. notifications are accepted from all processes systemd itself invoked,
  263. including all control processes.
  264. * .nspawn files gained support for defining overlay mounts using the
  265. Overlay= and OverlayReadOnly= options. Previously this functionality
  266. was only available on the systemd-nspawn command line.
  267. * systemd-nspawn's --bind= and --overlay= options gained support for
  268. bind/overlay mounts whose source lies within the container tree by
  269. prefixing the source path with "+".
  270. * systemd-nspawn's --bind= and --overlay= options gained support for
  271. automatically allocating a temporary source directory in /var/tmp
  272. that is removed when the container dies. Specifically, if the source
  273. directory is specified as empty string this mechanism is selected. An
  274. example usage is --overlay=+/var::/var, which creates an overlay
  275. mount based on the original /var contained in the image, overlayed
  276. with a temporary directory in the host's /var/tmp. This way changes
  277. to /var are automatically flushed when the container shuts down.
  278. * systemd-nspawn --image= option does now permit raw file system block
  279. devices (in addition to images containing partition tables, as
  280. before).
  281. * The disk image dissection logic in systemd-nspawn gained support for
  282. automatically setting up LUKS encrypted as well as Verity protected
  283. partitions. When a container is booted from an encrypted image the
  284. passphrase is queried at start-up time. When a container with Verity
  285. data is started, the root hash is search in a ".roothash" file
  286. accompanying the disk image (alternatively, pass the root hash via
  287. the new --root-hash= command line option).
  288. * A new tool /usr/lib/systemd/systemd-dissect has been added that may
  289. be used to dissect disk images the same way as systemd-nspawn does
  290. it, following the Bootable Partition Specification. It may even be
  291. used to mount disk images with complex partition setups (including
  292. LUKS and Verity partitions) to a local host directory, in order to
  293. inspect them. This tool is not considered public API (yet), and is
  294. thus not installed into /usr/bin. Please do not rely on its
  295. existence, since it might go away or be changed in later systemd
  296. versions.
  297. * A new generator "systemd-verity-generator" has been added, similar in
  298. style to "systemd-cryptsetup-generator", permitting automatic setup of
  299. Verity root partitions when systemd boots up. In order to make use of
  300. this your partition setup should follow the Discoverable Partitions
  301. Specification, and the GPT partition ID of the root file system
  302. partition should be identical to the upper 128bit of the Verity root
  303. hash. The GPT partition ID of the Verity partition protecting it
  304. should be the lower 128bit of the Verity root hash. If the partition
  305. image follows this model it is sufficient to specify a single
  306. "roothash=" kernel command line argument to both configure which root
  307. image and verity partition to use as well as the root hash for
  308. it. Note that systemd-nspawn's Verity support follows the same
  309. semantics, meaning that disk images with proper Verity data in place
  310. may be booted in containers with systemd-nspawn as well as on
  311. physical systems via the verity generator. Also note that the "mkosi"
  312. tool available at has been updated
  313. to generate Verity protected disk images following this scheme. In
  314. fact, it has been updated to generate disk images that optionally
  315. implement a complete UEFI SecureBoot trust chain, involving a signed
  316. kernel and initrd image that incorporates such a root hash as well as
  317. a Verity-enabled root partition.
  318. * The hardware database (hwdb) udev supports has been updated to carry
  319. accelerometer quirks.
  320. * All system services are now run with a fresh kernel keyring set up
  321. for them. The invocation ID is stored by default in it, thus
  322. providing a safe, non-overridable way to determine the invocation
  323. ID of each service.
  324. * Service unit files gained new BindPaths= and BindReadOnlyPaths=
  325. options for bind mounting arbitrary paths in a service-specific
  326. way. When these options are used, arbitrary host or service files and
  327. directories may be mounted to arbitrary locations in the service's
  328. view.
  329. * Documentation has been added that lists all of systemd's low-level
  330. environment variables:
  332. * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
  333. whether a specific socket file descriptor matches a specified socket
  334. address.
  335. * systemd-firstboot has been updated to check for the
  336. systemd.firstboot= kernel command line option. It accepts a boolean
  337. and when set to false the first boot questions are skipped.
  338. * systemd-fstab-generator has been updated to check for the
  339. systemd.volatile= kernel command line option, which either takes an
  340. optional boolean parameter or the special value "state". If used the
  341. system may be booted in a "volatile" boot mode. Specifically,
  342. "systemd.volatile" is used, the root directory will be mounted as
  343. tmpfs, and only /usr is mounted from the actual root file system. If
  344. "systemd.volatile=state" is used, the root directory will be mounted
  345. as usual, but /var is mounted as tmpfs. This concept provides similar
  346. functionality as systemd-nspawn's --volatile= option, but provides it
  347. on physical boots. Use this option for implementing stateless
  348. systems, or testing systems with all state and/or configuration reset
  349. to the defaults. (Note though that many distributions are not
  350. prepared to boot up without a populated /etc or /var, though.)
  351. * systemd-gpt-auto-generator gained support for LUKS encrypted root
  352. partitions. Previously it only supported LUKS encrypted partitions
  353. for all other uses, except for the root partition itself.
  354. * Socket units gained support for listening on AF_VSOCK sockets for
  355. communication in virtualized QEMU environments.
  356. * The "configure" script gained a new option --with-fallback-hostname=
  357. for specifying the fallback hostname to use if none is configured in
  358. /etc/hostname. For example, by specifying
  359. --with-fallback-hostname=fedora it is possible to default to a
  360. hostname of "fedora" on pristine installations.
  361. * systemd-cgls gained support for a new --unit= switch for listing only
  362. the control groups of a specific unit. Similar --user-unit= has been
  363. added for listing only the control groups of a specific user unit.
  364. * systemd-mount gained a new --umount switch for unmounting a mount or
  365. automount point (and all mount/automount points below it).
  366. * systemd will now refuse full configuration reloads (via systemctl
  367. daemon-reload and related calls) unless at least 16MiB of free space
  368. are available in /run. This is a safety precaution in order to ensure
  369. that generators can safely operate after the reload completed.
  370. * A new unit file option RootImage= has been added, which has a similar
  371. effect as RootDirectory= but mounts the service's root directory from
  372. a disk image instead of plain directory. This logic reuses the same
  373. image dissection and mount logic that systemd-nspawn already uses,
  374. and hence supports any disk images systemd-nspawn supports, including
  375. those following the Discoverable Partition Specification, as well as
  376. Verity enabled images. This option enables systemd to run system
  377. services directly off disk images acting as resource bundles,
  378. possibly even including full integrity data.
  379. * A new MountAPIVFS= unit file option has been added, taking a boolean
  380. argument. If enabled /proc, /sys and /dev (collectively called the
  381. "API VFS") will be mounted for the service. This is only relevant if
  382. RootDirectory= or RootImage= is used for the service, as these mounts
  383. are of course in place in the host mount namespace anyway.
  384. * systemd-nspawn gained support for a new --pivot-root= switch. If
  385. specified the root directory within the container image is pivoted to
  386. the specified mount point, while the original root disk is moved to a
  387. different place. This option enables booting of ostree images
  388. directly with systemd-nspawn.
  389. * The systemd build scripts will no longer complain if the NTP server
  390. addresses are not changed from the defaults. Google now supports
  391. these NTP servers officially. We still recommend downstreams to
  392. properly register an NTP pool with the NTP pool project though.
  393. * coredumpctl gained a new "--reverse" option for printing the list
  394. of coredumps in reverse order.
  395. * coredumpctl will now show additional information about truncated and
  396. inaccessible coredumps, as well as coredumps that are still being
  397. processed. It also gained a new --quiet switch for suppressing
  398. additional informational message in its output.
  399. * coredumpctl gained support for only showing coredumps newer and/or
  400. older than specific timestamps, using the new --since= and --until=
  401. options, reminiscent of journalctl's options by the same name.
  402. * The systemd-coredump logic has been improved so that it may be reused
  403. to collect backtraces in non-compiled languages, for example in
  404. scripting languages such as Python.
  405. * machinectl will now show the UID shift of local containers, if user
  406. namespacing is enabled for them.
  407. * systemd will now optionally run "environment generator" binaries at
  408. configuration load time. They may be used to add environment
  409. variables to the environment block passed to services invoked. One
  410. user environment generator is shipped by default that sets up
  411. environment variables based on files dropped into /etc/environment.d
  412. and ~/.config/environment.d/.
  413. * systemd-resolved now includes the new, recently published 2017 DNSSEC
  414. root key (KSK).
  415. * hostnamed has been updated to report a new chassis type of
  416. "convertible" to cover "foldable" laptops that can both act as a
  417. tablet and as a laptop, such as various Lenovo Yoga devices.
  418. Contributions from: Adrián López, Alexander Galanin, Alexander
  419. Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
  420. Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
  421. Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
  422. Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
  423. David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
  424. Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
  425. Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
  426. Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
  427. Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
  428. Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
  429. Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
  430. Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
  431. Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
  432. Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
  433. Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
  434. Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
  435. Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
  436. Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
  437. Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
  438. Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
  439. Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
  440. Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
  441. Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
  442. Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
  443. Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
  444. Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
  445. YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
  446. Тихонов
  447. — Berlin, 2017-03-01
  448. CHANGES WITH 232:
  449. * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
  450. RestrictAddressFamilies= enabled. These sandboxing options should
  451. generally be compatible with the various external udev call-out
  452. binaries we are aware of, however there may be exceptions, in
  453. particular when exotic languages for these call-outs are used. In
  454. this case, consider turning off these settings locally.
  455. * The new RemoveIPC= option can be used to remove IPC objects owned by
  456. the user or group of a service when that service exits.
  457. * The new ProtectKernelModules= option can be used to disable explicit
  458. load and unload operations of kernel modules by a service. In
  459. addition access to /usr/lib/modules is removed if this option is set.
  460. * ProtectSystem= option gained a new value "strict", which causes the
  461. whole file system tree with the exception of /dev, /proc, and /sys,
  462. to be remounted read-only for a service.
  463. * The new ProtectKernelTunables= option can be used to disable
  464. modification of configuration files in /sys and /proc by a service.
  465. Various directories and files are remounted read-only, so access is
  466. restricted even if the file permissions would allow it.
  467. * The new ProtectControlGroups= option can be used to disable write
  468. access by a service to /sys/fs/cgroup.
  469. * Various systemd services have been hardened with
  470. ProtectKernelTunables=yes, ProtectControlGroups=yes,
  471. RestrictAddressFamilies=.
  472. * Support for dynamically creating users for the lifetime of a service
  473. has been added. If DynamicUser=yes is specified, user and group IDs
  474. will be allocated from the range 61184..65519 for the lifetime of the
  475. service. They can be resolved using the new NSS
  476. module. The module must be enabled in /etc/nsswitch.conf. Services
  477. started in this way have PrivateTmp= and RemoveIPC= enabled, so that
  478. any resources allocated by the service will be cleaned up when the
  479. service exits. They also have ProtectHome=read-only and
  480. ProtectSystem=strict enabled, so they are not able to make any
  481. permanent modifications to the system.
  482. * The nss-systemd module also always resolves root and nobody, making
  483. it possible to have no /etc/passwd or /etc/group files in minimal
  484. container or chroot environments.
  485. * Services may be started with their own user namespace using the new
  486. boolean PrivateUsers= option. Only root, nobody, and the uid/gid
  487. under which the service is running are mapped. All other users are
  488. mapped to nobody.
  489. * Support for the cgroup namespace has been added to systemd-nspawn. If
  490. supported by kernel, the container system started by systemd-nspawn
  491. will have its own view of the cgroup hierarchy. This new behaviour
  492. can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
  493. * The new MemorySwapMax= option can be used to limit the maximum swap
  494. usage under the unified cgroup hierarchy.
  495. * Support for the CPU controller in the unified cgroup hierarchy has
  496. been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
  497. options. This controller requires out-of-tree patches for the kernel
  498. and the support is provisional.
  499. * Mount and automount units may now be created transiently
  500. (i.e. dynamically at runtime via the bus API, instead of requiring
  501. unit files in the file system).
  502. * systemd-mount is a new tool which may mount file systems – much like
  503. mount(8), optionally pulling in additional dependencies through
  504. transient .mount and .automount units. For example, this tool
  505. automatically runs fsck on a backing block device before mounting,
  506. and allows the automount logic to be used dynamically from the
  507. command line for establishing mount points. This tool is particularly
  508. useful when dealing with removable media, as it will ensure fsck is
  509. run – if necessary – before the first access and that the file system
  510. is quickly unmounted after each access by utilizing the automount
  511. logic. This maximizes the chance that the file system on the
  512. removable media stays in a clean state, and if it isn't in a clean
  513. state is fixed automatically.
  514. * LazyUnmount=yes option for mount units has been added to expose the
  515. umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
  516. option.
  517. * /efi will be used as the mount point of the EFI boot partition, if
  518. the directory is present, and the mount point was not configured
  519. through other means (e.g. fstab). If /efi directory does not exist,
  520. /boot will be used as before. This makes it easier to automatically
  521. mount the EFI partition on systems where /boot is used for something
  522. else.
  523. * When operating on GPT disk images for containers, systemd-nspawn will
  524. now mount the ESP to /boot or /efi according to the same rules as PID
  525. 1 running on a host. This allows tools like "bootctl" to operate
  526. correctly within such containers, in order to make container images
  527. bootable on physical systems.
  528. * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
  529. * Two new user session targets have been added to support running
  530. graphical sessions under the systemd --user instance:
  531. and See
  532. systemd.special(7) for a description of how those targets should be
  533. used.
  534. * The vconsole initialization code has been significantly reworked to
  535. use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
  536. support unicode keymaps. Font and keymap configuration will now be
  537. copied to all allocated virtual consoles.
  538. * FreeBSD's bhyve virtualization is now detected.
  539. * Information recorded in the journal for core dumps now includes the
  540. contents of /proc/mountinfo and the command line of the process at
  541. the top of the process hierarchy (which is usually the init process
  542. of the container).
  543. * systemd-journal-gatewayd learned the --directory= option to serve
  544. files from the specified location.
  545. * journalctl --root=… can be used to peruse the journal in the
  546. /var/log/ directories inside of a container tree. This is similar to
  547. the existing --machine= option, but does not require the container to
  548. be active.
  549. * The hardware database has been extended to support
  550. ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
  551. trackball devices.
  552. MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
  553. specify the click rate for mice which include a horizontal wheel with
  554. a click rate that is different than the one for the vertical wheel.
  555. * systemd-run gained a new --wait option that makes service execution
  556. synchronous. (Specifically, the command will not return until the
  557. specified service binary exited.)
  558. * systemctl gained a new --wait option that causes the start command to
  559. wait until the units being started have terminated again.
  560. * A new journal output mode "short-full" has been added which displays
  561. timestamps with abbreviated English day names and adds a timezone
  562. suffix. Those timestamps include more information than the default
  563. "short" output mode, and can be passed directly to journalctl's
  564. --since= and --until= options.
  565. * /etc/resolv.conf will be bind-mounted into containers started by
  566. systemd-nspawn, if possible, so any changes to resolv.conf contents
  567. are automatically propagated to the container.
  568. * The number of instances for socket-activated services originating
  569. from a single IP address can be limited with
  570. MaxConnectionsPerSource=, extending the existing setting of
  571. MaxConnections=.
  572. * systemd-networkd gained support for vcan ("Virtual CAN") interface
  573. configuration.
  574. * .netdev and .network configuration can now be extended through
  575. drop-ins.
  576. * UDP Segmentation Offload, TCP Segmentation Offload, Generic
  577. Segmentation Offload, Generic Receive Offload, Large Receive Offload
  578. can be enabled and disabled using the new UDPSegmentationOffload=,
  579. TCPSegmentationOffload=, GenericSegmentationOffload=,
  580. GenericReceiveOffload=, LargeReceiveOffload= options in the
  581. [Link] section of .link files.
  582. * The Spanning Tree Protocol, Priority, Aging Time, and the Default
  583. Port VLAN ID can be configured for bridge devices using the new STP=,
  584. Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
  585. section of .netdev files.
  586. * The route table to which routes received over DHCP or RA should be
  587. added can be configured with the new RouteTable= option in the [DHCP]
  588. and [IPv6AcceptRA] sections of .network files.
  589. * The Address Resolution Protocol can be disabled on links managed by
  590. systemd-networkd using the ARP=no setting in the [Link] section of
  591. .network files.
  592. * New environment variables $SERVICE_RESULT, $EXIT_CODE and
  593. $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
  594. encode information about the result and exit codes of the current
  595. service runtime cycle.
  596. * systemd-sysctl will now configure kernel parameters in the order
  597. they occur in the configuration files. This matches what sysctl
  598. has been traditionally doing.
  599. * kernel-install "plugins" that are executed to perform various
  600. tasks after a new kernel is added and before an old one is removed
  601. can now return a special value to terminate the procedure and
  602. prevent any later plugins from running.
  603. * Journald's SplitMode=login setting has been deprecated. It has been
  604. removed from documentation, and its use is discouraged. In a future
  605. release it will be completely removed, and made equivalent to current
  606. default of SplitMode=uid.
  607. * Storage=both option setting in /etc/systemd/coredump.conf has been
  608. removed. With fast LZ4 compression storing the core dump twice is not
  609. useful.
  610. * The --share-system systemd-nspawn option has been replaced with an
  611. (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
  612. this functionality is discouraged. In addition the variables
  614. $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
  615. individual namespaces.
  616. * "machinectl list" now shows the IP address of running containers in
  617. the output, as well as OS release information.
  618. * "loginctl list" now shows the TTY of each session in the output.
  619. * sd-bus gained new API calls sd_bus_track_set_recursive(),
  620. sd_bus_track_get_recursive(), sd_bus_track_count_name(),
  621. sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
  622. tracking objects in a "recursive" mode, where a single client can be
  623. counted multiple times, if it takes multiple references.
  624. * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
  625. sd_bus_get_exit_on_disconnect(). They may be used to to make a
  626. process using sd-bus automatically exit if the bus connection is
  627. severed.
  628. * Bus clients of the service manager may now "pin" loaded units into
  629. memory, by taking an explicit reference on them. This is useful to
  630. ensure the client can retrieve runtime data about the service even
  631. after the service completed execution. Taking such a reference is
  632. available only for privileged clients and should be helpful to watch
  633. running services in a race-free manner, and in particular collect
  634. information about exit statuses and results.
  635. * The nss-resolve module has been changed to strictly return UNAVAIL
  636. when communication via D-Bus with resolved failed, and NOTFOUND when
  637. a lookup completed but was negative. This means it is now possible to
  638. neatly configure fallbacks using nsswitch.conf result checking
  639. expressions. Taking benefit of this, the new recommended
  640. configuration line for the "hosts" entry in /etc/nsswitch.conf is:
  641. hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
  642. * A new setting CtrlAltDelBurstAction= has been added to
  643. /etc/systemd/system.conf which may be used to configure the precise
  644. behaviour if the user on the console presses Ctrl-Alt-Del more often
  645. than 7 times in 2s. Previously this would unconditionally result in
  646. an expedited, immediate reboot. With this new setting the precise
  647. operation may be configured in more detail, and also turned off
  648. entirely.
  649. * In .netdev files two new settings RemoteChecksumTx= and
  650. RemoteChecksumRx= are now understood that permit configuring the
  651. remote checksumming logic for VXLAN networks.
  652. * The service manager learnt a new "invocation ID" concept for invoked
  653. services. Each runtime cycle of a service will get a new invocation
  654. ID (a 128bit random UUID) assigned that identifies the current
  655. run of the service uniquely and globally. A new invocation ID
  656. is generated each time a service starts up. The journal will store
  657. the invocation ID of a service along with any logged messages, thus
  658. making the invocation ID useful for matching the online runtime of a
  659. service with the offline log data it generated in a safe way without
  660. relying on synchronized timestamps. In many ways this new service
  661. invocation ID concept is similar to the kernel's boot ID concept that
  662. uniquely and globally identifies the runtime of each boot. The
  663. invocation ID of a service is passed to the service itself via an
  664. environment variable ($INVOCATION_ID). A new bus call
  665. GetUnitByInvocationID() has been added that is similar to GetUnit()
  666. but instead of retrieving the bus path for a unit by its name
  667. retrieves it by its invocation ID. The returned path is valid only as
  668. long as the passed invocation ID is current.
  669. * systemd-resolved gained a new "DNSStubListener" setting in
  670. resolved.conf. It either takes a boolean value or the special values
  671. "udp" and "tcp", and configures whether to enable the stub DNS
  672. listener on
  673. * IP addresses configured via networkd may now carry additional
  674. configuration settings supported by the kernel. New options include:
  675. HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
  676. PrefixRoute=, AutoJoin=.
  677. * The PAM configuration fragment file for "user@.service" shipped with
  678. systemd (i.e. the --user instance of systemd) has been stripped to
  679. the minimum necessary to make the system boot. Previously, it
  680. contained Fedora-specific stanzas that did not apply to other
  681. distributions. It is expected that downstream distributions add
  682. additional configuration lines, matching their needs to this file,
  683. using it only as rough template of what systemd itself needs. Note
  684. that this reduced fragment does not even include an invocation of
  685. pam_limits which most distributions probably want to add, even though
  686. systemd itself does not need it. (There's also the new build time
  687. option --with-pamconfdir=no to disable installation of the PAM
  688. fragment entirely.)
  689. * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
  690. capability is now also dropped from its set (in addition to
  691. CAP_SYS_MKNOD as before).
  692. * In service unit files it is now possible to connect a specific named
  693. file descriptor with stdin/stdout/stdout of an executed service. The
  694. name may be specified in matching .socket units using the
  695. FileDescriptorName= setting.
  696. * A number of journal settings may now be configured on the kernel
  697. command line. Specifically, the following options are now understood:
  698. systemd.journald.max_level_console=,
  699. systemd.journald.max_level_store=,
  700. systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
  701. systemd.journald.max_level_wall=.
  702. * "systemctl is-enabled --full" will now show by which symlinks a unit
  703. file is enabled in the unit dependency tree.
  704. * Support for VeraCrypt encrypted partitions has been added to the
  705. "cryptsetup" logic and /etc/crypttab.
  706. * systemd-detect-virt gained support for a new --private-users switch
  707. that checks whether the invoking processes are running inside a user
  708. namespace. Similar, a new special value "private-users" for the
  709. existing ConditionVirtualization= setting has been added, permitting
  710. skipping of specific units in user namespace environments.
  711. Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
  712. Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
  713. Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
  714. Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
  715. Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
  716. Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
  717. Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
  718. Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
  719. Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
  720. Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
  721. Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
  722. Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
  723. Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
  724. Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
  725. Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
  726. Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
  727. Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
  728. Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
  729. Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
  730. Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
  731. Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
  732. Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
  733. Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
  734. Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
  735. E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
  736. Jędrzejewski-Szmek, Zeal Jagannatha
  737. — Santa Fe, 2016-11-03
  738. CHANGES WITH 231:
  739. * In service units the various ExecXYZ= settings have been extended
  740. with an additional special character as first argument of the
  741. assigned value: if the character '+' is used the specified command
  742. line it will be run with full privileges, regardless of User=,
  743. Group=, CapabilityBoundingSet= and similar options. The effect is
  744. similar to the existing PermissionsStartOnly= option, but allows
  745. configuration of this concept for each executed command line
  746. independently.
  747. * Services may now alter the service watchdog timeout at runtime by
  748. sending a WATCHDOG_USEC= message via sd_notify().
  749. * MemoryLimit= and related unit settings now optionally take percentage
  750. specifications. The percentage is taken relative to the amount of
  751. physical memory in the system (or in case of containers, the assigned
  752. amount of memory). This allows scaling service resources neatly with
  753. the amount of RAM available on the system. Similarly, systemd-logind's
  754. RuntimeDirectorySize= option now also optionally takes percentage
  755. values.
  756. * In similar fashion TasksMax= takes percentage values now, too. The
  757. value is taken relative to the configured maximum number of processes
  758. on the system. The per-service task maximum has been changed to 15%
  759. using this functionality. (Effectively this is an increase of 512 →
  760. 4915 for service units, given the kernel's default pid_max setting.)
  761. * Calendar time specifications in .timer units now understand a ".."
  762. syntax for time ranges. Example: "4..7:10" may now be used for
  763. defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
  764. 7:10am every day.
  765. * The InaccessableDirectories=, ReadOnlyDirectories= and
  766. ReadWriteDirectories= unit file settings have been renamed to
  767. InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
  768. applied to all kinds of file nodes, and not just directories, with
  769. the exception of symlinks. Specifically these settings may now be
  770. used on block and character device nodes, UNIX sockets and FIFOS as
  771. well as regular files. The old names of these settings remain
  772. available for compatibility.
  773. * systemd will now log about all service processes it kills forcibly
  774. (using SIGKILL) because they remained after the clean shutdown phase
  775. of the service completed. This should help identifying services that
  776. shut down uncleanly. Moreover if KillUserProcesses= is enabled in
  777. systemd-logind's configuration a similar log message is generated for
  778. processes killed at the end of each session due to this setting.
  779. * systemd will now set the $JOURNAL_STREAM environment variable for all
  780. services whose stdout/stderr are connected to the Journal (which
  781. effectively means by default: all services). The variable contains
  782. the device and inode number of the file descriptor used for
  783. stdout/stderr. This may be used by invoked programs to detect whether
  784. their stdout/stderr is connected to the Journal, in which case they
  785. can switch over to direct Journal communication, thus being able to
  786. pass extended, structured metadata along with their log messages. As
  787. one example, this is now used by glib's logging primitives.
  788. * When using systemd's default tmp.mount unit for /tmp, the mount point
  789. will now be established with the "nosuid" and "nodev" options. This
  790. avoids privilege escalation attacks that put traps and exploits into
  791. /tmp. However, this might cause problems if you e. g. put container
  792. images or overlays into /tmp; if you need this, override tmp.mount's
  793. "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
  794. desired options.
  795. * systemd now supports the "memory" cgroup controller also on
  796. cgroupsv2.
  797. * The systemd-cgtop tool now optionally takes a control group path as
  798. command line argument. If specified, the control group list shown is
  799. limited to subgroups of that group.
  800. * The SystemCallFilter= unit file setting gained support for
  801. pre-defined, named system call filter sets. For example
  802. SystemCallFilter=@clock is now an effective way to make all clock
  803. changing-related system calls unavailable to a service. A number of
  804. similar pre-defined groups are defined. Writing system call filters
  805. for system services is simplified substantially with this new
  806. concept. Accordingly, all of systemd's own, long-running services now
  807. enable system call filtering based on this, by default.
  808. * A new service setting MemoryDenyWriteExecute= has been added, taking
  809. a boolean value. If turned on, a service may no longer create memory
  810. mappings that are writable and executable at the same time. This
  811. enhances security for services where this is enabled as it becomes
  812. harder to dynamically write and then execute memory in exploited
  813. service processes. This option has been enabled for all of systemd's
  814. own long-running services.
  815. * A new RestrictRealtime= service setting has been added, taking a
  816. boolean argument. If set the service's processes may no longer
  817. acquire realtime scheduling. This improves security as realtime
  818. scheduling may otherwise be used to easily freeze the system.
  819. * systemd-nspawn gained a new switch --notify-ready= taking a boolean
  820. value. This may be used for requesting that the system manager inside
  821. of the container reports start-up completion to nspawn which then
  822. propagates this notification further to the service manager
  823. supervising nspawn itself. A related option NotifyReady= in .nspawn
  824. files has been added too. This functionality allows ordering of the
  825. start-up of multiple containers using the usual systemd ordering
  826. primitives.
  827. * machinectl gained a new command "stop" that is an alias for
  828. "terminate".
  829. * systemd-resolved gained support for contacting DNS servers on
  830. link-local IPv6 addresses.
  831. * If systemd-resolved receives the SIGUSR2 signal it will now flush all
  832. its caches. A method call for requesting the same operation has been
  833. added to the bus API too, and is made available via "systemd-resolve
  834. --flush-caches".
  835. * systemd-resolve gained a new --status switch. If passed a brief
  836. summary of the used DNS configuration with per-interface information
  837. is shown.
  838. * resolved.conf gained a new Cache= boolean option, defaulting to
  839. on. If turned off local DNS caching is disabled. This comes with a
  840. performance penalty in particular when DNSSEC is enabled. Note that
  841. resolved disables its internal caching implicitly anyway, when the
  842. configured DNS server is on a host-local IP address such as ::1 or
  843., thus automatically avoiding double local caching.
  844. * systemd-resolved now listens on the local IP address
  845. for DNS requests. This improves compatibility with local programs
  846. that do not use the libc NSS or systemd-resolved's bus APIs for name
  847. resolution. This minimal DNS service is only available to local
  848. programs and does not implement the full DNS protocol, but enough to
  849. cover local DNS clients. A new, static resolv.conf file, listing just
  850. this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
  851. now recommended to make /etc/resolv.conf a symlink to this file in
  852. order to route all DNS lookups to systemd-resolved, regardless if
  853. done via NSS, the bus API or raw DNS packets. Note that this local
  854. DNS service is not as fully featured as the libc NSS or
  855. systemd-resolved's bus APIs. For example, as unicast DNS cannot be
  856. used to deliver link-local address information (as this implies
  857. sending a local interface index along), LLMNR/mDNS support via this
  858. interface is severely restricted. It is thus strongly recommended for
  859. all applications to use the libc NSS API or native systemd-resolved
  860. bus API instead.
  861. * systemd-networkd's bridge support learned a new setting
  862. VLANFiltering= for controlling VLAN filtering. Moreover a new section
  863. in .network files has been added for configuring VLAN bridging in
  864. more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
  865. * systemd-networkd's IPv6 Router Advertisement code now makes use of
  866. the DNSSL and RDNSS options. This means IPv6 DNS configuration may
  867. now be acquired without relying on DHCPv6. Two new options
  868. UseDomains= and UseDNS= have been added to configure this behaviour.
  869. * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
  870. renamed IPv6AcceptRA=, without altering its behaviour. The old
  871. setting name remains available for compatibility reasons.
  872. * The systemd-networkd VTI/VTI6 tunneling support gained new options
  873. Key=, InputKey= and OutputKey=.
  874. * systemd-networkd gained support for VRF ("Virtual Routing Function")
  875. interface configuration.
  876. * "systemctl edit" may now be used to create new unit files by
  877. specifying the --force switch.
  878. * sd-event gained a new function sd_event_get_iteration() for
  879. requesting the current iteration counter of the event loop. It starts
  880. at zero and is increased by one with each event loop iteration.
  881. * A new rpm macro %systemd_ordering is provided by the macros.systemd
  882. file. It can be used in lieu of %systemd_requires in packages which
  883. don't use any systemd functionality and are intended to be installed
  884. in minimal containers without systemd present. This macro provides
  885. ordering dependencies to ensure that if the package is installed in
  886. the same rpm transaction as systemd, systemd will be installed before
  887. the scriptlets for the package are executed, allowing unit presets
  888. to be handled.
  889. New macros %_systemdgeneratordir and %_systemdusergeneratordir have
  890. been added to simplify packaging of generators.
  891. * The os-release file gained VERSION_CODENAME field for the
  892. distribution nickname (e.g. VERSION_CODENAME=woody).
  894. can be set to disable parsing of metadata and the creation
  895. of persistent symlinks for that device.
  896. * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
  897. to make them available to logged-in users has been reverted.
  898. * Much of the common code of the various systemd components is now
  899. built into an internal shared library
  900. (incorporating the systemd version number in the name, to be updated
  901. with future releases) that the components link to. This should
  902. decrease systemd footprint both in memory during runtime and on
  903. disk. Note that the shared library is not for public use, and is
  904. neither API not ABI stable, but is likely to change with every new
  905. released update. Packagers need to make sure that binaries
  906. linking to are updated in step with the
  907. library.
  908. * Configuration for "mkosi" is now part of the systemd
  909. repository. mkosi is a tool to easily build legacy-free OS images,
  910. and is available on github: If
  911. "mkosi" is invoked in the build tree a new raw OS image is generated
  912. incorporating the systemd sources currently being worked on and a
  913. clean, fresh distribution installation. The generated OS image may be
  914. booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
  915. UEFI PC. This functionality is particularly useful to easily test
  916. local changes made to systemd in a pristine, defined environment. See
  917. HACKING for details.
  918. * configure learned the --with-support-url= option to specify the
  919. distribution's bugtracker.
  920. Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
  921. Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
  922. Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
  923. Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
  924. Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
  925. Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
  926. Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
  927. Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
  928. Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
  929. Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
  930. Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
  931. Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
  932. Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
  933. Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
  934. Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
  935. Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
  936. Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
  937. Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
  938. WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
  939. — Berlin, 2016-07-25
  940. CHANGES WITH 230:
  941. * DNSSEC is now turned on by default in systemd-resolved (in
  942. "allow-downgrade" mode), but may be turned off during compile time by
  943. passing "--with-default-dnssec=no" to "configure" (and of course,
  944. during runtime with DNSSEC= in resolved.conf). We recommend
  945. downstreams to leave this on at least during development cycles and
  946. report any issues with the DNSSEC logic upstream. We are very
  947. interested in collecting feedback about the DNSSEC validator and its
  948. limitations in the wild. Note however, that DNSSEC support is
  949. probably nothing downstreams should turn on in stable distros just
  950. yet, as it might create incompatibilities with a few DNS servers and
  951. networks. We tried hard to make sure we downgrade to non-DNSSEC mode
  952. automatically whenever we detect such incompatible setups, but there
  953. might be systems we do not cover yet. Hence: please help us testing
  954. the DNSSEC code, leave this on where you can, report back, but then
  955. again don't consider turning this on in your stable, LTS or
  956. production release just yet. (Note that you have to enable
  957. nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
  958. and its DNSSEC mode for host name resolution from local
  959. applications.)
  960. * systemd-resolve conveniently resolves DANE records with the --tlsa
  961. option and OPENPGPKEY records with the --openpgp option. It also
  962. supports dumping raw DNS record data via the new --raw= switch.
  963. * systemd-logind will now by default terminate user processes that are
  964. part of the user session scope unit (session-XX.scope) when the user
  965. logs out. This behavior is controlled by the KillUserProcesses=
  966. setting in logind.conf, and the previous default of "no" is now
  967. changed to "yes". This means that user sessions will be properly
  968. cleaned up after, but additional steps are necessary to allow
  969. intentionally long-running processes to survive logout.
  970. While the user is logged in at least once, user@.service is running,
  971. and any service that should survive the end of any individual login
  972. session can be started at a user service or scope using systemd-run.
  973. systemd-run(1) man page has been extended with an example which shows
  974. how to run screen in a scope unit underneath user@.service. The same
  975. command works for tmux.
  976. After the user logs out of all sessions, user@.service will be
  977. terminated too, by default, unless the user has "lingering" enabled.
  978. To effectively allow users to run long-term tasks even if they are
  979. logged out, lingering must be enabled for them. See loginctl(1) for
  980. details. The default polkit policy was modified to allow users to
  981. set lingering for themselves without authentication.
  982. Previous defaults can be restored at compile time by the
  983. --without-kill-user-processes option to "configure".
  984. * systemd-logind gained new configuration settings SessionsMax= and
  985. InhibitorsMax=, both with a default of 8192. It will not register new
  986. user sessions or inhibitors above this limit.
  987. * systemd-logind will now reload configuration on SIGHUP.
  988. * The unified cgroup hierarchy added in Linux 4.5 is now supported.
  989. Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
  990. enable. Also, support for the "io" cgroup controller in the unified
  991. hierarchy has been added, so that the "memory", "pids" and "io" are
  992. now the controllers that are supported on the unified hierarchy.
  993. WARNING: it is not possible to use previous systemd versions with
  994. systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
  995. is necessary to also update systemd in the initramfs if using the
  996. unified hierarchy. An updated SELinux policy is also required.
  997. * LLDP support has been extended, and both passive (receive-only) and
  998. active (sender) modes are supported. Passive mode ("routers-only") is
  999. enabled by default in systemd-networkd. Active LLDP mode is enabled
  1000. by default for containers on the internal network. The "networkctl
  1001. lldp" command may be used to list information gathered. "networkctl
  1002. status" will also show basic LLDP information on connected peers now.
  1003. * The IAID and DUID unique identifier sent in DHCP requests may now be
  1004. configured for the system and each .network file managed by
  1005. systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
  1006. * systemd-networkd gained support for configuring proxy ARP support for
  1007. each interface, via the ProxyArp= setting in .network files. It also
  1008. gained support for configuring the multicast querier feature of
  1009. bridge devices, via the new MulticastQuerier= setting in .netdev
  1010. files. Similarly, snooping on the IGMP traffic can be controlled
  1011. via the new setting MulticastSnooping=.
  1012. A new setting PreferredLifetime= has been added for addresses
  1013. configured in .network file to configure the lifetime intended for an
  1014. address.
  1015. The systemd-networkd DHCP server gained the option EmitRouter=, which
  1016. defaults to yes, to configure whether the DHCP Option 3 (Router)
  1017. should be emitted.
  1018. * The testing tool /usr/lib/systemd/systemd-activate is renamed to
  1019. systemd-socket-activate and installed into /usr/bin. It is now fully
  1020. supported.
  1021. * systemd-journald now uses separate threads to flush changes to disk
  1022. when closing journal files, thus reducing impact of slow disk I/O on
  1023. logging performance.
  1024. * The sd-journal API gained two new calls
  1025. sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
  1026. can be used to open journal files using file descriptors instead of
  1027. file or directory paths. sd_journal_open_container() has been
  1028. deprecated, sd_journal_open_directory_fd() should be used instead
  1029. with the flag SD_JOURNAL_OS_ROOT.
  1030. * journalctl learned a new output mode "-o short-unix" that outputs log
  1031. lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
  1032. UTC). It also gained support for a new --no-hostname setting to
  1033. suppress the hostname column in the family of "short" output modes.
  1034. * systemd-ask-password now optionally skips printing of the password to
  1035. stdout with --no-output which can be useful in scripts.
  1036. * Framebuffer devices (/dev/fb*) and 3D printers and scanners
  1037. (devices tagged with ID_MAKER_TOOL) are now tagged with
  1038. "uaccess" and are available to logged in users.
  1039. * The DeviceAllow= unit setting now supports specifiers (with "%").
  1040. * "systemctl show" gained a new --value switch, which allows print a
  1041. only the contents of a specific unit property, without also printing
  1042. the property's name. Similar support was added to "show*" verbs
  1043. of loginctl and machinectl that output "key=value" lists.
  1044. * A new unit type "generated" was added for files dynamically generated
  1045. by generator tools. Similarly, a new unit type "transient" is used
  1046. for unit files created using the runtime API. "systemctl enable" will
  1047. refuse to operate on such files.
  1048. * A new command "systemctl revert" has been added that may be used to
  1049. revert to the vendor version of a unit file, in case local changes
  1050. have been made by adding drop-ins or overriding the unit file.
  1051. * "machinectl clean" gained a new verb to automatically remove all or
  1052. just hidden container images.
  1053. * systemd-tmpfiles gained support for a new line type "e" for emptying
  1054. directories, if they exist, without creating them if they don't.
  1055. * systemd-nspawn gained support for automatically patching the UID/GIDs
  1056. of the owners and the ACLs of all files and directories in a
  1057. container tree to match the UID/GID user namespacing range selected
  1058. for the container invocation. This mode is enabled via the new
  1059. --private-users-chown switch. It also gained support for
  1060. automatically choosing a free, previously unused UID/GID range when
  1061. starting a container, via the new --private-users=pick setting (which
  1062. implies --private-users-chown). Together, these options for the first
  1063. time make user namespacing for nspawn containers fully automatic and
  1064. thus deployable. The systemd-nspawn@.service template unit file has
  1065. been changed to use this functionality by default.
  1066. * systemd-nspawn gained a new --network-zone= switch, that allows
  1067. creating ad-hoc virtual Ethernet links between multiple containers,
  1068. that only exist as long as at least one container referencing them is
  1069. running. This allows easy connecting of multiple containers with a
  1070. common link that implements an Ethernet broadcast domain. Each of
  1071. these network "zones" may be named relatively freely by the user, and
  1072. may be referenced by any number of containers, but each container may
  1073. only reference one of these "zones". On the lower level, this is
  1074. implemented by an automatically managed bridge network interface for
  1075. each zone, that is created when the first container referencing its
  1076. zone is created and removed when the last one referencing its zone
  1077. terminates.
  1078. * The default start timeout may now be configured on the kernel command
  1079. line via systemd.default_timeout_start_sec=. It was already
  1080. configurable via the DefaultTimeoutStartSec= option in
  1081. /etc/systemd/system.conf.
  1082. * Socket units gained a new TriggerLimitIntervalSec= and
  1083. TriggerLimitBurst= setting to configure a limit on the activation
  1084. rate of the socket unit.
  1085. * The LimitNICE= setting now optionally takes normal UNIX nice values
  1086. in addition to the raw integer limit value. If the specified
  1087. parameter is prefixed with "+" or "-" and is in the range -20..19 the
  1088. value is understood as UNIX nice value. If not prefixed like this it
  1089. is understood as raw RLIMIT_NICE limit.
  1090. * Note that the effect of the PrivateDevices= unit file setting changed
  1091. slightly with this release: the per-device /dev file system will be
  1092. mounted read-only from this version on, and will have "noexec"
  1093. set. This (minor) change of behavior might cause some (exceptional)
  1094. legacy software to break, when PrivateDevices=yes is set for its
  1095. service. Please leave PrivateDevices= off if you run into problems
  1096. with this.
  1097. * systemd-bootchart has been split out to a separate repository:
  1099. * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
  1100. merged into the kernel in its current form.
  1101. * The compatibility libraries,
  1102.,, and
  1103. which have been deprecated since systemd-209 have been removed along
  1104. with the corresponding pkg-config files. All symbols provided by
  1105. those libraries are provided by
  1106. * The Capabilities= unit file setting has been removed (it is ignored
  1107. for backwards compatibility). AmbientCapabilities= and
  1108. CapabilityBoundingSet= should be used instead.
  1109. * A new special target has been added,,
  1110. which creates a synchronization point for dependencies of the root
  1111. device in early userspace. Initramfs builders must ensure that this
  1112. target is now included in early userspace.
  1113. Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
  1114. Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
  1115. Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
  1116. Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
  1117. Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
  1118. R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
  1119. Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
  1120. Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
  1121. Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
  1122. Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
  1123. John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
  1124. Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
  1125. Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
  1126. Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
  1127. Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
  1128. mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
  1129. Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
  1130. Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
  1131. Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
  1132. Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
  1133. Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
  1134. Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
  1135. Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
  1136. Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
  1137. Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
  1138. Jędrzejewski-Szmek
  1139. — Fairfax, 2016-05-21
  1140. CHANGES WITH 229:
  1141. * The systemd-resolved DNS resolver service has gained a substantial
  1142. set of new features, most prominently it may now act as a DNSSEC
  1143. validating stub resolver. DNSSEC mode is currently turned off by
  1144. default, but is expected to be turned on by default in one of the
  1145. next releases. For now, we invite everybody to test the DNSSEC logic
  1146. by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
  1147. service also gained a full set of D-Bus interfaces, including calls
  1148. to configure DNS and DNSSEC settings per link (for use by external
  1149. network management software). systemd-resolved and systemd-networkd
  1150. now distinguish between "search" and "routing" domains. The former
  1151. are used to qualify single-label names, the latter are used purely
  1152. for routing lookups within certain domains to specific links.
  1153. resolved now also synthesizes RRs for all entries from /etc/hosts.
  1154. * The systemd-resolve tool (which is a client utility for
  1155. systemd-resolved) has been improved considerably and is now fully
  1156. supported and documented. Hence it has moved from /usr/lib/systemd to
  1157. /usr/bin.
  1158. * /dev/disk/by-path/ symlink support has been (re-)added for virtio
  1159. devices.
  1160. * The coredump collection logic has been reworked: when a coredump is
  1161. collected it is now written to disk, compressed and processed
  1162. (including stacktrace extraction) from a new instantiated service
  1163. systemd-coredump@.service, instead of directly from the
  1164. /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
  1165. processing large coredumps can take up a substantial amount of
  1166. resources and time, and this previously happened entirely outside of
  1167. systemd's service supervision. With the new logic the core_pattern
  1168. hook only does minimal metadata collection before passing off control
  1169. to the new instantiated service, which is configured with a time
  1170. limit, a nice level and other settings to minimize negative impact on
  1171. the rest of the system. Also note that the new logic will honour the
  1172. RLIMIT_CORE setting of the crashed process, which now allows users
  1173. and processes to turn off coredumping for their processes by setting
  1174. this limit.
  1175. * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
  1176. and all forked processes by default. Previously, PID 1 would leave
  1177. the setting at "0" for all processes, as set by the kernel. Note that
  1178. the resource limit traditionally has no effect on the generated
  1179. coredumps on the system if the /proc/sys/kernel/core_pattern hook
  1180. logic is used. Since the limit is now honoured (see above) its
  1181. default has been changed so that the coredumping logic is enabled by
  1182. default for all processes, while allowing specific opt-out.
  1183. * When the stacktrace is extracted from processes of system users, this
  1184. is now done as "systemd-coredump" user, in order to sandbox this
  1185. potentially security sensitive parsing operation. (Note that when
  1186. processing coredumps of normal users this is done under the user ID
  1187. of process that crashed, as before.) Packagers should take notice
  1188. that it is now necessary to create the "systemd-coredump" system user
  1189. and group at package installation time.
  1190. * The systemd-activate socket activation testing tool gained support
  1191. for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
  1192. and --seqpacket switches. It also has been extended to support both
  1193. new-style and inetd-style file descriptor passing. Use the new
  1194. --inetd switch to request inetd-style file descriptor passing.
  1195. * Most systemd tools now honor a new $SYSTEMD_COLORS environment
  1196. variable, which takes a boolean value. If set to false, ANSI color
  1197. output is disabled in the tools even when run on a terminal that
  1198. supports it.
  1199. * The VXLAN support in networkd now supports two new settings
  1200. DestinationPort= and PortRange=.
  1201. * A new systemd.machine_id= kernel command line switch has been added,
  1202. that may be used to set the machine ID in /etc/machine-id if it is
  1203. not initialized yet. This command line option has no effect if the
  1204. file is already initialized.
  1205. * systemd-nspawn gained a new --as-pid2 switch that invokes any
  1206. specified command line as PID 2 rather than PID 1 in the
  1207. container. In this mode PID 1 is a minimal stub init process that
  1208. implements the special POSIX and Linux semantics of PID 1 regarding
  1209. signal and child process management. Note that this stub init process
  1210. is implemented in nspawn itself and requires no support from the
  1211. container image. This new logic is useful to support running
  1212. arbitrary commands in the container, as normal processes are
  1213. generally not prepared to run as PID 1.
  1214. * systemd-nspawn gained a new --chdir= switch for setting the current
  1215. working directory for the process started in the container.
  1216. * "journalctl /dev/sda" will now output all kernel log messages for
  1217. specified device from the current boot, in addition to all devices
  1218. that are parents of it. This should make log output about devices
  1219. pretty useful, as long as kernel drivers attach enough metadata to
  1220. the log messages. (The usual SATA drivers do.)
  1221. * The sd-journal API gained two new calls
  1222. sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
  1223. that report whether log data from /run or /var has been found.
  1224. * journalctl gained a new switch "--fields" that prints all journal
  1225. record field names currently in use in the journal. This is backed
  1226. by two new sd-journal API calls sd_journal_enumerate_fields() and
  1227. sd_journal_restart_fields().
  1228. * Most configurable timeouts in systemd now expect an argument of
  1229. "infinity" to turn them off, instead of "0" as before. The semantics
  1230. from now on is that a timeout of "0" means "now", and "infinity"
  1231. means "never". To maintain backwards compatibility, "0" continues to
  1232. turn off previously existing timeout settings.
  1233. * "systemctl reload-or-try-restart" has been renamed to "systemctl
  1234. try-reload-or-restart" to clarify what it actually does: the "try"
  1235. logic applies to both reloading and restarting, not just restarting.
  1236. The old name continues to be accepted for compatibility.
  1237. * On boot-up, when PID 1 detects that the system clock is behind the
  1238. release date of the systemd version in use, the clock is now set
  1239. to the latter. Previously, this was already done in timesyncd, in order
  1240. to avoid running with clocks set to the various clock epochs such as
  1241. 1902, 1938 or 1970. With this change the logic is now done in PID 1
  1242. in addition to timesyncd during early boot-up, so that it is enforced
  1243. before the first process is spawned by systemd. Note that the logic
  1244. in timesyncd remains, as it is more comprehensive and ensures
  1245. clock monotonicity by maintaining a persistent timestamp file in
  1246. /var. Since /var is generally not available in earliest boot or the
  1247. initrd, this part of the logic remains in timesyncd, and is not done
  1248. by PID 1.
  1249. * Support for tweaking details in net_cls.class_id through the
  1250. NetClass= configuration directive has been removed, as the kernel
  1251. people have decided to deprecate that controller in cgroup v2.
  1252. Userspace tools such as nftables are moving over to setting rules
  1253. that are specific to the full cgroup path of a task, which obsoletes
  1254. these controllers anyway. The NetClass= directive is kept around for
  1255. legacy compatibility reasons. For a more in-depth description of the
  1256. kernel change, please refer to the respective upstream commit:
  1258. * A new service setting RuntimeMaxSec= has been added that may be used
  1259. to specify a maximum runtime for a service. If the timeout is hit, the
  1260. service is terminated and put into a failure state.
  1261. * A new service setting AmbientCapabilities= has been added. It allows
  1262. configuration of additional Linux process capabilities that are
  1263. passed to the activated processes. This is only available on very
  1264. recent kernels.
  1265. * The process resource limit settings in service units may now be used
  1266. to configure hard and soft limits individually.
  1267. * The various libsystemd APIs such as sd-bus or sd-event now publicly
  1268. expose support for gcc's __attribute__((cleanup())) C extension.
  1269. Specifically, for many object destructor functions alternative
  1270. versions have been added that have names suffixed with "p" and take a
  1271. pointer to a pointer to the object to destroy, instead of just a
  1272. pointer to the object itself. This is useful because these destructor
  1273. functions may be used directly as parameters to the cleanup
  1274. construct. Internally, systemd has been a heavy user of this GCC
  1275. extension for a long time, and with this change similar support is
  1276. now available to consumers of the library outside of systemd. Note
  1277. that by using this extension in your sources compatibility with old
  1278. and strictly ANSI compatible C compilers is lost. However, all gcc or
  1279. LLVM versions of recent years support this extension.
  1280. * Timer units gained support for a new setting RandomizedDelaySec= that
  1281. allows configuring some additional randomized delay to the configured
  1282. time. This is useful to spread out timer events to avoid load peaks in
  1283. clusters or larger setups.
  1284. * Calendar time specifications now support sub-second accuracy.
  1285. * Socket units now support listening on SCTP and UDP-lite protocol
  1286. sockets.
  1287. * The sd-event API now comes with a full set of man pages.
  1288. * Older versions of systemd contained experimental support for
  1289. compressing journal files and coredumps with the LZ4 compressor that
  1290. was not compatible with the lz4 binary (due to API limitations of the
  1291. lz4 library). This support has been removed; only support for files
  1292. compatible with the lz4 binary remains. This LZ4 logic is now
  1293. officially supported and no longer considered experimental.
  1294. * The dkr image import logic has been removed again from importd. dkr's
  1295. micro-services focus doesn't fit into the machine image focus of
  1296. importd, and quickly got out of date with the upstream dkr API.
  1297. * Creation of the /run/lock/lockdev/ directory was dropped from
  1298. tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
  1299. been available for many years. If you still need this, you need to
  1300. create your own tmpfiles.d config file with:
  1301. d /run/lock/lockdev 0775 root lock -
  1302. Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
  1303. Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
  1304. Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
  1305. Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
  1306. Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
  1307. David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
  1308. Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
  1309. Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
  1310. Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
  1311. Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
  1312. Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
  1313. lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
  1314. Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
  1315. Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
  1316. Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
  1317. Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
  1318. Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
  1319. Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
  1320. Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
  1321. — Berlin, 2016-02-11
  1322. CHANGES WITH 228:
  1323. * A number of properties previously only settable in unit
  1324. files are now also available as properties to set when
  1325. creating transient units programmatically via the bus, as it
  1326. is exposed with systemd-run's --property=
  1327. setting. Specifically, these are: SyslogIdentifier=,
  1328. SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
  1329. EnvironmentFile=, ReadWriteDirectories=,
  1330. ReadOnlyDirectories=, InaccessibleDirectories=,
  1331. ProtectSystem=, ProtectHome=, RuntimeDirectory=.
  1332. * When creating transient services via the bus API it is now
  1333. possible to pass in a set of file descriptors to use as
  1334. STDIN/STDOUT/STDERR for the invoked process.
  1335. * Slice units may now be created transiently via the bus APIs,
  1336. similar to the way service and scope units may already be
  1337. created transiently.
  1338. * Wherever systemd expects a calendar timestamp specification
  1339. (like in journalctl's --since= and --until= switches) UTC
  1340. timestamps are now supported. Timestamps suffixed with "UTC"
  1341. are now considered to be in Universal Time Coordinated
  1342. instead of the local timezone. Also, timestamps may now
  1343. optionally be specified with sub-second accuracy. Both of
  1344. these additions also apply to recurring calendar event
  1345. specification, such as OnCalendar= in timer units.
  1346. * journalctl gained a new "--sync" switch that asks the
  1347. journal daemon to write all so far unwritten log messages to
  1348. disk and sync the files, before returning.
  1349. * systemd-tmpfiles learned two new line types "q" and "Q" that
  1350. operate like "v", but also set up a basic btrfs quota
  1351. hierarchy when used on a btrfs file system with quota
  1352. enabled.
  1353. * tmpfiles' "v", "q" and "Q" will now create a plain directory
  1354. instead of a subvolume (even on a btrfs file system) if the
  1355. root directory is a plain directory, and not a
  1356. subvolume. This should simplify things with certain chroot()
  1357. environments which are not aware of the concept of btrfs
  1358. subvolumes.
  1359. * systemd-detect-virt gained a new --chroot switch to detect
  1360. whether execution takes place in a chroot() environment.
  1361. * CPUAffinity= now takes CPU index ranges in addition to
  1362. individual indexes.
  1363. * The various memory-related resource limit settings (such as
  1364. LimitAS=) now understand the usual K, M, G, ... suffixes to
  1365. the base of 1024 (IEC). Similar, the time-related resource
  1366. limit settings understand the usual min, h, day, ...
  1367. suffixes now.
  1368. * There's a new system.conf setting DefaultTasksMax= to
  1369. control the default TasksMax= setting for services and
  1370. scopes running on the system. (TasksMax= is the primary
  1371. setting that exposes the "pids" cgroup controller on systemd
  1372. and was introduced in the previous systemd release.) The
  1373. setting now defaults to 512, which means services that are
  1374. not explicitly configured otherwise will only be able to
  1375. create 512 processes or threads at maximum, from this
  1376. version on. Note that this means that thread- or
  1377. process-heavy services might need to be reconfigured to set
  1378. TasksMax= to a higher value. It is sufficient to set
  1379. TasksMax= in these specific unit files to a higher value, or
  1380. even "infinity". Similar, there's now a logind.conf setting
  1381. UserTasksMax= that defaults to 4096 and limits the total
  1382. number of processes or tasks each user may own
  1383. concurrently. nspawn containers also have the TasksMax=
  1384. value set by default now, to 8192. Note that all of this
  1385. only has an effect if the "pids" cgroup controller is
  1386. enabled in the kernel. The general benefit of these changes
  1387. should be a more robust and safer system, that provides a
  1388. certain amount of per-service fork() bomb protection.
  1389. * systemd-nspawn gained the new --network-veth-extra= switch
  1390. to define additional and arbitrarily-named virtual Ethernet
  1391. links between the host and the container.
  1392. * A new service execution setting PassEnvironment= has been
  1393. added that allows importing select environment variables
  1394. from PID1's environment block into the environment block of
  1395. the service.
  1396. * Timer units gained support for a new RemainAfterElapse=
  1397. setting which takes a boolean argument. It defaults to on,
  1398. exposing behaviour unchanged to previous releases. If set to
  1399. off, timer units are unloaded after they elapsed if they
  1400. cannot elapse again. This is particularly useful for
  1401. transient timer units, which shall not stay around longer
  1402. than until they first elapse.
  1403. * systemd will now bump the net.unix.max_dgram_qlen to 512 by
  1404. default now (the kernel default is 16). This is beneficial
  1405. for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
  1406. allows substantially larger numbers of queued
  1407. datagrams. This should increase the capability of systemd to
  1408. parallelize boot-up, as logging and sd_notify() are unlikely
  1409. to stall execution anymore. If you need to change the value
  1410. from the new defaults, use the usual sysctl.d/ snippets.
  1411. * The compression framing format used by the journal or
  1412. coredump processing has changed to be in line with what the
  1413. official LZ4 tools generate. LZ4 compression support in
  1414. systemd was considered unsupported previously, as the format
  1415. was not compatible with the normal tools. With this release
  1416. this has changed now, and it is hence safe for downstream
  1417. distributions to turn it on. While not compressing as well
  1418. as the XZ, LZ4 is substantially faster, which makes
  1419. it a good default choice for the compression logic in the
  1420. journal and in coredump handling.
  1421. * Any reference to /etc/mtab has been dropped from
  1422. systemd. The file has been obsolete since a while, but
  1423. systemd refused to work on systems where it was incorrectly
  1424. set up (it should be a symlink or non-existent). Please make
  1425. sure to update to util-linux 2.27.1 or newer in conjunction
  1426. with this systemd release, which also drops any reference to
  1427. /etc/mtab. If you maintain a distribution make sure that no
  1428. software you package still references it, as this is a
  1429. likely source of bugs. There's also a glibc bug pending,
  1430. asking for removal of any reference to this obsolete file:
  1432. Note that only util-linux versions built with
  1433. --enable-libmount-force-mountinfo are supported.
  1434. * Support for the ".snapshot" unit type has been removed. This
  1435. feature turned out to be little useful and little used, and
  1436. has now been removed from the core and from systemctl.
  1437. * The dependency types RequiresOverridable= and
  1438. RequisiteOverridable= have been removed from systemd. They
  1439. have been used only very sparingly to our knowledge and
  1440. other options that provide a similar effect (such as
  1441. systemctl --mode=ignore-dependencies) are much more useful
  1442. and commonly used. Moreover, they were only half-way
  1443. implemented as the option to control behaviour regarding
  1444. these dependencies was never added to systemctl. By removing
  1445. these dependency types the execution engine becomes a bit
  1446. simpler. Unit files that use these dependencies should be
  1447. changed to use the non-Overridable dependency types
  1448. instead. In fact, when parsing unit files with these
  1449. options, that's what systemd will automatically convert them
  1450. too, but it will also warn, asking users to fix the unit
  1451. files accordingly. Removal of these dependency types should
  1452. only affect a negligible number of unit files in the wild.
  1453. * Behaviour of networkd's IPForward= option changed
  1454. (again). It will no longer maintain a per-interface setting,
  1455. but propagate one way from interfaces where this is enabled
  1456. to the global kernel setting. The global setting will be
  1457. enabled when requested by a network that is set up, but
  1458. never be disabled again. This change was made to make sure
  1459. IPv4 and IPv6 behaviour regarding packet forwarding is
  1460. similar (as the Linux IPv6 stack does not support
  1461. per-interface control of this setting) and to minimize
  1462. surprises.
  1463. * In unit files the behaviour of %u, %U, %h, %s has
  1464. changed. These specifiers will now unconditionally resolve
  1465. to the various user database fields of the user that the
  1466. systemd instance is running as, instead of the user
  1467. configured in the specific unit via User=. Note that this
  1468. effectively doesn't change much, as resolving of these
  1469. specifiers was already turned off in the --system instance
  1470. of systemd, as we cannot do NSS lookups from PID 1. In the
  1471. --user instance of systemd these specifiers where correctly
  1472. resolved, but hardly made any sense, since the user instance
  1473. lacks privileges to do user switches anyway, and User= is
  1474. hence useless. Moreover, even in the --user instance of
  1475. systemd behaviour was awkward as it would only take settings
  1476. from User= assignment placed before the specifier into
  1477. account. In order to unify and simplify the logic around
  1478. this the specifiers will now always resolve to the
  1479. credentials of the user invoking the manager (which in case
  1480. of PID 1 is the root user).
  1481. Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
  1482. Yang, Daniel Machon, Daniel Mack, David Herrmann, David
  1483. Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
  1484. Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
  1485. Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
  1486. Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
  1487. Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
  1488. Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
  1489. Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
  1490. Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
  1491. Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
  1492. Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
  1493. Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
  1494. Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
  1495. Jędrzejewski-Szmek
  1496. — Berlin, 2015-11-18
  1497. CHANGES WITH 227:
  1498. * systemd now depends on util-linux v2.27. More specifically,
  1499. the newly added mount monitor feature in libmount now
  1500. replaces systemd's former own implementation.
  1501. * libmount mandates /etc/mtab not to be regular file, and
  1502. systemd now enforces this condition at early boot.
  1503. /etc/mtab has been deprecated and warned about for a very
  1504. long time, so systems running systemd should already have
  1505. stopped having this file around as anything else than a
  1506. symlink to /proc/self/mounts.
  1507. * Support for the "pids" cgroup controller has been added. It
  1508. allows accounting the number of tasks in a cgroup and
  1509. enforcing limits on it. This adds two new setting
  1510. TasksAccounting= and TasksMax= to each unit, as well as a
  1511. global option DefaultTasksAccounting=.
  1512. * Support for the "net_cls" cgroup controller has been added.
  1513. It allows assigning a net class ID to each task in the
  1514. cgroup, which can then be used in firewall rules and traffic
  1515. shaping configurations. Note that the kernel netfilter net
  1516. class code does not currently work reliably for ingress
  1517. packets on unestablished sockets.
  1518. This adds a new config directive called NetClass= to CGroup
  1519. enabled units. Allowed values are positive numbers for fixed
  1520. assignments and "auto" for picking a free value
  1521. automatically.
  1522. * 'systemctl is-system-running' now returns 'offline' if the
  1523. system is not booted with systemd. This command can now be
  1524. used as a substitute for 'systemd-notify --booted'.
  1525. * Watchdog timeouts have been increased to 3 minutes for all
  1526. in-tree service files. Apparently, disk IO issues are more
  1527. frequent than we hoped, and user reported >1 minute waiting
  1528. for disk IO.
  1529. * 'machine-id-commit' functionality has been merged into
  1530. 'machine-id-setup --commit'. The separate binary has been
  1531. removed.
  1532. * The WorkingDirectory= directive in unit files may now be set
  1533. to the special value '~'. In this case, the working
  1534. directory is set to the home directory of the user
  1535. configured in User=.
  1536. * "machinectl shell" will now open the shell in the home
  1537. directory of the selected user by default.
  1538. * The CrashChVT= configuration file setting is renamed to
  1539. CrashChangeVT=, following our usual logic of not
  1540. abbreviating unnecessarily. The old directive is still
  1541. supported for compat reasons. Also, this directive now takes
  1542. an integer value between 1 and 63, or a boolean value. The
  1543. formerly supported '-1' value for disabling stays around for
  1544. compat reasons.
  1545. * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
  1546. NoNewPrivileges=, TTYPath=, WorkingDirectory= and
  1547. RootDirectory= properties can now be set for transient
  1548. units.
  1549. * The systemd-analyze tool gained a new "set-log-target" verb
  1550. to change the logging target the system manager logs to
  1551. dynamically during runtime. This is similar to how
  1552. "systemd-analyze set-log-level" already changes the log
  1553. level.
  1554. * In nspawn /sys is now mounted as tmpfs, with only a selected
  1555. set of subdirectories mounted in from the real sysfs. This
  1556. enhances security slightly, and is useful for ensuring user
  1557. namespaces work correctly.
  1558. * Support for USB FunctionFS activation has been added. This
  1559. allows implementation of USB gadget services that are
  1560. activated as soon as they are requested, so that they don't
  1561. have to run continuously, similar to classic socket
  1562. activation.
  1563. * The "systemctl exit" command now optionally takes an
  1564. additional parameter that sets the exit code to return from
  1565. the systemd manager when exiting. This is only relevant when
  1566. running the systemd user instance, or when running the
  1567. system instance in a container.
  1568. * sd-bus gained the new API calls sd_bus_path_encode_many()
  1569. and sd_bus_path_decode_many() that allow easy encoding and
  1570. decoding of multiple identifier strings inside a D-Bus
  1571. object path. Another new call sd_bus_default_flush_close()
  1572. has been added to flush and close per-thread default
  1573. connections.
  1574. * systemd-cgtop gained support for a -M/--machine= switch to
  1575. show the control groups within a certain container only.
  1576. * "systemctl kill" gained support for an optional --fail
  1577. switch. If specified the requested operation will fail of no
  1578. processes have been killed, because the unit had no
  1579. processes attached, or similar.
  1580. * A new systemd.crash_reboot=1 kernel command line option has
  1581. been added that triggers a reboot after crashing. This can
  1582. also be set through CrashReboot= in systemd.conf.
  1583. * The RuntimeDirectory= setting now understands unit
  1584. specifiers like %i or %f.
  1585. * A new (still internal) library API sd-ipv4acd has been added,
  1586. that implements address conflict detection for IPv4. It's
  1587. based on code from sd-ipv4ll, and will be useful for
  1588. detecting DHCP address conflicts.
  1589. * File descriptors passed during socket activation may now be
  1590. named. A new API sd_listen_fds_with_names() is added to
  1591. access the names. The default names may be overridden,
  1592. either in the .socket file using the FileDescriptorName=
  1593. parameter, or by passing FDNAME= when storing the file
  1594. descriptors using sd_notify().
  1595. * systemd-networkd gained support for:
  1596. - Setting the IPv6 Router Advertisement settings via
  1597. IPv6AcceptRouterAdvertisements= in .network files.
  1598. - Configuring the HelloTimeSec=, MaxAgeSec= and
  1599. ForwardDelaySec= bridge parameters in .netdev files.
  1600. - Configuring PreferredSource= for static routes in
  1601. .network files.
  1602. * The "ask-password" framework used to query for LUKS harddisk
  1603. passwords or SSL passwords during boot gained support for
  1604. caching passwords in the kernel keyring, if it is
  1605. available. This makes sure that the user only has to type in
  1606. a passphrase once if there are multiple objects to unlock
  1607. with the same one. Previously, such password caching was
  1608. available only when Plymouth was used; this moves the
  1609. caching logic into the systemd codebase itself. The
  1610. "systemd-ask-password" utility gained a new --keyname=
  1611. switch to control which kernel keyring key to use for
  1612. caching a password in. This functionality is also useful for
  1613. enabling display managers such as gdm to automatically
  1614. unlock the user's GNOME keyring if its passphrase, the
  1615. user's password and the harddisk password are the same, if
  1616. gdm-autologin is used.
  1617. * When downloading tar or raw images using "machinectl
  1618. pull-tar" or "machinectl pull-raw", a matching ".nspawn"
  1619. file is now also downloaded, if it is available and stored
  1620. next to the image file.
  1621. * Units of type ".socket" gained a new boolean setting
  1622. Writable= which is only useful in conjunction with
  1623. ListenSpecial=. If true, enables opening the specified
  1624. special file in O_RDWR mode rather than O_RDONLY mode.
  1625. * systemd-rfkill has been reworked to become a singleton
  1626. service that is activated through /dev/rfkill on each rfkill
  1627. state change and saves the settings to disk. This way,
  1628. systemd-rfkill is now compatible with devices that exist
  1629. only intermittendly, and even restores state if the previous
  1630. system shutdown was abrupt rather than clean.
  1631. * The journal daemon gained support for vacuuming old journal
  1632. files controlled by the number of files that shall remain,
  1633. in addition to the already existing control by size and by
  1634. date. This is useful as journal interleaving performance
  1635. degrades with too many separate journal files, and allows
  1636. putting an effective limit on them. The new setting defaults
  1637. to 100, but this may be changed by setting SystemMaxFiles=
  1638. and RuntimeMaxFiles= in journald.conf. Also, the
  1639. "journalctl" tool gained the new --vacuum-files= switch to
  1640. manually vacuum journal files to leave only the specified
  1641. number of files in place.
  1642. * udev will now create /dev/disk/by-path links for ATA devices
  1643. on kernels where that is supported.
  1644. * Galician, Serbian, Turkish and Korean translations were added.
  1645. Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
  1646. Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
  1647. (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
  1648. Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
  1649. Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
  1650. de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
  1651. Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
  1652. Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
  1653. Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
  1654. Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
  1655. Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
  1656. Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
  1657. Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
  1658. Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
  1659. Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
  1660. Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
  1661. Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
  1662. Zbigniew Jędrzejewski-Szmek, Марко М. Костић
  1663. — Berlin, 2015-10-07
  1664. CHANGES WITH 226:
  1665. * The DHCP implementation of systemd-networkd gained a set of
  1666. new features:
  1667. - The DHCP server now supports emitting DNS and NTP
  1668. information. It may be enabled and configured via
  1669. EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
  1670. and NTP information is enabled, but no servers are
  1671. configured, the corresponding uplink information (if there
  1672. is any) is propagated.
  1673. - Server and client now support transmission and reception
  1674. of timezone information. It can be configured via the
  1675. newly introduced network options UseTimezone=,
  1676. EmitTimezone=, and Timezone=. Transmission of timezone
  1677. information is enabled between host and containers by
  1678. default now: the container will change its local timezone
  1679. to what the host has set.
  1680. - Lease timeouts can now be configured via
  1681. MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
  1682. - The DHCP server improved on the stability of
  1683. leases. Clients are more likely to get the same lease
  1684. information back, even if the server loses state.
  1685. - The DHCP server supports two new configuration options to
  1686. control the lease address pool metrics, PoolOffset= and
  1687. PoolSize=.
  1688. * The encapsulation limit of tunnels in systemd-networkd may
  1689. now be configured via 'EncapsulationLimit='. It allows
  1690. modifying the maximum additional levels of encapsulation
  1691. that are permitted to be prepended to a packet.
  1692. * systemd now supports the concept of user buses replacing
  1693. session buses, if used with dbus-1.10 (and enabled via dbus
  1694. --enable-user-session). It previously only supported this on
  1695. kdbus-enabled systems, and this release expands this to
  1696. 'dbus-daemon' systems.
  1697. * systemd-networkd now supports predictable interface names
  1698. for virtio devices.
  1699. * systemd now optionally supports the new Linux kernel
  1700. "unified" control group hierarchy. If enabled via the kernel
  1701. command-line option 'systemd.unified_cgroup_hierarchy=1',
  1702. systemd will try to mount the unified cgroup hierarchy
  1703. directly on /sys/fs/cgroup. If not enabled, or not
  1704. available, systemd will fall back to the legacy cgroup
  1705. hierarchy setup, as before. Host system and containers can
  1706. mix and match legacy and unified hierarchies as they
  1707. wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
  1708. environment variable to individually select the hierarchy to
  1709. use for executed containers. By default, nspawn will use the
  1710. unified hierarchy for the containers if the host uses the
  1711. unified hierarchy, and the legacy hierarchy otherwise.
  1712. Please note that at this point the unified hierarchy is an
  1713. experimental kernel feature and is likely to change in one
  1714. of the next kernel releases. Therefore, it should not be
  1715. enabled by default in downstream distributions yet. The
  1716. minimum required kernel version for the unified hierarchy to
  1717. work is 4.2. Note that when the unified hierarchy is used
  1718. for the first time delegated access to controllers is
  1719. safe. Because of this systemd-nspawn containers will get
  1720. access to controllers now, as will systemd user
  1721. sessions. This means containers and user sessions may now
  1722. manage their own resources, partitioning up what the system
  1723. grants them.
  1724. * A new special scope unit "init.scope" has been introduced
  1725. that encapsulates PID 1 of the system. It may be used to
  1726. determine resource usage and enforce resource limits on PID
  1727. 1 itself. PID 1 hence moved out of the root of the control
  1728. group tree.
  1729. * The cgtop tool gained support for filtering out kernel
  1730. threads when counting tasks in a control group. Also, the
  1731. count of processes is now recursively summed up by
  1732. default. Two options -k and --recursive= have been added to
  1733. revert to old behaviour. The tool has also been updated to
  1734. work correctly in containers now.
  1735. * systemd-nspawn's --bind= and --bind-ro= options have been
  1736. extended to allow creation of non-recursive bind mounts.
  1737. * libsystemd gained two new calls sd_pid_get_cgroup() and
  1738. sd_peer_get_cgroup() which return the control group path of
  1739. a process or peer of a connected AF_UNIX socket. This
  1740. function call is particularly useful when implementing
  1741. delegated subtrees support in the control group hierarchy.
  1742. * The "sd-event" event loop API of libsystemd now supports
  1743. correct dequeuing of real-time signals, without losing
  1744. signal events.
  1745. * When systemd requests a PolicyKit decision when managing
  1746. units it will now add additional fields to the request,
  1747. including unit name and desired operation. This enables more
  1748. powerful PolicyKit policies, that make decisions depending
  1749. on these parameters.
  1750. * nspawn learnt support for .nspawn settings files, that may
  1751. accompany the image files or directories of containers, and
  1752. may contain additional settings for the container. This is
  1753. an alternative to configuring container parameters via the
  1754. nspawn command line.
  1755. Contributions from: Cristian Rodríguez, Daniel Mack, David
  1756. Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
  1757. Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
  1758. Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
  1759. Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
  1760. Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
  1761. Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
  1762. Andersen, Tom Gundersen, Torstein Husebø
  1763. — Berlin, 2015-09-08
  1764. CHANGES WITH 225:
  1765. * machinectl gained a new verb 'shell' which opens a fresh
  1766. shell on the target container or the host. It is similar to
  1767. the existing 'login' command of machinectl, but spawns the
  1768. shell directly without prompting for username or
  1769. password. The pseudo machine '.host' now refers to the local
  1770. host and is used by default. Hence, 'machinectl shell' can
  1771. be used as replacement for 'su -' which spawns a session as
  1772. a fresh systemd unit in a way that is fully isolated from
  1773. the originating session.
  1774. * systemd-networkd learned to cope with private-zone DHCP
  1775. options and allows other programs to query the values.
  1776. * SELinux access control when enabling/disabling units is no
  1777. longer enforced with this release. The previous
  1778. implementation was incorrect, and a new corrected
  1779. implementation is not yet available. As unit file operations
  1780. are still protected via PolicyKit and D-Bus policy this is
  1781. not a security problem. Yet, distributions which care about
  1782. optimal SELinux support should probably not stabilize on
  1783. this release.
  1784. * sd-bus gained support for matches of type "arg0has=", that
  1785. test for membership of strings in string arrays sent in bus
  1786. messages.
  1787. * systemd-resolved now dumps the contents of its DNS and LLMNR
  1788. caches to the logs on reception of the SIGUSR1 signal. This
  1789. is useful to debug DNS behaviour.
  1790. * The coredumpctl tool gained a new --directory= option to
  1791. operate on journal files in a specific directory.
  1792. * "systemctl reboot" and related commands gained a new
  1793. "--message=" option which may be used to set a free-text
  1794. wall message when shutting down or rebooting the
  1795. system. This message is also logged, which is useful for
  1796. figuring out the reason for a reboot or shutdown a
  1797. posteriori.
  1798. * The "systemd-resolve-host" tool's -i switch now takes
  1799. network interface numbers as alternative to interface names.
  1800. * A new unit file setting for services has been introduced:
  1801. UtmpMode= allows configuration of how precisely systemd
  1802. handles utmp and wtmp entries for the service if this is
  1803. enabled. This allows writing services that appear similar to
  1804. user sessions in the output of the "w", "who", "last" and
  1805. "lastlog" tools.
  1806. * systemd-resolved will now locally synthesize DNS resource
  1807. records for the "localhost" and "gateway" domains as well as
  1808. the local hostname. This should ensure that clients querying
  1809. RRs via resolved will get similar results as those going via
  1810. NSS, if nss-myhostname is enabled.
  1811. Contributions from: Alastair Hughes, Alex Crawford, Daniel
  1812. Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
  1813. Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
  1814. Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
  1815. Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
  1816. Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
  1817. Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
  1818. Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
  1819. reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
  1820. Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
  1821. Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
  1822. WaLyong Cho, Zbigniew Jędrzejewski-Szmek
  1823. — Berlin, 2015-08-27
  1824. CHANGES WITH 224:
  1825. * The systemd-efi-boot-generator functionality was merged into
  1826. systemd-gpt-auto-generator.
  1827. * systemd-networkd now supports Group Policy for vxlan
  1828. devices. It can be enabled via the new boolean configuration
  1829. option called 'GroupPolicyExtension='.
  1830. Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
  1831. Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
  1832. Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
  1833. — Berlin, 2015-07-31
  1834. CHANGES WITH 223:
  1835. * The python-systemd code has been removed from the systemd repository.
  1836. A new repository has been created which accommodates the code from
  1837. now on, and we kindly ask distributions to create a separate package
  1838. for this:
  1839. * The systemd daemon will now reload its main configuration
  1840. (/etc/systemd/system.conf) on daemon-reload.
  1841. * sd-dhcp now exposes vendor specific extensions via
  1842. sd_dhcp_lease_get_vendor_specific().
  1843. * systemd-networkd gained a number of new configuration options.
  1844. - A new boolean configuration option for TAP devices called
  1845. 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
  1846. device, thus allowing to send and receive GSO packets.
  1847. - A new tunnel configuration option called 'CopyDSCP='.
  1848. If enabled, the DSCP field of ip6 tunnels is copied into the
  1849. decapsulated packet.
  1850. - A set of boolean bridge configuration options were added.
  1851. 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
  1852. and 'UnicastFlood=' are now parsed by networkd and applied to the
  1853. respective bridge link device via the respective IFLA_BRPORT_*
  1854. netlink attribute.
  1855. - A new string configuration option to override the hostname sent
  1856. to a DHCP server, called 'Hostname='. If set and 'SendHostname='
  1857. is true, networkd will use the configured hostname instead of the
  1858. system hostname when sending DHCP requests.
  1859. - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
  1860. networkd will configure the IPv6 flow-label of the tunnel device
  1861. according to RFC2460.
  1862. - The 'macvtap' virtual network devices are now supported, similar to
  1863. the already supported 'macvlan' devices.
  1864. * systemd-resolved now implements RFC5452 to improve resilience against
  1865. cache poisoning. Additionally, source port randomization is enabled
  1866. by default to further protect against DNS spoofing attacks.
  1867. * nss-mymachines now supports translating UIDs and GIDs of running
  1868. containers with user-namespaces enabled. If a container 'foo'
  1869. translates a host uid 'UID' to the container uid 'TUID', then
  1870. nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
  1871. (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
  1872. mapped as 'vg-foo-TGID'.
  1873. Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
  1874. Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
  1875. HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
  1876. Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
  1877. Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
  1878. Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
  1879. Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
  1880. Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
  1881. Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
  1882. Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
  1883. — Berlin, 2015-07-29
  1884. CHANGES WITH 222:
  1885. * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
  1886. There are no known issues with current sysfs, and udev does not need
  1887. or should be used to work around such bugs.
  1888. * udev does no longer enable USB HID power management. Several reports
  1889. indicate, that some devices cannot handle that setting.
  1890. * The udev accelerometer helper was removed. The functionality
  1891. is now fully included in iio-sensor-proxy. But this means,
  1892. older iio-sensor-proxy versions will no longer provide
  1893. accelerometer/orientation data with this systemd version.
  1894. Please upgrade iio-sensor-proxy to version 1.0.
  1895. * networkd gained a new configuration option IPv6PrivacyExtensions=
  1896. which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
  1897. for Stateless Address") on selected networks.
  1898. * For the sake of fewer build-time dependencies and less code in the
  1899. main repository, the python bindings are about to be removed in the
  1900. next release. A new repository has been created which accommodates
  1901. the code from now on, and we kindly ask distributions to create a
  1902. separate package for this. The removal will take place in v223.
  1904. Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
  1905. Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
  1906. daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
  1907. Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
  1908. Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
  1909. (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
  1910. Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
  1911. Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
  1912. Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
  1913. Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
  1914. — Berlin, 2015-07-07
  1915. CHANGES WITH 221:
  1916. * The sd-bus.h and sd-event.h APIs have now been declared
  1917. stable and have been added to the official interface of
  1918. sd-bus implements an alternative D-Bus client
  1919. library, that is relatively easy to use, very efficient and
  1920. supports both classic D-Bus as well as kdbus as transport
  1921. backend. sd-event is a generic event loop abstraction that
  1922. is built around Linux epoll, but adds features such as event
  1923. prioritization or efficient timer handling. Both APIs are good
  1924. choices for C programs looking for a bus and/or event loop
  1925. implementation that is minimal and does not have to be
  1926. portable to other kernels.
  1927. * kdbus support is no longer compile-time optional. It is now
  1928. always built-in. However, it can still be disabled at
  1929. runtime using the kdbus=0 kernel command line setting, and
  1930. that setting may be changed to default to off, by specifying
  1931. --disable-kdbus at build-time. Note though that the kernel
  1932. command line setting has no effect if the kdbus.ko kernel
  1933. module is not installed, in which case kdbus is (obviously)
  1934. also disabled. We encourage all downstream distributions to
  1935. begin testing kdbus by adding it to the kernel images in the
  1936. development distributions, and leaving kdbus support in
  1937. systemd enabled.
  1938. * The minimal required util-linux version has been bumped to
  1939. 2.26.
  1940. * Support for chkconfig (--enable-chkconfig) was removed in
  1941. favor of calling an abstraction tool
  1942. /lib/systemd/systemd-sysv-install. This needs to be
  1943. implemented for your distribution. See "SYSV INIT.D SCRIPTS"
  1944. in README for details.
  1945. * If there's a systemd unit and a SysV init script for the
  1946. same service name, and the user executes "systemctl enable"
  1947. for it (or a related call), then this will now enable both
  1948. (or execute the related operation on both), not just the
  1949. unit.
  1950. * The libudev API documentation has been converted from gtkdoc
  1951. into man pages.
  1952. * gudev has been removed from the systemd tree, it is now an
  1953. external project.
  1954. * The systemd-cgtop tool learnt a new --raw switch to generate
  1955. "raw" (machine parsable) output.
  1956. * networkd's IPForwarding= .network file setting learnt the
  1957. new setting "kernel", which ensures that networkd does not
  1958. change the IP forwarding sysctl from the default kernel
  1959. state.
  1960. * The systemd-logind bus API now exposes a new boolean
  1961. property "Docked" that reports whether logind considers the
  1962. system "docked", i.e. connected to a docking station or not.
  1963. Contributions from: Alex Crawford, Andreas Pokorny, Andrei
  1964. Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
  1965. Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
  1966. David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
  1967. Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
  1968. Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
  1969. Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
  1970. Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
  1971. Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
  1972. Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
  1973. Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
  1974. Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
  1975. Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
  1976. Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
  1977. Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
  1978. Fink, Zbigniew Jędrzejewski-Szmek
  1979. — Berlin, 2015-06-19
  1980. CHANGES WITH 220:
  1981. * The gudev library has been extracted into a separate repository
  1982. available at:
  1983. It is now managed as part of the Gnome project. Distributions
  1984. are recommended to pass --disable-gudev to systemd and use
  1985. gudev from the Gnome project instead. gudev is still included
  1986. in systemd, for now. It will be removed soon, though. Please
  1987. also see the announcement-thread on systemd-devel:
  1989. * systemd now exposes a CPUUsageNSec= property for each
  1990. service unit on the bus, that contains the overall consumed
  1991. CPU time of a service (the sum of what each process of the
  1992. service consumed). This value is only available if
  1993. CPUAccounting= is turned on for a service, and is then shown
  1994. in the "systemctl status" output.
  1995. * Support for configuring alternative mappings of the old SysV
  1996. runlevels to systemd targets has been removed. They are now
  1997. hardcoded in a way that runlevels 2, 3, 4 all map to
  1998. and 5 to (which
  1999. previously was already the default behaviour).
  2000. * The auto-mounter logic gained support for mount point
  2001. expiry, using a new TimeoutIdleSec= setting in .automount
  2002. units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
  2003. * The EFI System Partition (ESP) as mounted to /boot by
  2004. systemd-efi-boot-generator will now be unmounted
  2005. automatically after 2 minutes of not being used. This should
  2006. minimize the risk of ESP corruptions.
  2007. * New /etc/fstab options x-systemd.requires= and
  2008. x-systemd.requires-mounts-for= are now supported to express
  2009. additional dependencies for mounts. This is useful for
  2010. journalling file systems that support external journal
  2011. devices or overlay file systems that require underlying file
  2012. systems to be mounted.
  2013. * systemd does not support direct live-upgrades (via systemctl
  2014. daemon-reexec) from versions older than v44 anymore. As no
  2015. distribution we are aware of shipped such old versions in a
  2016. stable release this should not be problematic.
  2017. * When systemd forks off a new per-connection service instance
  2018. it will now set the $REMOTE_ADDR environment variable to the
  2019. remote IP address, and $REMOTE_PORT environment variable to
  2020. the remote IP port. This behaviour is similar to the
  2021. corresponding environment variables defined by CGI.
  2022. * systemd-networkd gained support for uplink failure
  2023. detection. The BindCarrier= option allows binding interface
  2024. configuration dynamically to the link sense of other
  2025. interfaces. This is useful to achieve behaviour like in
  2026. network switches.
  2027. * systemd-networkd gained support for configuring the DHCP
  2028. client identifier to use when requesting leases.
  2029. * systemd-networkd now has a per-network UseNTP= option to
  2030. configure whether NTP server information acquired via DHCP
  2031. is passed on to services like systemd-timesyncd.
  2032. * systemd-networkd gained support for vti6 tunnels.
  2033. * Note that systemd-networkd manages the sysctl variable
  2034. /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
  2035. it is configured for since v219. The variable controls IP
  2036. forwarding, and is a per-interface alternative to the global
  2037. /proc/sys/net/ipv[46]/ip_forward. This setting is
  2038. configurable in the IPForward= option, which defaults to
  2039. "no". This means if networkd is used for an interface it is
  2040. no longer sufficient to set the global sysctl option to turn
  2041. on IP forwarding! Instead, the .network file option
  2042. IPForward= needs to be turned on! Note that the
  2043. implementation of this behaviour was broken in v219 and has
  2044. been fixed in v220.
  2045. * Many bonding and vxlan options are now configurable in
  2046. systemd-networkd.
  2047. * systemd-nspawn gained a new --property= setting to set unit
  2048. properties for the container scope. This is useful for
  2049. setting resource parameters (e.g. "CPUShares=500") on
  2050. containers started from the command line.
  2051. * systemd-nspawn gained a new --private-users= switch to make
  2052. use of user namespacing available on recent Linux kernels.
  2053. * systemd-nspawn may now be called as part of a shell pipeline
  2054. in which case the pipes used for stdin and stdout are passed
  2055. directly to the process invoked in the container, without
  2056. indirection via a pseudo tty.
  2057. * systemd-nspawn gained a new switch to control the UNIX
  2058. signal to use when killing the init process of the container
  2059. when shutting down.
  2060. * systemd-nspawn gained a new --overlay= switch for mounting
  2061. overlay file systems into the container using the new kernel
  2062. overlayfs support.
  2063. * When a container image is imported via systemd-importd and
  2064. the host file system is not btrfs, a loopback block device
  2065. file is created in /var/lib/machines.raw with a btrfs file
  2066. system inside. It is then mounted to /var/lib/machines to
  2067. enable btrfs features for container management. The loopback
  2068. file and btrfs file system is grown as needed when container
  2069. images are imported via systemd-importd.
  2070. * systemd-machined/systemd-importd gained support for btrfs
  2071. quota, to enforce container disk space limits on disk. This
  2072. is exposed in "machinectl set-limit".
  2073. * systemd-importd now can import containers from local .tar,
  2074. .raw and .qcow2 images, and export them to .tar and .raw. It
  2075. can also import dkr v2 images now from the network (on top
  2076. of v1 as before).
  2077. * systemd-importd gained support for verifying downloaded
  2078. images with gpg2 (previously only gpg1 was supported).
  2079. * systemd-machined, systemd-logind, systemd: most bus calls
  2080. are now accessible to unprivileged processes via
  2081. PolicyKit. Also, systemd-logind will now allow users to kill
  2082. their own sessions without further privileges or
  2083. authorization.
  2084. * systemd-shutdownd has been removed. This service was
  2085. previously responsible for implementing scheduled shutdowns
  2086. as exposed in /usr/bin/shutdown's time parameter. This
  2087. functionality has now been moved into systemd-logind and is
  2088. accessible via a bus interface.
  2089. * "systemctl reboot" gained a new switch --firmware-setup that
  2090. can be used to reboot into the EFI firmware setup, if that
  2091. is available. systemd-logind now exposes an API on the bus
  2092. to trigger such reboots, in case graphical desktop UIs want
  2093. to cover this functionality.
  2094. * "systemctl enable", "systemctl disable" and "systemctl mask"
  2095. now support a new "--now" switch. If specified the units
  2096. that are enabled will also be started, and the ones
  2097. disabled/masked also stopped.
  2098. * The Gummiboot EFI boot loader tool has been merged into
  2099. systemd, and renamed to "systemd-boot". The bootctl tool has been
  2100. updated to support systemd-boot.
  2101. * An EFI kernel stub has been added that may be used to create
  2102. kernel EFI binaries that contain not only the actual kernel,
  2103. but also an initrd, boot splash, command line and OS release
  2104. information. This combined binary can then be signed as a
  2105. single image, so that the firmware can verify it all in one
  2106. step. systemd-boot has special support for EFI binaries created
  2107. like this and can extract OS release information from them
  2108. and show them in the boot menu. This functionality is useful
  2109. to implement cryptographically verified boot schemes.
  2110. * Optional support has been added to systemd-fsck to pass
  2111. fsck's progress report to an AF_UNIX socket in the file
  2112. system.
  2113. * udev will no longer create device symlinks for all block
  2114. devices by default. A blacklist for excluding special block
  2115. devices from this logic has been turned into a whitelist
  2116. that requires picking block devices explicitly that require
  2117. device symlinks.
  2118. * A new (currently still internal) API sd-device.h has been
  2119. added to libsystemd. This modernized API is supposed to
  2120. replace libudev eventually. In fact, already much of libudev
  2121. is now just a wrapper around sd-device.h.
  2122. * A new hwdb database for storing metadata about pointing
  2123. stick devices has been added.
  2124. * systemd-tmpfiles gained support for setting file attributes
  2125. similar to the "chattr" tool with new 'h' and 'H' lines.
  2126. * systemd-journald will no longer unconditionally set the
  2127. btrfs NOCOW flag on new journal files. This is instead done
  2128. with tmpfiles snippet using the new 'h' line type. This
  2129. allows easy disabling of this logic, by masking the
  2130. journal-nocow.conf tmpfiles file.
  2131. * systemd-journald will now translate audit message types to
  2132. human readable identifiers when writing them to the
  2133. journal. This should improve readability of audit messages.
  2134. * The LUKS logic gained support for the offset= and skip=
  2135. options in /etc/crypttab, as previously implemented by
  2136. Debian.
  2137. * /usr/lib/os-release gained a new optional field VARIANT= for
  2138. distributions that support multiple variants (such as a
  2139. desktop edition, a server edition, ...)
  2140. Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
  2141. Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
  2142. Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
  2143. Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
  2144. Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
  2145. Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
  2146. Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
  2147. Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
  2148. Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
  2149. Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
  2150. Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
  2151. Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
  2152. Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
  2153. Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
  2154. De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
  2155. Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
  2156. Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
  2157. Michael Biebl, Michael Marineau, Michael Olbrich, Michal
  2158. Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
  2159. Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
  2160. Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
  2161. Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
  2162. Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
  2163. Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
  2164. Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
  2165. Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
  2166. Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
  2167. — Berlin, 2015-05-22
  2168. CHANGES WITH 219:
  2169. * Introduce a new API "sd-hwdb.h" for querying the hardware
  2170. metadata database. With this minimal interface one can query
  2171. and enumerate the udev hwdb, decoupled from the old libudev
  2172. library. libudev's interface for this is now only a wrapper
  2173. around sd-hwdb. A new tool systemd-hwdb has been added to
  2174. interface with and update the database.
  2175. * When any of systemd's tools copies files (for example due to
  2176. tmpfiles' C lines) a btrfs reflink will attempted first,
  2177. before bytewise copying is done.
  2178. * systemd-nspawn gained a new --ephemeral switch. When
  2179. specified a btrfs snapshot is taken of the container's root
  2180. directory, and immediately removed when the container
  2181. terminates again. Thus, a container can be started whose
  2182. changes never alter the container's root directory, and are
  2183. lost on container termination. This switch can also be used
  2184. for starting a container off the root file system of the
  2185. host without affecting the host OS. This switch is only
  2186. available on btrfs file systems.
  2187. * systemd-nspawn gained a new --template= switch. It takes the
  2188. path to a container tree to use as template for the tree
  2189. specified via --directory=, should that directory be
  2190. missing. This allows instantiating containers dynamically,
  2191. on first run. This switch is only available on btrfs file
  2192. systems.
  2193. * When a .mount unit refers to a mount point on which multiple
  2194. mounts are stacked, and the .mount unit is stopped all of
  2195. the stacked mount points will now be unmounted until no
  2196. mount point remains.
  2197. * systemd now has an explicit notion of supported and
  2198. unsupported unit types. Jobs enqueued for unsupported unit
  2199. types will now fail with an "unsupported" error code. More
  2200. specifically .swap, .automount and .device units are not
  2201. supported in containers, .busname units are not supported on
  2202. non-kdbus systems. .swap and .automount are also not
  2203. supported if their respective kernel compile time options
  2204. are disabled.
  2205. * machinectl gained support for two new "copy-from" and
  2206. "copy-to" commands for copying files from a running
  2207. container to the host or vice versa.
  2208. * machinectl gained support for a new "bind" command to bind
  2209. mount host directories into local containers. This is
  2210. currently only supported for nspawn containers.
  2211. * networkd gained support for configuring bridge forwarding
  2212. database entries (fdb) from .network files.
  2213. * A new tiny daemon "systemd-importd" has been added that can
  2214. download container images in tar, raw, qcow2 or dkr formats,
  2215. and make them available locally in /var/lib/machines, so
  2216. that they can run as nspawn containers. The daemon can GPG
  2217. verify the downloads (not supported for dkr, since it has no
  2218. provisions for verifying downloads). It will transparently
  2219. decompress bz2, xz, gzip compressed downloads if necessary,
  2220. and restore sparse files on disk. The daemon uses privilege
  2221. separation to ensure the actual download logic runs with
  2222. fewer privileges than the daemon itself. machinectl has
  2223. gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
  2224. make the functionality of importd available to the
  2225. user. With this in place the Fedora and Ubuntu "Cloud"
  2226. images can be downloaded and booted as containers unmodified
  2227. (the Fedora images lack the appropriate GPG signature files
  2228. currently, so they cannot be verified, but this will change
  2229. soon, hopefully). Note that downloading images is currently
  2230. only fully supported on btrfs.
  2231. * machinectl is now able to list container images found in
  2232. /var/lib/machines, along with some metadata about sizes of
  2233. disk and similar. If the directory is located on btrfs and
  2234. quota is enabled, this includes quota display. A new command
  2235. "image-status" has been added that shows additional
  2236. information about images.
  2237. * machinectl is now able to clone container images
  2238. efficiently, if the underlying file system (btrfs) supports
  2239. it, with the new "machinectl clone" command. It also
  2240. gained commands for renaming and removing images, as well as
  2241. marking them read-only or read-write (supported also on
  2242. legacy file systems).
  2243. * networkd gained support for collecting LLDP network
  2244. announcements, from hardware that supports this. This is
  2245. shown in networkctl output.
  2246. * systemd-run gained support for a new -t (--pty) switch for
  2247. invoking a binary on a pty whose input and output is
  2248. connected to the invoking terminal. This allows executing
  2249. processes as system services while interactively
  2250. communicating with them via the terminal. Most interestingly
  2251. this is supported across container boundaries. Invoking
  2252. "systemd-run -t /bin/bash" is an alternative to running a
  2253. full login session, the difference being that the former
  2254. will not register a session, nor go through the PAM session
  2255. setup.
  2256. * tmpfiles gained support for a new "v" line type for creating
  2257. btrfs subvolumes. If the underlying file system is a legacy
  2258. file system, this automatically degrades to creating a
  2259. normal directory. Among others /var/lib/machines is now
  2260. created like this at boot, should it be missing.
  2261. * The directory /var/lib/containers/ has been deprecated and
  2262. been replaced by /var/lib/machines. The term "machines" has
  2263. been used in the systemd context as generic term for both
  2264. VMs and containers, and hence appears more appropriate for
  2265. this, as the directory can also contain raw images bootable
  2266. via qemu/kvm.
  2267. * systemd-nspawn when invoked with -M but without --directory=
  2268. or --image= is now capable of searching for the container
  2269. root directory, subvolume or disk image automatically, in
  2270. /var/lib/machines. systemd-nspawn@.service has been updated
  2271. to make use of this, thus allowing it to be used for raw
  2272. disk images, too.
  2273. * A new unit has been introduced that is
  2274. supposed to group all containers/VMs invoked as services on
  2275. the system. systemd-nspawn@.service has been updated to
  2276. integrate with that.
  2277. * machinectl gained a new "start" command, for invoking a
  2278. container as a service. "machinectl start foo" is mostly
  2279. equivalent to "systemctl start systemd-nspawn@foo.service",
  2280. but handles escaping in a nicer way.
  2281. * systemd-nspawn will now mount most of the cgroupfs tree
  2282. read-only into each container, with the exception of the
  2283. container's own subtree in the name=systemd hierarchy.
  2284. * journald now sets the special FS_NOCOW file flag for its
  2285. journal files. This should improve performance on btrfs, by
  2286. avoiding heavy fragmentation when journald's write-pattern
  2287. is used on COW file systems. It degrades btrfs' data
  2288. integrity guarantees for the files to the same levels as for
  2289. ext3/ext4 however. This should be OK though as journald does
  2290. its own data integrity checks and all its objects are
  2291. checksummed on disk. Also, journald should handle btrfs disk
  2292. full events a lot more gracefully now, by processing SIGBUS
  2293. errors, and not relying on fallocate() anymore.
  2294. * When journald detects that journal files it is writing to
  2295. have been deleted it will immediately start new journal
  2296. files.
  2297. * systemd now provides a way to store file descriptors
  2298. per-service in PID 1. This is useful for daemons to ensure
  2299. that fds they require are not lost during a daemon
  2300. restart. The fds are passed to the daemon on the next
  2301. invocation in the same way socket activation fds are
  2302. passed. This is now used by journald to ensure that the
  2303. various sockets connected to all the system's stdout/stderr
  2304. are not lost when journald is restarted. File descriptors
  2305. may be stored in PID 1 via the sd_pid_notify_with_fds() API,
  2306. an extension to sd_notify(). Note that a limit is enforced
  2307. on the number of fds a service can store in PID 1, and it
  2308. defaults to 0, so that no fds may be stored, unless this is
  2309. explicitly turned on.
  2310. * The default TERM variable to use for units connected to a
  2311. terminal, when no other value is explicitly is set is now
  2312. vt220 rather than vt102. This should be fairly safe still,
  2313. but allows PgUp/PgDn work.
  2314. * The /etc/crypttab option header= as known from Debian is now
  2315. supported.
  2316. * "loginctl user-status" and "loginctl session-status" will
  2317. now show the last 10 lines of log messages of the
  2318. user/session following the status output. Similar,
  2319. "machinectl status" will show the last 10 log lines
  2320. associated with a virtual machine or container
  2321. service. (Note that this is usually not the log messages
  2322. done in the VM/container itself, but simply what the
  2323. container manager logs. For nspawn this includes all console
  2324. output however.)
  2325. * "loginctl session-status" without further argument will now
  2326. show the status of the session of the caller. Similar,
  2327. "lock-session", "unlock-session", "activate",
  2328. "enable-linger", "disable-linger" may now be called without
  2329. session/user parameter in which case they apply to the
  2330. caller's session/user.
  2331. * An X11 session scriptlet is now shipped that uploads
  2332. $DISPLAY and $XAUTHORITY into the environment of the systemd
  2333. --user daemon if a session begins. This should improve
  2334. compatibility with X11 enabled applications run as systemd
  2335. user services.
  2336. * Generators are now subject to masking via /etc and /run, the
  2337. same way as unit files.
  2338. * networkd .network files gained support for configuring
  2339. per-link IPv4/IPv6 packet forwarding as well as IPv4
  2340. masquerading. This is by default turned on for veth links to
  2341. containers, as registered by systemd-nspawn. This means that
  2342. nspawn containers run with --network-veth will now get
  2343. automatic routed access to the host's networks without any
  2344. further configuration or setup, as long as networkd runs on
  2345. the host.
  2346. * systemd-nspawn gained the --port= (-p) switch to expose TCP
  2347. or UDP posts of a container on the host. With this in place
  2348. it is possible to run containers with private veth links
  2349. (--network-veth), and have their functionality exposed on
  2350. the host as if their services were running directly on the
  2351. host.
  2352. * systemd-nspawn's --network-veth switch now gained a short
  2353. version "-n", since with the changes above it is now truly
  2354. useful out-of-the-box. The systemd-nspawn@.service has been
  2355. updated to make use of it too by default.
  2356. * systemd-nspawn will now maintain a per-image R/W lock, to
  2357. ensure that the same image is not started more than once
  2358. writable. (It's OK to run an image multiple times
  2359. simultaneously in read-only mode.)
  2360. * systemd-nspawn's --image= option is now capable of
  2361. dissecting and booting MBR and GPT disk images that contain
  2362. only a single active Linux partition. Previously it
  2363. supported only GPT disk images with proper GPT type
  2364. IDs. This allows running cloud images from major
  2365. distributions directly with systemd-nspawn, without
  2366. modification.
  2367. * In addition to collecting mouse dpi data in the udev
  2368. hardware database, there's now support for collecting angle
  2369. information for mouse scroll wheels. The database is
  2370. supposed to guarantee similar scrolling behavior on mice
  2371. that it knows about. There's also support for collecting
  2372. information about Touchpad types.
  2373. * udev's input_id built-in will now also collect touch screen
  2374. dimension data and attach it to probed devices.
  2375. * /etc/os-release gained support for a Distribution Privacy
  2376. Policy link field.
  2377. * networkd gained support for creating "ipvlan", "gretap",
  2378. "ip6gre", "ip6gretap" and "ip6tnl" network devices.
  2379. * systemd-tmpfiles gained support for "a" lines for setting
  2380. ACLs on files.
  2381. * systemd-nspawn will now mount /tmp in the container to
  2382. tmpfs, automatically.
  2383. * systemd now exposes the memory.usage_in_bytes cgroup
  2384. attribute and shows it for each service in the "systemctl
  2385. status" output, if available.
  2386. * When the user presses Ctrl-Alt-Del more than 7x within 2s an
  2387. immediate reboot is triggered. This useful if shutdown is
  2388. hung and is unable to complete, to expedite the
  2389. operation. Note that this kind of reboot will still unmount
  2390. all file systems, and hence should not result in fsck being
  2391. run on next reboot.
  2392. * A .device unit for an optical block device will now be
  2393. considered active only when a medium is in the drive. Also,
  2394. mount units are now bound to their backing devices thus
  2395. triggering automatic unmounting when devices become
  2396. unavailable. With this in place systemd will now
  2397. automatically unmount left-over mounts when a CD-ROM is
  2398. ejected or an USB stick is yanked from the system.
  2399. * networkd-wait-online now has support for waiting for
  2400. specific interfaces only (with globbing), and for giving up
  2401. after a configurable timeout.
  2402. * networkd now exits when idle. It will be automatically
  2403. restarted as soon as interfaces show up, are removed or
  2404. change state. networkd will stay around as long as there is
  2405. at least one DHCP state machine or similar around, that keep
  2406. it non-idle.
  2407. * networkd may now configure IPv6 link-local addressing in
  2408. addition to IPv4 link-local addressing.
  2409. * The IPv6 "token" for use in SLAAC may now be configured for
  2410. each .network interface in networkd.
  2411. * Routes configured with networkd may now be assigned a scope
  2412. in .network files.
  2413. * networkd's [Match] sections now support globbing and lists
  2414. of multiple space-separated matches per item.
  2415. Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
  2416. Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
  2417. Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
  2418. Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
  2419. Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
  2420. Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
  2421. Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
  2422. Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
  2423. Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
  2424. Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
  2425. Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
  2426. Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
  2427. Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
  2428. Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
  2429. Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
  2430. Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
  2431. Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
  2432. Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
  2433. Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
  2434. Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
  2435. Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
  2436. Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
  2437. Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
  2438. Hoffmann, Zbigniew Jędrzejewski-Szmek
  2439. — Berlin, 2015-02-16
  2440. CHANGES WITH 218:
  2441. * When querying unit file enablement status (for example via
  2442. "systemctl is-enabled"), a new state "indirect" is now known
  2443. which indicates that a unit might not be enabled itself, but
  2444. another unit listed in its Also= setting might be.
  2445. * Similar to the various existing ConditionXYZ= settings for
  2446. units, there are now matching AssertXYZ= settings. While
  2447. failing conditions cause a unit to be skipped, but its job
  2448. to succeed, failing assertions declared like this will cause
  2449. a unit start operation and its job to fail.
  2450. * hostnamed now knows a new chassis type "embedded".
  2451. * systemctl gained a new "edit" command. When used on a unit
  2452. file, this allows extending unit files with .d/ drop-in
  2453. configuration snippets or editing the full file (after
  2454. copying it from /usr/lib to /etc). This will invoke the
  2455. user's editor (as configured with $EDITOR), and reload the
  2456. modified configuration after editing.
  2457. * "systemctl status" now shows the suggested enablement state
  2458. for a unit, as declared in the (usually vendor-supplied)
  2459. system preset files.
  2460. * nss-myhostname will now resolve the single-label host name
  2461. "gateway" to the locally configured default IP routing
  2462. gateways, ordered by their metrics. This assigns a stable
  2463. name to the used gateways, regardless which ones are
  2464. currently configured. Note that the name will only be
  2465. resolved after all other name sources (if nss-myhostname is
  2466. configured properly) and should hence not negatively impact
  2467. systems that use the single-label host name "gateway" in
  2468. other contexts.
  2469. * systemd-inhibit now allows filtering by mode when listing
  2470. inhibitors.
  2471. * Scope and service units gained a new "Delegate" boolean
  2472. property, which, when set, allows processes running inside the
  2473. unit to further partition resources. This is primarily
  2474. useful for systemd user instances as well as container
  2475. managers.
  2476. * journald will now pick up audit messages directly from
  2477. the kernel, and log them like any other log message. The
  2478. audit fields are split up and fully indexed. This means that
  2479. journalctl in many ways is now a (nicer!) alternative to
  2480. ausearch, the traditional audit client. Note that this
  2481. implements only a minimal audit client. If you want the
  2482. special audit modes like reboot-on-log-overflow, please use
  2483. the traditional auditd instead, which can be used in
  2484. parallel to journald.
  2485. * The ConditionSecurity= unit file option now understands the
  2486. special string "audit" to check whether auditing is
  2487. available.
  2488. * journalctl gained two new commands --vacuum-size= and
  2489. --vacuum-time= to delete old journal files until the
  2490. remaining ones take up no more than the specified size on disk,
  2491. or are not older than the specified time.
  2492. * A new, native PPPoE library has been added to sd-network,
  2493. systemd's library of light-weight networking protocols. This
  2494. library will be used in a future version of networkd to
  2495. enable PPPoE communication without an external pppd daemon.
  2496. * The busctl tool now understands a new "capture" verb that
  2497. works similar to "monitor", but writes a packet capture
  2498. trace to STDOUT that can be redirected to a file which is
  2499. compatible with libcap's capture file format. This can then
  2500. be loaded in Wireshark and similar tools to inspect bus
  2501. communication.
  2502. * The busctl tool now understands a new "tree" verb that shows
  2503. the object trees of a specific service on the bus, or of all
  2504. services.
  2505. * The busctl tool now understands a new "introspect" verb that
  2506. shows all interfaces and members of objects on the bus,
  2507. including their signature and values. This is particularly
  2508. useful to get more information about bus objects shown by
  2509. the new "busctl tree" command.
  2510. * The busctl tool now understands new verbs "call",
  2511. "set-property" and "get-property" for invoking bus method
  2512. calls, setting and getting bus object properties in a
  2513. friendly way.
  2514. * busctl gained a new --augment-creds= argument that controls
  2515. whether the tool shall augment credential information it
  2516. gets from the bus with data from /proc, in a possibly
  2517. race-ful way.
  2518. * nspawn's --link-journal= switch gained two new values
  2519. "try-guest" and "try-host" that work like "guest" and
  2520. "host", but do not fail if the host has no persistent
  2521. journalling enabled. -j is now equivalent to
  2522. --link-journal=try-guest.
  2523. * macvlan network devices created by nspawn will now have
  2524. stable MAC addresses.
  2525. * A new SmackProcessLabel= unit setting has been added, which
  2526. controls the SMACK security label processes forked off by
  2527. the respective unit shall use.
  2528. * If compiled with --enable-xkbcommon, systemd-localed will
  2529. verify x11 keymap settings by compiling the given keymap. It
  2530. will spew out warnings if the compilation fails. This
  2531. requires libxkbcommon to be installed.
  2532. * When a coredump is collected, a larger number of metadata
  2533. fields is now collected and included in the journal records
  2534. created for it. More specifically, control group membership,
  2535. environment variables, memory maps, working directory,
  2536. chroot directory, /proc/$PID/status, and a list of open file
  2537. descriptors is now stored in the log entry.
  2538. * The udev hwdb now contains DPI information for mice. For
  2539. details see:
  2541. * All systemd programs that read standalone configuration
  2542. files in /etc now also support a corresponding series of
  2543. .conf.d configuration directories in /etc/, /run/,
  2544. /usr/local/lib/, /usr/lib/, and (if configured with
  2545. --enable-split-usr) /lib/. In particular, the following
  2546. configuration files now have corresponding configuration
  2547. directories: system.conf user.conf, logind.conf,
  2548. journald.conf, sleep.conf, bootchart.conf, coredump.conf,
  2549. resolved.conf, timesyncd.conf, journal-remote.conf, and
  2550. journal-upload.conf. Note that distributions should use the
  2551. configuration directories in /usr/lib/; the directories in
  2552. /etc/ are reserved for the system administrator.
  2553. * systemd-rfkill will no longer take the rfkill device name
  2554. into account when storing rfkill state on disk, as the name
  2555. might be dynamically assigned and not stable. Instead, the
  2556. ID_PATH udev variable combined with the rfkill type (wlan,
  2557. bluetooth, ...) is used.
  2558. * A new service systemd-machine-id-commit.service has been
  2559. added. When used on systems where /etc is read-only during
  2560. boot, and /etc/machine-id is not initialized (but an empty
  2561. file), this service will copy the temporary machine ID
  2562. created as replacement into /etc after the system is fully
  2563. booted up. This is useful for systems that are freshly
  2564. installed with a non-initialized machine ID, but should get
  2565. a fixed machine ID for subsequent boots.
  2566. * networkd's .netdev files now provide a large set of
  2567. configuration parameters for VXLAN devices. Similarly, the
  2568. bridge port cost parameter is now configurable in .network
  2569. files. There's also new support for configuring IP source
  2570. routing. networkd .link files gained support for a new
  2571. OriginalName= match that is useful to match against the
  2572. original interface name the kernel assigned. .network files
  2573. may include MTU= and MACAddress= fields for altering the MTU
  2574. and MAC address while being connected to a specific network
  2575. interface.
  2576. * The LUKS logic gained supported for configuring
  2577. UUID-specific key files. There's also new support for naming
  2578. LUKS device from the kernel command line, using the new
  2579. argument.
  2580. * Timer units may now be transiently created via the bus API
  2581. (this was previously already available for scope and service
  2582. units). In addition it is now possible to create multiple
  2583. transient units at the same time with a single bus call. The
  2584. "systemd-run" tool has been updated to make use of this for
  2585. running commands on a specified time, in at(1)-style.
  2586. * tmpfiles gained support for "t" lines, for assigning
  2587. extended attributes to files. Among other uses this may be
  2588. used to assign SMACK labels to files.
  2589. Contributions from: Alin Rauta, Alison Chaiken, Andrej
  2590. Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
  2591. Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
  2592. Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
  2593. Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
  2594. Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
  2595. Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
  2596. Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
  2597. Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
  2598. Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
  2599. Michael Biebl, Michael Chapman, Michael Marineau, Michal
  2600. Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
  2601. Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
  2602. Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
  2603. Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
  2604. Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
  2605. Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
  2606. Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
  2607. — Berlin, 2014-12-10
  2608. CHANGES WITH 217:
  2609. * journalctl gained the new options -t/--identifier= to match
  2610. on the syslog identifier (aka "tag"), as well as --utc to
  2611. show log timestamps in the UTC timezone. journalctl now also
  2612. accepts -n/--lines=all to disable line capping in a pager.
  2613. * journalctl gained a new switch, --flush, that synchronously
  2614. flushes logs from /run/log/journal to /var/log/journal if
  2615. persistent storage is enabled. systemd-journal-flush.service
  2616. now waits until the operation is complete.
  2617. * Services can notify the manager before they start a reload
  2618. (by sending RELOADING=1) or shutdown (by sending
  2619. STOPPING=1). This allows the manager to track and show the
  2620. internal state of daemons and closes a race condition when
  2621. the process is still running but has closed its D-Bus
  2622. connection.
  2623. * Services with Type=oneshot do not have to have any ExecStart
  2624. commands anymore.
  2625. * User units are now loaded also from
  2626. $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
  2627. /run/systemd/user directory that was already previously
  2628. supported, but is under the control of the user.
  2629. * Job timeouts (i.e. time-outs on the time a job that is
  2630. queued stays in the run queue) can now optionally result in
  2631. immediate reboot or power-off actions (JobTimeoutAction= and
  2632. JobTimeoutRebootArgument=). This is useful on ".target"
  2633. units, to limit the maximum time a target remains
  2634. undispatched in the run queue, and to trigger an emergency
  2635. operation in such a case. This is now used by default to
  2636. turn off the system if boot-up (as defined by everything in
  2637. hangs and does not complete for at least
  2638. 15min. Also, if power-off or reboot hang for at least 30min
  2639. an immediate power-off/reboot operation is triggered. This
  2640. functionality is particularly useful to increase reliability
  2641. on embedded devices, but also on laptops which might
  2642. accidentally get powered on when carried in a backpack and
  2643. whose boot stays stuck in a hard disk encryption passphrase
  2644. question.
  2645. * systemd-logind can be configured to also handle lid switch
  2646. events even when the machine is docked or multiple displays
  2647. are attached (HandleLidSwitchDocked= option).
  2648. * A helper binary and a service have been added which can be
  2649. used to resume from hibernation in the initramfs. A
  2650. generator will parse the resume= option on the kernel
  2651. command line to trigger resume.
  2652. * A user console daemon systemd-consoled has been
  2653. added. Currently, it is a preview, and will so far open a
  2654. single terminal on each session of the user marked as
  2655. Desktop=systemd-console.
  2656. * Route metrics can be specified for DHCP routes added by
  2657. systemd-networkd.
  2658. * The SELinux context of socket-activated services can be set
  2659. from the information provided by the networking stack
  2660. (SELinuxContextFromNet= option).
  2661. * Userspace firmware loading support has been removed and
  2662. the minimum supported kernel version is thus bumped to 3.7.
  2663. * Timeout for udev workers has been increased from 1 to 3
  2664. minutes, but a warning will be printed after 1 minute to
  2665. help diagnose kernel modules that take a long time to load.
  2666. * Udev rules can now remove tags on devices with TAG-="foobar".
  2667. * systemd's readahead implementation has been removed. In many
  2668. circumstances it didn't give expected benefits even for
  2669. rotational disk drives and was becoming less relevant in the
  2670. age of SSDs. As none of the developers has been using
  2671. rotating media anymore, and nobody stepped up to actively
  2672. maintain this component of systemd it has now been removed.
  2673. * Swap units can use Options= to specify discard options.
  2674. Discard options specified for swaps in /etc/fstab are now
  2675. respected.
  2676. * Docker containers are now detected as a separate type of
  2677. virtualization.
  2678. * The Password Agent protocol gained support for queries where
  2679. the user input is shown, useful e.g. for user names.
  2680. systemd-ask-password gained a new --echo option to turn that
  2681. on.
  2682. * The default sysctl.d/ snippets will now set:
  2683. net.core.default_qdisc = fq_codel
  2684. This selects Fair Queuing Controlled Delay as the default
  2685. queuing discipline for network interfaces. fq_codel helps
  2686. fight the network bufferbloat problem. It is believed to be
  2687. a good default with no tuning required for most workloads.
  2688. Downstream distributions may override this choice. On 10Gbit
  2689. servers that do not do forwarding, "fq" may perform better.
  2690. Systems without a good clocksource should use "pfifo_fast".
  2691. * If kdbus is enabled during build a new option BusPolicy= is
  2692. available for service units, that allows locking all service
  2693. processes into a stricter bus policy, in order to limit
  2694. access to various bus services, or even hide most of them
  2695. from the service's view entirely.
  2696. * networkctl will now show the .network and .link file
  2697. networkd has applied to a specific interface.
  2698. * sd-login gained a new API call sd_session_get_desktop() to
  2699. query which desktop environment has been selected for a
  2700. session.
  2701. * UNIX utmp support is now compile-time optional to support
  2702. legacy-free systems.
  2703. * systemctl gained two new commands "add-wants" and
  2704. "add-requires" for pulling in units from specific targets
  2705. easily.
  2706. * If the word "rescue" is specified on the kernel command line
  2707. the system will now boot into rescue mode (aka
  2708., which was previously available only by
  2709. specifying "1" or "" on the kernel
  2710. command line. This new kernel command line option nicely
  2711. mirrors the already existing "emergency" kernel command line
  2712. option.
  2713. * New kernel command line options mount.usr=, mount.usrflags=,
  2714. mount.usrfstype= have been added that match root=, rootflags=,
  2715. rootfstype= but allow mounting a specific file system to
  2716. /usr.
  2717. * The $NOTIFY_SOCKET is now also passed to control processes of
  2718. services, not only the main process.
  2719. * This version reenables support for fsck's -l switch. This
  2720. means at least version v2.25 of util-linux is required for
  2721. operation, otherwise dead-locks on device nodes may
  2722. occur. Again: you need to update util-linux to at least
  2723. v2.25 when updating systemd to v217.
  2724. * The "multi-seat-x" tool has been removed from systemd, as
  2725. its functionality has been integrated into X servers 1.16,
  2726. and the tool is hence redundant. It is recommended to update
  2727. display managers invoking this tool to simply invoke X
  2728. directly from now on, again.
  2729. * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
  2730. message flag has been added for all of systemd's PolicyKit
  2731. authenticated method calls has been added. In particular
  2732. this now allows optional interactive authorization via
  2733. PolicyKit for many of PID1's privileged operations such as
  2734. unit file enabling and disabling.
  2735. * "udevadm hwdb --update" learnt a new switch "--usr" for
  2736. placing the rebuilt hardware database in /usr instead of
  2737. /etc. When used only hardware database entries stored in
  2738. /usr will be used, and any user database entries in /etc are
  2739. ignored. This functionality is useful for vendors to ship a
  2740. pre-built database on systems where local configuration is
  2741. unnecessary or unlikely.
  2742. * Calendar time specifications in .timer units now also
  2743. understand the strings "semi-annually", "quarterly" and
  2744. "minutely" as shortcuts (in addition to the preexisting
  2745. "anually", "hourly", ...).
  2746. * systemd-tmpfiles will now correctly create files in /dev
  2747. at boot which are marked for creation only at boot. It is
  2748. recommended to always create static device nodes with 'c!'
  2749. and 'b!', so that they are created only at boot and not
  2750. overwritten at runtime.
  2751. * When the watchdog logic is used for a service (WatchdogSec=)
  2752. and the watchdog timeout is hit the service will now be
  2753. terminated with SIGABRT (instead of just SIGTERM), in order
  2754. to make sure a proper coredump and backtrace is
  2755. generated. This ensures that hanging services will result in
  2756. similar coredump/backtrace behaviour as services that hit a
  2757. segmentation fault.
  2758. Contributions from: Andreas Henriksson, Andrei Borzenkov,
  2759. Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
  2760. Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
  2761. Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
  2762. Herrmann, David Sommerseth, David Strauss, Emil Renner
  2763. Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
  2764. Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
  2765. Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
  2766. Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
  2767. Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
  2768. Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
  2769. Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
  2770. Michael Marineau, Michael Olbrich, Michael Scherer, Michal
  2771. Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
  2772. Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
  2773. Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
  2774. Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
  2775. Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
  2776. Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
  2777. Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
  2778. Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
  2779. Jędrzejewski-Szmek
  2780. — Berlin, 2014-10-28
  2781. CHANGES WITH 216:
  2782. * timedated no longer reads NTP implementation unit names from
  2783. /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
  2784. implementations should add a
  2785. Conflicts=systemd-timesyncd.service
  2786. to their unit files to take over and replace systemd's NTP
  2787. default functionality.
  2788. * systemd-sysusers gained a new line type "r" for configuring
  2789. which UID/GID ranges to allocate system users/groups
  2790. from. Lines of type "u" may now add an additional column
  2791. that specifies the home directory for the system user to be
  2792. created. Also, systemd-sysusers may now optionally read user
  2793. information from STDIN instead of a file. This is useful for
  2794. invoking it from RPM preinst scriptlets that need to create
  2795. users before the first RPM file is installed since these
  2796. files might need to be owned by them. A new
  2797. %sysusers_create_inline RPM macro has been introduced to do
  2798. just that. systemd-sysusers now updates the shadow files as
  2799. well as the user/group databases, which should enhance
  2800. compatibility with certain tools like grpck.
  2801. * A number of bus APIs of PID 1 now optionally consult
  2802. PolicyKit to permit access for otherwise unprivileged
  2803. clients under certain conditions. Note that this currently
  2804. doesn't support interactive authentication yet, but this is
  2805. expected to be added eventually, too.
  2806. * /etc/machine-info now has new fields for configuring the
  2807. deployment environment of the machine, as well as the
  2808. location of the machine. hostnamectl has been updated with
  2809. new command to update these fields.
  2810. * systemd-timesyncd has been updated to automatically acquire
  2811. NTP server information from systemd-networkd, which might
  2812. have been discovered via DHCP.
  2813. * systemd-resolved now includes a caching DNS stub resolver
  2814. and a complete LLMNR name resolution implementation. A new
  2815. NSS module "nss-resolve" has been added which can be used
  2816. instead of glibc's own "nss-dns" to resolve hostnames via
  2817. systemd-resolved. Hostnames, addresses and arbitrary RRs may
  2818. be resolved via systemd-resolved D-Bus APIs. In contrast to
  2819. the glibc internal resolver systemd-resolved is aware of
  2820. multi-homed system, and keeps DNS server and caches separate
  2821. and per-interface. Queries are sent simultaneously on all
  2822. interfaces that have DNS servers configured, in order to
  2823. properly handle VPNs and local LANs which might resolve
  2824. separate sets of domain names. systemd-resolved may acquire
  2825. DNS server information from systemd-networkd automatically,
  2826. which in turn might have discovered them via DHCP. A tool
  2827. "systemd-resolve-host" has been added that may be used to
  2828. query the DNS logic in resolved. systemd-resolved implements
  2829. IDNA and automatically uses IDNA or UTF-8 encoding depending
  2830. on whether classic DNS or LLMNR is used as transport. In the
  2831. next releases we intend to add a DNSSEC and mDNS/DNS-SD
  2832. implementation to systemd-resolved.
  2833. * A new NSS module nss-mymachines has been added, that
  2834. automatically resolves the names of all local registered
  2835. containers to their respective IP addresses.
  2836. * A new client tool "networkctl" for systemd-networkd has been
  2837. added. It currently is entirely passive and will query
  2838. networking configuration from udev, rtnetlink and networkd,
  2839. and present it to the user in a very friendly
  2840. way. Eventually, we hope to extend it to become a full
  2841. control utility for networkd.
  2842. * .socket units gained a new DeferAcceptSec= setting that
  2843. controls the kernels' TCP_DEFER_ACCEPT sockopt for
  2844. TCP. Similarly, support for controlling TCP keep-alive
  2845. settings has been added (KeepAliveTimeSec=,
  2846. KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
  2847. turning off Nagle's algorithm on TCP has been added
  2848. (NoDelay=).
  2849. * logind learned a new session type "web", for use in projects
  2850. like Cockpit which register web clients as PAM sessions.
  2851. * timer units with at least one OnCalendar= setting will now
  2852. be started only after has been
  2853. reached. This way they will not elapse before the system
  2854. clock has been corrected by a local NTP client or
  2855. similar. This is particular useful on RTC-less embedded
  2856. machines, that come up with an invalid system clock.
  2857. * systemd-nspawn's --network-veth= switch should now result in
  2858. stable MAC addresses for both the outer and the inner side
  2859. of the link.
  2860. * systemd-nspawn gained a new --volatile= switch for running
  2861. container instances with /etc or /var unpopulated.
  2862. * The kdbus client code has been updated to use the new Linux
  2863. 3.17 memfd subsystem instead of the old kdbus-specific one.
  2864. * systemd-networkd's DHCP client and server now support
  2865. FORCERENEW. There are also new configuration options to
  2866. configure the vendor client identifier and broadcast mode
  2867. for DHCP.
  2868. * systemd will no longer inform the kernel about the current
  2869. timezone, as this is necessarily incorrect and racy as the
  2870. kernel has no understanding of DST and similar
  2871. concepts. This hence means FAT timestamps will be always
  2872. considered UTC, similar to what Android is already
  2873. doing. Also, when the RTC is configured to the local time
  2874. (rather than UTC) systemd will never synchronize back to it,
  2875. as this might confuse Windows at a later boot.
  2876. * systemd-analyze gained a new command "verify" for offline
  2877. validation of unit files.
  2878. * systemd-networkd gained support for a couple of additional
  2879. settings for bonding networking setups. Also, the metric for
  2880. statically configured routes may now be configured. For
  2881. network interfaces where this is appropriate the peer IP
  2882. address may now be configured.
  2883. * systemd-networkd's DHCP client will no longer request
  2884. broadcasting by default, as this tripped up some networks.
  2885. For hardware where broadcast is required the feature should
  2886. be switched back on using RequestBroadcast=yes.
  2887. * systemd-networkd will now set up IPv4LL addresses (when
  2888. enabled) even if DHCP is configured successfully.
  2889. * udev will now default to respect network device names given
  2890. by the kernel when the kernel indicates that these are
  2891. predictable. This behavior can be tweaked by changing
  2892. NamePolicy= in the relevant .link file.
  2893. * A new library systemd-terminal has been added that
  2894. implements full TTY stream parsing and rendering. This
  2895. library is supposed to be used later on for implementing a
  2896. full userspace VT subsystem, replacing the current kernel
  2897. implementation.
  2898. * A new tool systemd-journal-upload has been added to push
  2899. journal data to a remote system running
  2900. systemd-journal-remote.
  2901. * journald will no longer forward all local data to another
  2902. running syslog daemon. This change has been made because
  2903. rsyslog (which appears to be the most commonly used syslog
  2904. implementation these days) no longer makes use of this, and
  2905. instead pulls the data out of the journal on its own. Since
  2906. forwarding the messages to a non-existent syslog server is
  2907. more expensive than we assumed we have now turned this
  2908. off. If you run a syslog server that is not a recent rsyslog
  2909. version, you have to turn this option on again
  2910. (ForwardToSyslog= in journald.conf).
  2911. * journald now optionally supports the LZ4 compressor for
  2912. larger journal fields. This compressor should perform much
  2913. better than XZ which was the previous default.
  2914. * machinectl now shows the IP addresses of local containers,
  2915. if it knows them, plus the interface name of the container.
  2916. * A new tool "systemd-escape" has been added that makes it
  2917. easy to escape strings to build unit names and similar.
  2918. * sd_notify() messages may now include a new ERRNO= field
  2919. which is parsed and collected by systemd and shown among the
  2920. "systemctl status" output for a service.
  2921. * A new component "systemd-firstboot" has been added that
  2922. queries the most basic systemd information (timezone,
  2923. hostname, root password) interactively on first
  2924. boot. Alternatively it may also be used to provision these
  2925. things offline on OS images installed into directories.
  2926. * The default sysctl.d/ snippets will now set
  2927. net.ipv4.conf.default.promote_secondaries=1
  2928. This has the benefit of no flushing secondary IP addresses
  2929. when primary addresses are removed.
  2930. Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
  2931. Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
  2932. Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
  2933. Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
  2934. Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
  2935. B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
  2936. Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
  2937. Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
  2938. Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
  2939. Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
  2940. Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
  2941. Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
  2942. Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
  2943. Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
  2944. Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
  2945. — Berlin, 2014-08-19
  2946. CHANGES WITH 215:
  2947. * A new tool systemd-sysusers has been added. This tool
  2948. creates system users and groups in /etc/passwd and
  2949. /etc/group, based on static declarative system user/group
  2950. definitions in /usr/lib/sysusers.d/. This is useful to
  2951. enable factory resets and volatile systems that boot up with
  2952. an empty /etc directory, and thus need system users and
  2953. groups created during early boot. systemd now also ships
  2954. with two default sysusers.d/ files for the most basic
  2955. users and groups systemd and the core operating system
  2956. require.
  2957. * A new tmpfiles snippet has been added that rebuilds the
  2958. essential files in /etc on boot, should they be missing.
  2959. * A directive for ensuring automatic clean-up of
  2960. /var/cache/man/ has been removed from the default
  2961. configuration. This line should now be shipped by the man
  2962. implementation. The necessary change has been made to the
  2963. man-db implementation. Note that you need to update your man
  2964. implementation to one that ships this line, otherwise no
  2965. automatic clean-up of /var/cache/man will take place.
  2966. * A new condition ConditionNeedsUpdate= has been added that
  2967. may conditionalize services to only run when /etc or /var
  2968. are "older" than the vendor operating system resources in
  2969. /usr. This is useful for reconstructing or updating /etc
  2970. after an offline update of /usr or a factory reset, on the
  2971. next reboot. Services that want to run once after such an
  2972. update or reset should use this condition and order
  2973. themselves before the new systemd-update-done.service, which
  2974. will mark the two directories as fully updated. A number of
  2975. service files have been added making use of this, to rebuild
  2976. the udev hardware database, the journald message catalog and
  2977. dynamic loader cache (ldconfig). The systemd-sysusers tool
  2978. described above also makes use of this now. With this in
  2979. place it is now possible to start up a minimal operating
  2980. system with /etc empty cleanly. For more information on the
  2981. concepts involved see this recent blog story:
  2983. * A new system group "input" has been introduced, and all
  2984. input device nodes get this group assigned. This is useful
  2985. for system-level software to get access to input devices. It
  2986. complements what is already done for "audio" and "video".
  2987. * systemd-networkd learnt minimal DHCPv4 server support in
  2988. addition to the existing DHCPv4 client support. It also
  2989. learnt DHCPv6 client and IPv6 Router Solicitation client
  2990. support. The DHCPv4 client gained support for static routes
  2991. passed in from the server. Note that the [DHCPv4] section
  2992. known in older systemd-networkd versions has been renamed to
  2993. [DHCP] and is now also used by the DHCPv6 client. Existing
  2994. .network files using settings of this section should be
  2995. updated, though compatibility is maintained. Optionally, the
  2996. client hostname may now be sent to the DHCP server.
  2997. * networkd gained support for vxlan virtual networks as well
  2998. as tun/tap and dummy devices.
  2999. * networkd gained support for automatic allocation of address
  3000. ranges for interfaces from a system-wide pool of
  3001. addresses. This is useful for dynamically managing a large
  3002. number of interfaces with a single network configuration
  3003. file. In particular this is useful to easily assign
  3004. appropriate IP addresses to the veth links of a large number
  3005. of nspawn instances.
  3006. * RPM macros for processing sysusers, sysctl and binfmt
  3007. drop-in snippets at package installation time have been
  3008. added.
  3009. * The /etc/os-release file should now be placed in
  3010. /usr/lib/os-release. The old location is automatically
  3011. created as symlink. /usr/lib is the more appropriate
  3012. location of this file, since it shall actually describe the
  3013. vendor operating system shipped in /usr, and not the
  3014. configuration stored in /etc.
  3015. * .mount units gained a new boolean SloppyOptions= setting
  3016. that maps to mount(8)'s -s option which enables permissive
  3017. parsing of unknown mount options.
  3018. * tmpfiles learnt a new "L+" directive which creates a symlink
  3019. but (unlike "L") deletes a pre-existing file first, should
  3020. it already exist and not already be the correct
  3021. symlink. Similarly, "b+", "c+" and "p+" directives have been
  3022. added as well, which create block and character devices, as
  3023. well as fifos in the filesystem, possibly removing any
  3024. pre-existing files of different types.
  3025. * For tmpfiles' "L", "L+", "C" and "C+" directives the final
  3026. 'argument' field (which so far specified the source to
  3027. symlink/copy the files from) is now optional. If omitted the
  3028. same file os copied from /usr/share/factory/ suffixed by the
  3029. full destination path. This is useful for populating /etc
  3030. with essential files, by copying them from vendor defaults
  3031. shipped in /usr/share/factory/etc.
  3032. * A new command "systemctl preset-all" has been added that
  3033. applies the service preset settings to all installed unit
  3034. files. A new switch --preset-mode= has been added that
  3035. controls whether only enable or only disable operations
  3036. shall be executed.
  3037. * A new command "systemctl is-system-running" has been added
  3038. that allows checking the overall state of the system, for
  3039. example whether it is fully up and running.
  3040. * When the system boots up with an empty /etc, the equivalent
  3041. to "systemctl preset-all" is executed during early boot, to
  3042. make sure all default services are enabled after a factory
  3043. reset.
  3044. * systemd now contains a minimal preset file that enables the
  3045. most basic services systemd ships by default.
  3046. * Unit files' [Install] section gained a new DefaultInstance=
  3047. field for defining the default instance to create if a
  3048. template unit is enabled with no instance specified.
  3049. * A new passive target has been added
  3050. that may be used by services that need to make they run and
  3051. finish before the first LUKS cryptographic device is set up.
  3052. * The /dev/loop-control and /dev/btrfs-control device nodes
  3053. are now owned by the "disk" group by default, opening up
  3054. access to this group.
  3055. * systemd-coredump will now automatically generate a
  3056. stack trace of all core dumps taking place on the system,
  3057. based on elfutils' libdw library. This stack trace is logged
  3058. to the journal.
  3059. * systemd-coredump may now optionally store coredumps directly
  3060. on disk (in /var/lib/systemd/coredump, possibly compressed),
  3061. instead of storing them unconditionally in the journal. This
  3062. mode is the new default. A new configuration file
  3063. /etc/systemd/coredump.conf has been added to configure this
  3064. and other parameters of systemd-coredump.
  3065. * coredumpctl gained a new "info" verb to show details about a
  3066. specific coredump. A new switch "-1" has also been added
  3067. that makes sure to only show information about the most
  3068. recent entry instead of all entries. Also, as the tool is
  3069. generally useful now the "systemd-" prefix of the binary
  3070. name has been removed. Distributions that want to maintain
  3071. compatibility with the old name should add a symlink from
  3072. the old name to the new name.
  3073. * journald's SplitMode= now defaults to "uid". This makes sure
  3074. that unprivileged users can access their own coredumps with
  3075. coredumpctl without restrictions.
  3076. * New kernel command line options "systemd.wants=" (for
  3077. pulling an additional unit during boot), "systemd.mask="
  3078. (for masking a specific unit for the boot), and
  3079. "systemd.debug-shell" (for enabling the debug shell on tty9)
  3080. have been added. This is implemented in the new generator
  3081. "systemd-debug-generator".
  3082. * systemd-nspawn will now by default filter a couple of
  3083. syscalls for containers, among them those required for
  3084. kernel module loading, direct x86 IO port access, swap
  3085. management, and kexec. Most importantly though
  3086. open_by_handle_at() is now prohibited for containers,
  3087. closing a hole similar to a recently discussed vulnerability
  3088. in docker regarding access to files on file hierarchies the
  3089. container should normally not have access to. Note that, for
  3090. nspawn, we generally make no security claims anyway (and
  3091. this is explicitly documented in the man page), so this is
  3092. just a fix for one of the most obvious problems.
  3093. * A new man page file-hierarchy(7) has been added that
  3094. contains a minimized, modernized version of the file system
  3095. layout systemd expects, similar in style to the FHS
  3096. specification or hier(5). A new tool systemd-path(1) has
  3097. been added to query many of these paths for the local
  3098. machine and user.
  3099. * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
  3100. longer done. Since the directory now has a per-user size
  3101. limit, and is cleaned on logout this appears unnecessary,
  3102. in particular since this now brings the lifecycle of this
  3103. directory closer in line with how IPC objects are handled.
  3104. * systemd.pc now exports a number of additional directories,
  3105. including $libdir (which is useful to identify the library
  3106. path for the primary architecture of the system), and a
  3107. couple of drop-in directories.
  3108. * udev's predictable network interface names now use the dev_port
  3109. sysfs attribute, introduced in linux 3.15 instead of dev_id to
  3110. distinguish between ports of the same PCI function. dev_id should
  3111. only be used for ports using the same HW address, hence the need
  3112. for dev_port.
  3113. * machined has been updated to export the OS version of a
  3114. container (read from /etc/os-release and
  3115. /usr/lib/os-release) on the bus. This is now shown in
  3116. "machinectl status" for a machine.
  3117. * A new service setting RestartForceExitStatus= has been
  3118. added. If configured to a set of exit signals or process
  3119. return values, the service will be restarted when the main
  3120. daemon process exits with any of them, regardless of the
  3121. Restart= setting.
  3122. * systemctl's -H switch for connecting to remote systemd
  3123. machines has been extended so that it may be used to
  3124. directly connect to a specific container on the
  3125. host. "systemctl -H root@foobar:waldi" will now connect as
  3126. user "root" to host "foobar", and then proceed directly to
  3127. the container named "waldi". Note that currently you have to
  3128. authenticate as user "root" for this to work, as entering
  3129. containers is a privileged operation.
  3130. Contributions from: Andreas Henriksson, Benjamin Steinwender,
  3131. Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
  3132. Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
  3133. Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
  3134. Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
  3135. Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
  3136. Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
  3137. Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
  3138. Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
  3139. Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
  3140. Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
  3141. Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
  3142. — Berlin, 2014-07-03
  3143. CHANGES WITH 214:
  3144. * As an experimental feature, udev now tries to lock the
  3145. disk device node (flock(LOCK_SH|LOCK_NB)) while it
  3146. executes events for the disk or any of its partitions.
  3147. Applications like partitioning programs can lock the
  3148. disk device node (flock(LOCK_EX)) and claim temporary
  3149. device ownership that way; udev will entirely skip all event
  3150. handling for this disk and its partitions. If the disk
  3151. was opened for writing, the close will trigger a partition
  3152. table rescan in udev's "watch" facility, and if needed
  3153. synthesize "change" events for the disk and all its partitions.
  3154. This is now unconditionally enabled, and if it turns out to
  3155. cause major problems, we might turn it on only for specific
  3156. devices, or might need to disable it entirely. Device Mapper
  3157. devices are excluded from this logic.
  3158. * We temporarily dropped the "-l" switch for fsck invocations,
  3159. since they collide with the flock() logic above. util-linux
  3160. upstream has been changed already to avoid this conflict,
  3161. and we will readd "-l" as soon as util-linux with this
  3162. change has been released.
  3163. * The dependency on libattr has been removed. Since a long
  3164. time, the extended attribute calls have moved to glibc, and
  3165. libattr is thus unnecessary.
  3166. * Virtualization detection works without privileges now. This
  3167. means the systemd-detect-virt binary no longer requires
  3168. CAP_SYS_PTRACE file capabilities, and our daemons can run
  3169. with fewer privileges.
  3170. * systemd-networkd now runs under its own "systemd-network"
  3171. user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
  3172. CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
  3173. loses the ability to write to files owned by root this way.
  3174. * Similarly, systemd-resolved now runs under its own
  3175. "systemd-resolve" user with no capabilities remaining.
  3176. * Similarly, systemd-bus-proxyd now runs under its own
  3177. "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
  3178. * systemd-networkd gained support for setting up "veth"
  3179. virtual Ethernet devices for container connectivity, as well
  3180. as GRE and VTI tunnels.
  3181. * systemd-networkd will no longer automatically attempt to
  3182. manually load kernel modules necessary for certain tunnel
  3183. transports. Instead, it is assumed the kernel loads them
  3184. automatically when required. This only works correctly on
  3185. very new kernels. On older kernels, please consider adding
  3186. the kernel modules to /etc/modules-load.d/ as a work-around.
  3187. * The resolv.conf file systemd-resolved generates has been
  3188. moved to /run/systemd/resolve/. If you have a symlink from
  3189. /etc/resolv.conf, it might be necessary to correct it.
  3190. * Two new service settings, ProtectHome= and ProtectSystem=,
  3191. have been added. When enabled, they will make the user data
  3192. (such as /home) inaccessible or read-only and the system
  3193. (such as /usr) read-only, for specific services. This allows
  3194. very light-weight per-service sandboxing to avoid
  3195. modifications of user data or system files from
  3196. services. These two new switches have been enabled for all
  3197. of systemd's long-running services, where appropriate.
  3198. * Socket units gained new SocketUser= and SocketGroup=
  3199. settings to set the owner user and group of AF_UNIX sockets
  3200. and FIFOs in the file system.
  3201. * Socket units gained a new RemoveOnStop= setting. If enabled,
  3202. all FIFOS and sockets in the file system will be removed
  3203. when the specific socket unit is stopped.
  3204. * Socket units gained a new Symlinks= setting. It takes a list
  3205. of symlinks to create to file system sockets or FIFOs
  3206. created by the specific Unix sockets. This is useful to
  3207. manage symlinks to socket nodes with the same life-cycle as
  3208. the socket itself.
  3209. * The /dev/log socket and /dev/initctl FIFO have been moved to
  3210. /run, and have been replaced by symlinks. This allows
  3211. connecting to these facilities even if PrivateDevices=yes is
  3212. used for a service (which makes /dev/log itself unavailable,
  3213. but /run is left). This also has the benefit of ensuring
  3214. that /dev only contains device nodes, directories and
  3215. symlinks, and nothing else.
  3216. * sd-daemon gained two new calls sd_pid_notify() and
  3217. sd_pid_notifyf(). They are similar to sd_notify() and
  3218. sd_notifyf(), but allow overriding of the source PID of
  3219. notification messages if permissions permit this. This is
  3220. useful to send notify messages on behalf of a different
  3221. process (for example, the parent process). The
  3222. systemd-notify tool has been updated to make use of this
  3223. when sending messages (so that notification messages now
  3224. originate from the shell script invoking systemd-notify and
  3225. not the systemd-notify process itself. This should minimize
  3226. a race where systemd fails to associate notification
  3227. messages to services when the originating process already
  3228. vanished.
  3229. * A new "on-abnormal" setting for Restart= has been added. If
  3230. set, it will result in automatic restarts on all "abnormal"
  3231. reasons for a process to exit, which includes unclean
  3232. signals, core dumps, timeouts and watchdog timeouts, but
  3233. does not include clean and unclean exit codes or clean
  3234. signals. Restart=on-abnormal is an alternative for
  3235. Restart=on-failure for services that shall be able to
  3236. terminate and avoid restarts on certain errors, by
  3237. indicating so with an unclean exit code. Restart=on-failure
  3238. or Restart=on-abnormal is now the recommended setting for
  3239. all long-running services.
  3240. * If the InaccessibleDirectories= service setting points to a
  3241. mount point (or if there are any submounts contained within
  3242. it), it is now attempted to completely unmount it, to make
  3243. the file systems truly unavailable for the respective
  3244. service.
  3245. * The ReadOnlyDirectories= service setting and
  3246. systemd-nspawn's --read-only parameter are now recursively
  3247. applied to all submounts, too.
  3248. * Mount units may now be created transiently via the bus APIs.
  3249. * The support for SysV and LSB init scripts has been removed
  3250. from the systemd daemon itself. Instead, it is now
  3251. implemented as a generator that creates native systemd units
  3252. from these scripts when needed. This enables us to remove a
  3253. substantial amount of legacy code from PID 1, following the
  3254. fact that many distributions only ship a very small number
  3255. of LSB/SysV init scripts nowadays.
  3256. * Privileged Xen (dom0) domains are not considered
  3257. virtualization anymore by the virtualization detection
  3258. logic. After all, they generally have unrestricted access to
  3259. the hardware and usually are used to manage the unprivileged
  3260. (domU) domains.
  3261. * systemd-tmpfiles gained a new "C" line type, for copying
  3262. files or entire directories.
  3263. * systemd-tmpfiles "m" lines are now fully equivalent to "z"
  3264. lines. So far, they have been non-globbing versions of the
  3265. latter, and have thus been redundant. In future, it is
  3266. recommended to only use "z". "m" has hence been removed
  3267. from the documentation, even though it stays supported.
  3268. * A tmpfiles snippet to recreate the most basic structure in
  3269. /var has been added. This is enough to create the /var/run →
  3270. /run symlink and create a couple of structural
  3271. directories. This allows systems to boot up with an empty or
  3272. volatile /var. Of course, while with this change, the core OS
  3273. now is capable with dealing with a volatile /var, not all
  3274. user services are ready for it. However, we hope that sooner
  3275. or later, many service daemons will be changed upstream so
  3276. that they are able to automatically create their necessary
  3277. directories in /var at boot, should they be missing. This is
  3278. the first step to allow state-less systems that only require
  3279. the vendor image for /usr to boot.
  3280. * systemd-nspawn has gained a new --tmpfs= switch to mount an
  3281. empty tmpfs instance to a specific directory. This is
  3282. particularly useful for making use of the automatic
  3283. reconstruction of /var (see above), by passing --tmpfs=/var.
  3284. * Access modes specified in tmpfiles snippets may now be
  3285. prefixed with "~", which indicates that they shall be masked
  3286. by whether the existing file or directory is currently
  3287. writable, readable or executable at all. Also, if specified,
  3288. the sgid/suid/sticky bits will be masked for all
  3289. non-directories.
  3290. * A new passive target unit "" has been
  3291. added which is useful for services that shall run before any
  3292. network is configured, for example firewall scripts.
  3293. * The "floppy" group that previously owned the /dev/fd*
  3294. devices is no longer used. The "disk" group is now used
  3295. instead. Distributions should probably deprecate usage of
  3296. this group.
  3297. Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
  3298. King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
  3299. Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
  3300. Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
  3301. Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
  3302. Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
  3303. Jędrzejewski-Szmek
  3304. — Berlin, 2014-06-11
  3305. CHANGES WITH 213:
  3306. * A new "systemd-timesyncd" daemon has been added for
  3307. synchronizing the system clock across the network. It
  3308. implements an SNTP client. In contrast to NTP
  3309. implementations such as chrony or the NTP reference server,
  3310. this only implements a client side, and does not bother with
  3311. the full NTP complexity, focusing only on querying time from
  3312. one remote server and synchronizing the local clock to
  3313. it. Unless you intend to serve NTP to networked clients or
  3314. want to connect to local hardware clocks, this simple NTP
  3315. client should be more than appropriate for most
  3316. installations. The daemon runs with minimal privileges, and
  3317. has been hooked up with networkd to only operate when
  3318. network connectivity is available. The daemon saves the
  3319. current clock to disk every time a new NTP sync has been
  3320. acquired, and uses this to possibly correct the system clock
  3321. early at bootup, in order to accommodate for systems that
  3322. lack an RTC such as the Raspberry Pi and embedded devices,
  3323. and to make sure that time monotonically progresses on these
  3324. systems, even if it is not always correct. To make use of
  3325. this daemon, a new system user and group "systemd-timesync"
  3326. needs to be created on installation of systemd.
  3327. * The queue "seqnum" interface of libudev has been disabled, as
  3328. it was generally incompatible with device namespacing as
  3329. sequence numbers of devices go "missing" if the devices are
  3330. part of a different namespace.
  3331. * "systemctl list-timers" and "systemctl list-sockets" gained
  3332. a --recursive switch for showing units of these types also
  3333. for all local containers, similar in style to the already
  3334. supported --recursive switch for "systemctl list-units".
  3335. * A new RebootArgument= setting has been added for service
  3336. units, which may be used to specify a kernel reboot argument
  3337. to use when triggering reboots with StartLimitAction=.
  3338. * A new FailureAction= setting has been added for service
  3339. units which may be used to specify an operation to trigger
  3340. when a service fails. This works similarly to
  3341. StartLimitAction=, but unlike it, controls what is done
  3342. immediately rather than only after several attempts to
  3343. restart the service in question.
  3344. * hostnamed got updated to also expose the kernel name,
  3345. release, and version on the bus. This is useful for
  3346. executing commands like hostnamectl with the -H switch.
  3347. systemd-analyze makes use of this to properly display
  3348. details when running non-locally.
  3349. * The bootchart tool can now show cgroup information in the
  3350. graphs it generates.
  3351. * The CFS CPU quota cgroup attribute is now exposed for
  3352. services. The new CPUQuota= switch has been added for this
  3353. which takes a percentage value. Setting this will have the
  3354. result that a service may never get more CPU time than the
  3355. specified percentage, even if the machine is otherwise idle.
  3356. * systemd-networkd learned IPIP and SIT tunnel support.
  3357. * LSB init scripts exposing a dependency on $network will now
  3358. get a dependency on rather than simply
  3359. This should bring LSB handling closer to
  3360. what it was on SysV systems.
  3361. * A new kernel option has been added to control
  3362. how fsck shall deal with unclean file systems at boot.
  3363. * The (.ini) configuration file parser will now silently
  3364. ignore sections whose name begins with "X-". This may be
  3365. used to maintain application-specific extension sections in unit
  3366. files.
  3367. * machined gained a new API to query the IP addresses of
  3368. registered containers. "machinectl status" has been updated
  3369. to show these addresses in its output.
  3370. * A new call sd_uid_get_display() has been added to the
  3371. sd-login APIs for querying the "primary" session of a
  3372. user. The "primary" session of the user is elected from the
  3373. user's sessions and generally a graphical session is
  3374. preferred over a text one.
  3375. * A minimal systemd-resolved daemon has been added. It
  3376. currently simply acts as a companion to systemd-networkd and
  3377. manages resolv.conf based on per-interface DNS
  3378. configuration, possibly supplied via DHCP. In the long run
  3379. we hope to extend this into a local DNSSEC enabled DNS and
  3380. mDNS cache.
  3381. * The systemd-networkd-wait-online tool is now enabled by
  3382. default. It will delay until a network
  3383. connection has been configured. The tool primarily integrates
  3384. with networkd, but will also make a best effort to make sense
  3385. of network configuration performed in some other way.
  3386. * Two new service options StartupCPUShares= and
  3387. StartupBlockIOWeight= have been added that work similarly to
  3388. CPUShares= and BlockIOWeight= however only apply during
  3389. system startup. This is useful to prioritize certain services
  3390. differently during bootup than during normal runtime.
  3391. * hostnamed has been changed to prefer the statically
  3392. configured hostname in /etc/hostname (unless set to
  3393. 'localhost' or empty) over any dynamic one supplied by
  3394. dhcp. With this change, the rules for picking the hostname
  3395. match more closely the rules of other configuration settings
  3396. where the local administrator's configuration in /etc always
  3397. overrides any other settings.
  3398. Contributions fron: Ali H. Caliskan, Alison Chaiken, Bas van
  3399. den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
  3400. Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
  3401. David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
  3402. Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
  3403. Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
  3404. Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
  3405. Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
  3406. Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
  3407. Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
  3408. Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
  3409. Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
  3410. Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
  3411. Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
  3412. Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
  3413. Lindskog, WaLyong Cho, Will Woods, Zbigniew
  3414. Jędrzejewski-Szmek
  3415. — Beijing, 2014-05-28
  3416. CHANGES WITH 212:
  3417. * When restoring the screen brightness at boot, stay away from
  3418. the darkest setting or from the lowest 5% of the available
  3419. range, depending on which is the larger value of both. This
  3420. should effectively protect the user from rebooting into a
  3421. black screen, should the brightness have been set to minimum
  3422. by accident.
  3423. * sd-login gained a new sd_machine_get_class() call to
  3424. determine the class ("vm" or "container") of a machine
  3425. registered with machined.
  3426. * sd-login gained new calls
  3427. sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
  3428. to query the identity of the peer of a local AF_UNIX
  3429. connection. They operate similarly to their sd_pid_get_xyz()
  3430. counterparts.
  3431. * PID 1 will now maintain a system-wide system state engine
  3432. with the states "starting", "running", "degraded",
  3433. "maintenance", "stopping". These states are bound to system
  3434. startup, normal runtime, runtime with at least one failed
  3435. service, rescue/emergency mode and system shutdown. This
  3436. state is shown in the "systemctl status" output when no unit
  3437. name is passed. It is useful to determine system state, in
  3438. particularly when doing so for many systems or containers at
  3439. once.
  3440. * A new command "list-machines" has been added to "systemctl"
  3441. that lists all local OS containers and shows their system
  3442. state (see above), if systemd runs inside of them.
  3443. * systemctl gained a new "-r" switch to recursively enumerate
  3444. units on all local containers, when used with the
  3445. "list-unit" command (which is the default one that is
  3446. executed when no parameters are specified).
  3447. * The GPT automatic partition discovery logic will now honour
  3448. two GPT partition flags: one may be set on a partition to
  3449. cause it to be mounted read-only, and the other may be set
  3450. on a partition to ignore it during automatic discovery.
  3451. * Two new GPT type UUIDs have been added for automatic root
  3452. partition discovery, for 32-bit and 64-bit ARM. This is not
  3453. particularly useful for discovering the root directory on
  3454. these architectures during bare-metal boots (since UEFI is
  3455. not common there), but still very useful to allow booting of
  3456. ARM disk images in nspawn with the -i option.
  3457. * MAC addresses of interfaces created with nspawn's
  3458. --network-interface= switch will now be generated from the
  3459. machine name, and thus be stable between multiple invocations
  3460. of the container.
  3461. * logind will now automatically remove all IPC objects owned
  3462. by a user if she or he fully logs out. This makes sure that
  3463. users who are logged out cannot continue to consume IPC
  3464. resources. This covers SysV memory, semaphores and message
  3465. queues as well as POSIX shared memory and message
  3466. queues. Traditionally, SysV and POSIX IPC had no life-cycle
  3467. limits. With this functionality, that is corrected. This may
  3468. be turned off by using the RemoveIPC= switch of logind.conf.
  3469. * The systemd-machine-id-setup and tmpfiles tools gained a
  3470. --root= switch to operate on a specific root directory,
  3471. instead of /.
  3472. * journald can now forward logged messages to the TTYs of all
  3473. logged in users ("wall"). This is the default for all
  3474. emergency messages now.
  3475. * A new tool systemd-journal-remote has been added to stream
  3476. journal log messages across the network.
  3477. * /sys/fs/cgroup/ is now mounted read-only after all cgroup
  3478. controller trees are mounted into it. Note that the
  3479. directories mounted beneath it are not read-only. This is a
  3480. security measure and is particularly useful because glibc
  3481. actually includes a search logic to pick any tmpfs it can
  3482. find to implement shm_open() if /dev/shm is not available
  3483. (which it might very well be in namespaced setups).
  3484. * machinectl gained a new "poweroff" command to cleanly power
  3485. down a local OS container.
  3486. * The PrivateDevices= unit file setting will now also drop the
  3487. CAP_MKNOD capability from the capability bound set, and
  3488. imply DevicePolicy=closed.
  3489. * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
  3490. comprehensively on all long-running systemd services where
  3491. this is appropriate.
  3492. * systemd-udevd will now run in a disassociated mount
  3493. namespace. To mount directories from udev rules, make sure to
  3494. pull in mount units via SYSTEMD_WANTS properties.
  3495. * The kdbus support gained support for uploading policy into
  3496. the kernel. sd-bus gained support for creating "monitoring"
  3497. connections that can eavesdrop into all bus communication
  3498. for debugging purposes.
  3499. * Timestamps may now be specified in seconds since the UNIX
  3500. epoch Jan 1st, 1970 by specifying "@" followed by the value
  3501. in seconds.
  3502. * Native tcpwrap support in systemd has been removed. tcpwrap
  3503. is old code, not really maintained anymore and has serious
  3504. shortcomings, and better options such as firewalls
  3505. exist. For setups that require tcpwrap usage, please
  3506. consider invoking your socket-activated service via tcpd,
  3507. like on traditional inetd.
  3508. * A new system.conf configuration option
  3509. DefaultTimerAccuracySec= has been added that controls the
  3510. default AccuracySec= setting of .timer units.
  3511. * Timer units gained a new WakeSystem= switch. If enabled,
  3512. timers configured this way will cause the system to resume
  3513. from system suspend (if the system supports that, which most
  3514. do these days).
  3515. * Timer units gained a new Persistent= switch. If enabled,
  3516. timers configured this way will save to disk when they have
  3517. been last triggered. This information is then used on next
  3518. reboot to possible execute overdue timer events, that
  3519. could not take place because the system was powered off.
  3520. This enables simple anacron-like behaviour for timer units.
  3521. * systemctl's "list-timers" will now also list the time a
  3522. timer unit was last triggered in addition to the next time
  3523. it will be triggered.
  3524. * systemd-networkd will now assign predictable IPv4LL
  3525. addresses to its local interfaces.
  3526. Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
  3527. Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
  3528. Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
  3529. Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
  3530. Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
  3531. Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
  3532. Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
  3533. Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
  3534. Jędrzejewski-Szmek
  3535. — Berlin, 2014-03-25
  3536. CHANGES WITH 211:
  3537. * A new unit file setting RestrictAddressFamilies= has been
  3538. added to restrict which socket address families unit
  3539. processes gain access to. This takes address family names
  3540. like "AF_INET" or "AF_UNIX", and is useful to minimize the
  3541. attack surface of services via exotic protocol stacks. This
  3542. is built on seccomp system call filters.
  3543. * Two new unit file settings RuntimeDirectory= and
  3544. RuntimeDirectoryMode= have been added that may be used to
  3545. manage a per-daemon runtime directories below /run. This is
  3546. an alternative for setting up directory permissions with
  3547. tmpfiles snippets, and has the advantage that the runtime
  3548. directory's lifetime is bound to the daemon runtime and that
  3549. the daemon starts up with an empty directory each time. This
  3550. is particularly useful when writing services that drop
  3551. privileges using the User= or Group= setting.
  3552. * The DeviceAllow= unit setting now supports globbing for
  3553. matching against device group names.
  3554. * The systemd configuration file system.conf gained new
  3555. settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
  3556. DefaultMemoryAccounting= to globally turn on/off accounting
  3557. for specific resources (cgroups) for all units. These
  3558. settings may still be overridden individually in each unit
  3559. though.
  3560. * systemd-gpt-auto-generator is now able to discover /srv and
  3561. root partitions in addition to /home and swap partitions. It
  3562. also supports LUKS-encrypted partitions now. With this in
  3563. place, automatic discovery of partitions to mount following
  3564. the Discoverable Partitions Specification
  3565. (
  3566. is now a lot more complete. This allows booting without
  3567. /etc/fstab and without root= on the kernel command line on
  3568. systems prepared appropriately.
  3569. * systemd-nspawn gained a new --image= switch which allows
  3570. booting up disk images and Linux installations on any block
  3571. device that follow the Discoverable Partitions Specification
  3572. (see above). This means that installations made with
  3573. appropriately updated installers may now be started and
  3574. deployed using container managers, completely
  3575. unmodified. (We hope that libvirt-lxc will add support for
  3576. this feature soon, too.)
  3577. * systemd-nspawn gained a new --network-macvlan= setting to
  3578. set up a private macvlan interface for the
  3579. container. Similarly, systemd-networkd gained a new
  3580. Kind=macvlan setting in .netdev files.
  3581. * systemd-networkd now supports configuring local addresses
  3582. using IPv4LL.
  3583. * A new tool systemd-network-wait-online has been added to
  3584. synchronously wait for network connectivity using
  3585. systemd-networkd.
  3586. * The sd-bus.h bus API gained a new sd_bus_track object for
  3587. tracking the life-cycle of bus peers. Note that sd-bus.h is
  3588. still not a public API though (unless you specify
  3589. --enable-kdbus on the configure command line, which however
  3590. voids your warranty and you get no API stability guarantee).
  3591. * The $XDG_RUNTIME_DIR runtime directories for each user are
  3592. now individual tmpfs instances, which has the benefit of
  3593. introducing separate pools for each user, with individual
  3594. size limits, and thus making sure that unprivileged clients
  3595. can no longer negatively impact the system or other users by
  3596. filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
  3597. RuntimeDirectorySize= has been introduced that allows
  3598. controlling the default size limit for all users. It
  3599. defaults to 10% of the available physical memory. This is no
  3600. replacement for quotas on tmpfs though (which the kernel
  3601. still does not support), as /dev/shm and /tmp are still
  3602. shared resources used by both the system and unprivileged
  3603. users.
  3604. * logind will now automatically turn off automatic suspending
  3605. on laptop lid close when more than one display is
  3606. connected. This was previously expected to be implemented
  3607. individually in desktop environments (such as GNOME),
  3608. however has been added to logind now, in order to fix a
  3609. boot-time race where a desktop environment might not have
  3610. been started yet and thus not been able to take an inhibitor
  3611. lock at the time where logind already suspends the system
  3612. due to a closed lid.
  3613. * logind will now wait at least 30s after each system
  3614. suspend/resume cycle, and 3min after system boot before
  3615. suspending the system due to a closed laptop lid. This
  3616. should give USB docking stations and similar enough time to
  3617. be probed and configured after system resume and boot in
  3618. order to then act as suspend blocker.
  3619. * systemd-run gained a new --property= setting which allows
  3620. initialization of resource control properties (and others)
  3621. for the created scope or service unit. Example: "systemd-run
  3622. --property=BlockIOWeight=10 updatedb" may be used to run
  3623. updatedb at a low block IO scheduling weight.
  3624. * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
  3625. now also work in --scope mode.
  3626. * When systemd is compiled with kdbus support, basic support
  3627. for enforced policies is now in place. (Note that enabling
  3628. kdbus still voids your warranty and no API compatibility
  3629. promises are made.)
  3630. Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
  3631. K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
  3632. Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
  3633. Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
  3634. Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
  3635. Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
  3636. Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
  3637. Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
  3638. Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
  3639. Zbigniew Jędrzejewski-Szmek
  3640. — Berlin, 2014-03-12
  3641. CHANGES WITH 210:
  3642. * systemd will now relabel /dev after loading the SMACK policy
  3643. according to SMACK rules.
  3644. * A new unit file option AppArmorProfile= has been added to
  3645. set the AppArmor profile for the processes of a unit.
  3646. * A new condition check ConditionArchitecture= has been added
  3647. to conditionalize units based on the system architecture, as
  3648. reported by uname()'s "machine" field.
  3649. * systemd-networkd now supports matching on the system
  3650. virtualization, architecture, kernel command line, host name
  3651. and machine ID.
  3652. * logind is now a lot more aggressive when suspending the
  3653. machine due to a closed laptop lid. Instead of acting only
  3654. on the lid close action, it will continuously watch the lid
  3655. status and act on it. This is useful for laptops where the
  3656. power button is on the outside of the chassis so that it can
  3657. be reached without opening the lid (such as the Lenovo
  3658. Yoga). On those machines, logind will now immediately
  3659. re-suspend the machine if the power button has been
  3660. accidentally pressed while the laptop was suspended and in a
  3661. backpack or similar.
  3662. * logind will now watch SW_DOCK switches and inhibit reaction
  3663. to the lid switch if it is pressed. This means that logind
  3664. will not suspend the machine anymore if the lid is closed
  3665. and the system is docked, if the laptop supports SW_DOCK
  3666. notifications via the input layer. Note that ACPI docking
  3667. stations do not generate this currently. Also note that this
  3668. logic is usually not fully sufficient and Desktop
  3669. Environments should take a lid switch inhibitor lock when an
  3670. external display is connected, as systemd will not watch
  3671. this on its own.
  3672. * nspawn will now make use of the devices cgroup controller by
  3673. default, and only permit creation of and access to the usual
  3674. API device nodes like /dev/null or /dev/random, as well as
  3675. access to (but not creation of) the pty devices.
  3676. * We will now ship a default .network file for
  3677. systemd-networkd that automatically configures DHCP for
  3678. network interfaces created by nspawn's --network-veth or
  3679. --network-bridge= switches.
  3680. * systemd will now understand the usual M, K, G, T suffixes
  3681. according to SI conventions (i.e. to the base 1000) when
  3682. referring to throughput and hardware metrics. It will stay
  3683. with IEC conventions (i.e. to the base 1024) for software
  3684. metrics, according to what is customary according to
  3685. Wikipedia. We explicitly document which base applies for
  3686. each configuration option.
  3687. * The DeviceAllow= setting in unit files now supports a syntax
  3688. to whitelist an entire group of devices node majors at once,
  3689. based on the /proc/devices listing. For example, with the
  3690. string "char-pts", it is now possible to whitelist all
  3691. current and future pseudo-TTYs at once.
  3692. * sd-event learned a new "post" event source. Event sources of
  3693. this type are triggered by the dispatching of any event
  3694. source of a type that is not "post". This is useful for
  3695. implementing clean-up and check event sources that are
  3696. triggered by other work being done in the program.
  3697. * systemd-networkd is no longer statically enabled, but uses
  3698. the usual [Install] sections so that it can be
  3699. enabled/disabled using systemctl. It still is enabled by
  3700. default however.
  3701. * When creating a veth interface pair with systemd-nspawn, the
  3702. host side will now be prefixed with "vb-" if
  3703. --network-bridge= is used, and with "ve-" if --network-veth
  3704. is used. This way, it is easy to distinguish these cases on
  3705. the host, for example to apply different configuration to
  3706. them with systemd-networkd.
  3707. * The compatibility libraries for,
  3708., and
  3709. do not make use of IFUNC
  3710. anymore. Instead, we now build multiple times
  3711. under these alternative names. This means that the footprint
  3712. is drastically increased, but given that these are
  3713. transitional compatibility libraries, this should not matter
  3714. much. This change has been made necessary to support the ARM
  3715. platform for these compatibility libraries, as the ARM
  3716. toolchain is not really at the same level as the toolchain
  3717. for other architectures like x86 and does not support
  3718. IFUNC. Please make sure to use --enable-compat-libs only
  3719. during a transitional period!
  3720. Contributions from: Andreas Fuchs, Armin K., Colin Walters,
  3721. Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
  3722. Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
  3723. St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
  3724. Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
  3725. Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
  3726. Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
  3727. Zbigniew Jędrzejewski-Szmek
  3728. — Berlin, 2014-02-24
  3729. CHANGES WITH 209:
  3730. * A new component "systemd-networkd" has been added that can
  3731. be used to configure local network interfaces statically or
  3732. via DHCP. It is capable of bringing up bridges, VLANs, and
  3733. bonding. Currently, no hook-ups for interactive network
  3734. configuration are provided. Use this for your initrd,
  3735. container, embedded, or server setup if you need a simple,
  3736. yet powerful, network configuration solution. This
  3737. configuration subsystem is quite nifty, as it allows wildcard
  3738. hotplug matching in interfaces. For example, with a single
  3739. configuration snippet, you can configure that all Ethernet
  3740. interfaces showing up are automatically added to a bridge,
  3741. or similar. It supports link-sensing and more.
  3742. * A new tool "systemd-socket-proxyd" has been added which can
  3743. act as a bidirectional proxy for TCP sockets. This is
  3744. useful for adding socket activation support to services that
  3745. do not actually support socket activation, including virtual
  3746. machines and the like.
  3747. * Add a new tool to save/restore rfkill state on
  3748. shutdown/boot.
  3749. * Save/restore state of keyboard backlights in addition to
  3750. display backlights on shutdown/boot.
  3751. * udev learned a new SECLABEL{} construct to label device
  3752. nodes with a specific security label when they appear. For
  3753. now, only SECLABEL{selinux} is supported, but the syntax is
  3754. prepared for additional security frameworks.
  3755. * udev gained a new scheme to configure link-level attributes
  3756. from files in /etc/systemd/network/*.link. These files can
  3757. match against MAC address, device path, driver name and type,
  3758. and will apply attributes like the naming policy, link speed,
  3759. MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
  3760. address assignment policy (randomized, ...).
  3761. * The configuration of network interface naming rules for
  3762. "permanent interface names" has changed: a new NamePolicy=
  3763. setting in the [Link] section of .link files determines the
  3764. priority of possible naming schemes (onboard, slot, MAC,
  3765. path). The default value of this setting is determined by
  3766. /usr/lib/net/links/ Old
  3767. 80-net-name-slot.rules udev configuration file has been
  3768. removed, so local configuration overriding this file should
  3769. be adapted to override instead.
  3770. * When the User= switch is used in a unit file, also
  3771. initialize $SHELL= based on the user database entry.
  3772. * systemd no longer depends on libdbus. All communication is
  3773. now done with sd-bus, systemd's low-level bus library
  3774. implementation.
  3775. * kdbus support has been added to PID 1 itself. When kdbus is
  3776. enabled, this causes PID 1 to set up the system bus and
  3777. enable support for a new ".busname" unit type that
  3778. encapsulates bus name activation on kdbus. It works a little
  3779. bit like ".socket" units, except for bus names. A new
  3780. generator has been added that converts classic dbus1 service
  3781. activation files automatically into native systemd .busname
  3782. and .service units.
  3783. * sd-bus: add a light-weight vtable implementation that allows
  3784. defining objects on the bus with a simple static const
  3785. vtable array of its methods, signals and properties.
  3786. * systemd will not generate or install static dbus
  3787. introspection data anymore to /usr/share/dbus-1/interfaces,
  3788. as the precise format of these files is unclear, and
  3789. nothing makes use of it.
  3790. * A proxy daemon is now provided to proxy clients connecting
  3791. via classic D-Bus AF_UNIX sockets to kdbus, to provide full
  3792. compatibility with classic D-Bus.
  3793. * A bus driver implementation has been added that supports the
  3794. classic D-Bus bus driver calls on kdbus, also for
  3795. compatibility purposes.
  3796. * A new API "sd-event.h" has been added that implements a
  3797. minimal event loop API built around epoll. It provides a
  3798. couple of features that direct epoll usage is lacking:
  3799. prioritization of events, scales to large numbers of timer
  3800. events, per-event timer slack (accuracy), system-wide
  3801. coalescing of timer events, exit handlers, watchdog
  3802. supervision support using systemd's sd_notify() API, child
  3803. process handling.
  3804. * A new API "sd-rntl.h" has been added that provides an API
  3805. around the route netlink interface of the kernel, similar in
  3806. style to "sd-bus.h".
  3807. * A new API "sd-dhcp-client.h" has been added that provides a
  3808. small DHCPv4 client-side implementation. This is used by
  3809. "systemd-networkd".
  3810. * There is a new kernel command line option
  3811. "systemd.restore_state=0|1". When set to "0", none of the
  3812. systemd tools will restore saved runtime state to hardware
  3813. devices. More specifically, the rfkill and backlight states
  3814. are not restored.
  3815. * The FsckPassNo= compatibility option in mount/service units
  3816. has been removed. The fstab generator will now add the
  3817. necessary dependencies automatically, and does not require
  3818. PID1's support for that anymore.
  3819. * journalctl gained a new switch, --list-boots, that lists
  3820. recent boots with their times and boot IDs.
  3821. * The various tools like systemctl, loginctl, timedatectl,
  3822. busctl, systemd-run, ... have gained a new switch "-M" to
  3823. connect to a specific, local OS container (as direct
  3824. connection, without requiring SSH). This works on any
  3825. container that is registered with machined, such as those
  3826. created by libvirt-lxc or nspawn.
  3827. * systemd-run and systemd-analyze also gained support for "-H"
  3828. to connect to remote hosts via SSH. This is particularly
  3829. useful for systemd-run because it enables queuing of jobs
  3830. onto remote systems.
  3831. * machinectl gained a new command "login" to open a getty
  3832. login in any local container. This works with any container
  3833. that is registered with machined (such as those created by
  3834. libvirt-lxc or nspawn), and which runs systemd inside.
  3835. * machinectl gained a new "reboot" command that may be used to
  3836. trigger a reboot on a specific container that is registered
  3837. with machined. This works on any container that runs an init
  3838. system of some kind.
  3839. * systemctl gained a new "list-timers" command to print a nice
  3840. listing of installed timer units with the times they elapse
  3841. next.
  3842. * Alternative reboot() parameters may now be specified on the
  3843. "systemctl reboot" command line and are passed to the
  3844. reboot() system call.
  3845. * systemctl gained a new --job-mode= switch to configure the
  3846. mode to queue a job with. This is a more generic version of
  3847. --fail, --irreversible, and --ignore-dependencies, which are
  3848. still available but not advertised anymore.
  3849. * /etc/systemd/system.conf gained new settings to configure
  3850. various default timeouts of units, as well as the default
  3851. start limit interval and burst. These may still be overridden
  3852. within each Unit.
  3853. * PID1 will now export on the bus profile data of the security
  3854. policy upload process (such as the SELinux policy upload to
  3855. the kernel).
  3856. * journald: when forwarding logs to the console, include
  3857. timestamps (following the setting in
  3858. /sys/module/printk/parameters/time).
  3859. * OnCalendar= in timer units now understands the special
  3860. strings "yearly" and "annually". (Both are equivalent)
  3861. * The accuracy of timer units is now configurable with the new
  3862. AccuracySec= setting. It defaults to 1min.
  3863. * A new dependency type JoinsNamespaceOf= has been added that
  3864. allows running two services within the same /tmp and network
  3865. namespace, if PrivateNetwork= or PrivateTmp= are used.
  3866. * A new command "cat" has been added to systemctl. It outputs
  3867. the original unit file of a unit, and concatenates the
  3868. contents of additional "drop-in" unit file snippets, so that
  3869. the full configuration is shown.
  3870. * systemctl now supports globbing on the various "list-xyz"
  3871. commands, like "list-units" or "list-sockets", as well as on
  3872. those commands which take multiple unit names.
  3873. * journalctl's --unit= switch gained support for globbing.
  3874. * All systemd daemons now make use of the watchdog logic so
  3875. that systemd automatically notices when they hang.
  3876. * If the $container_ttys environment variable is set,
  3877. getty-generator will automatically spawn a getty for each
  3878. listed tty. This is useful for container managers to request
  3879. login gettys to be spawned on as many ttys as needed.
  3880. * %h, %s, %U specifier support is not available anymore when
  3881. used in unit files for PID 1. This is because NSS calls are
  3882. not safe from PID 1. They stay available for --user
  3883. instances of systemd, and as special case for the root user.
  3884. * loginctl gained a new "--no-legend" switch to turn off output
  3885. of the legend text.
  3886. * The "sd-login.h" API gained three new calls:
  3887. sd_session_is_remote(), sd_session_get_remote_user(),
  3888. sd_session_get_remote_host() to query information about
  3889. remote sessions.
  3890. * The udev hardware database now also carries vendor/product
  3891. information of SDIO devices.
  3892. * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
  3893. determine whether watchdog notifications are requested by
  3894. the system manager.
  3895. * Socket-activated per-connection services now include a
  3896. short description of the connection parameters in the
  3897. description.
  3898. * tmpfiles gained a new "--boot" option. When this is not used,
  3899. only lines where the command character is not suffixed with
  3900. "!" are executed. When this option is specified, those
  3901. options are executed too. This partitions tmpfiles
  3902. directives into those that can be safely executed at any
  3903. time, and those which should be run only at boot (for
  3904. example, a line that creates /run/nologin).
  3905. * A new API "sd-resolve.h" has been added which provides a simple
  3906. asynchronous wrapper around glibc NSS host name resolution
  3907. calls, such as getaddrinfo(). In contrast to glibc's
  3908. getaddrinfo_a(), it does not use signals. In contrast to most
  3909. other asynchronous name resolution libraries, this one does
  3910. not reimplement DNS, but reuses NSS, so that alternate
  3911. host name resolution systems continue to work, such as mDNS,
  3912. LDAP, etc. This API is based on libasyncns, but it has been
  3913. cleaned up for inclusion in systemd.
  3914. * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
  3915. "sd-daemon.h" are no longer found in individual libraries
  3917., Instead, we have
  3918. merged them into a single library,, which
  3919. provides all symbols. The reason for this is cyclic
  3920. dependencies, as these libraries tend to use each other's
  3921. symbols. So far, we have managed to workaround that by linking
  3922. a copy of a good part of our code into each of these
  3923. libraries again and again, which, however, makes certain
  3924. things hard to do, like sharing static variables. Also, it
  3925. substantially increases footprint. With this change, there
  3926. is only one library for the basic APIs systemd
  3927. provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
  3928. "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
  3929. library as well, however are subject to the --enable-kdbus
  3930. switch (see below). Note that "sd-dhcp-client.h" is not part
  3931. of this library (this is because it only consumes, never
  3932. provides, services of/to other APIs). To make the transition
  3933. easy from the separate libraries to the unified one, we
  3934. provide the --enable-compat-libs compile-time switch which
  3935. will generate stub libraries that are compatible with the
  3936. old ones but redirect all calls to the new one.
  3937. * All of the kdbus logic and the new APIs "sd-bus.h",
  3938. "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
  3939. and "sd-utf8.h" are compile-time optional via the
  3940. "--enable-kdbus" switch, and they are not compiled in by
  3941. default. To make use of kdbus, you have to explicitly enable
  3942. the switch. Note however, that neither the kernel nor the
  3943. userspace API for all of this is considered stable yet. We
  3944. want to maintain the freedom to still change the APIs for
  3945. now. By specifying this build-time switch, you acknowledge
  3946. that you are aware of the instability of the current
  3947. APIs.
  3948. * Also, note that while kdbus is pretty much complete,
  3949. it lacks one thing: proper policy support. This means you
  3950. can build a fully working system with all features; however,
  3951. it will be highly insecure. Policy support will be added in
  3952. one of the next releases, at the same time that we will
  3953. declare the APIs stable.
  3954. * When the kernel command line argument "kdbus" is specified,
  3955. systemd will automatically load the kdbus.ko kernel module. At
  3956. this stage of development, it is only useful for testing kdbus
  3957. and should not be used in production. Note: if "--enable-kdbus"
  3958. is specified, and the kdbus.ko kernel module is available, and
  3959. "kdbus" is added to the kernel command line, the entire system
  3960. runs with kdbus instead of dbus-daemon, with the above mentioned
  3961. problem of missing the system policy enforcement. Also a future
  3962. version of kdbus.ko or a newer systemd will not be compatible with
  3963. each other, and will unlikely be able to boot the machine if only
  3964. one of them is updated.
  3965. * systemctl gained a new "import-environment" command which
  3966. uploads the caller's environment (or parts thereof) into the
  3967. service manager so that it is inherited by services started
  3968. by the manager. This is useful to upload variables like
  3969. $DISPLAY into the user service manager.
  3970. * A new PrivateDevices= switch has been added to service units
  3971. which allows running a service with a namespaced /dev
  3972. directory that does not contain any device nodes for
  3973. physical devices. More specifically, it only includes devices
  3974. such as /dev/null, /dev/urandom, and /dev/zero which are API
  3975. entry points.
  3976. * logind has been extended to support behaviour like VT
  3977. switching on seats that do not support a VT. This makes
  3978. multi-session available on seats that are not the first seat
  3979. (seat0), and on systems where kernel support for VTs has
  3980. been disabled at compile-time.
  3981. * If a process holds a delay lock for system sleep or shutdown
  3982. and fails to release it in time, we will now log its
  3983. identity. This makes it easier to identify processes that
  3984. cause slow suspends or power-offs.
  3985. * When parsing /etc/crypttab, support for a new key-slot=
  3986. option as supported by Debian is added. It allows indicating
  3987. which LUKS slot to use on disk, speeding up key loading.
  3988. * The sd_journald_sendv() API call has been checked and
  3989. officially declared to be async-signal-safe so that it may
  3990. be invoked from signal handlers for logging purposes.
  3991. * Boot-time status output is now enabled automatically after a
  3992. short timeout if boot does not progress, in order to give
  3993. the user an indication what she or he is waiting for.
  3994. * The boot-time output has been improved to show how much time
  3995. remains until jobs expire.
  3996. * The KillMode= switch in service units gained a new possible
  3997. value "mixed". If set, and the unit is shut down, then the
  3998. initial SIGTERM signal is sent only to the main daemon
  3999. process, while the following SIGKILL signal is sent to
  4000. all remaining processes of the service.
  4001. * When a scope unit is registered, a new property "Controller"
  4002. may be set. If set to a valid bus name, systemd will send a
  4003. RequestStop() signal to this name when it would like to shut
  4004. down the scope. This may be used to hook manager logic into
  4005. the shutdown logic of scope units. Also, scope units may now
  4006. be put in a special "abandoned" state, in which case the
  4007. manager process which created them takes no further
  4008. responsibilities for it.
  4009. * When reading unit files, systemd will now verify
  4010. the access mode of these files, and warn about certain
  4011. suspicious combinations. This has been added to make it
  4012. easier to track down packaging bugs where unit files are
  4013. marked executable or world-writable.
  4014. * systemd-nspawn gained a new "--setenv=" switch to set
  4015. container-wide environment variables. The similar option in
  4016. systemd-activate was renamed from "--environment=" to
  4017. "--setenv=" for consistency.
  4018. * systemd-nspawn has been updated to create a new kdbus domain
  4019. for each container that is invoked, thus allowing each
  4020. container to have its own set of system and user buses,
  4021. independent of the host.
  4022. * systemd-nspawn gained a new --drop-capability= switch to run
  4023. the container with less capabilities than the default. Both
  4024. --drop-capability= and --capability= now take the special
  4025. string "all" for dropping or keeping all capabilities.
  4026. * systemd-nspawn gained new switches for executing containers
  4027. with specific SELinux labels set.
  4028. * systemd-nspawn gained a new --quiet switch to not generate
  4029. any additional output but the container's own console
  4030. output.
  4031. * systemd-nspawn gained a new --share-system switch to run a
  4032. container without PID namespacing enabled.
  4033. * systemd-nspawn gained a new --register= switch to control
  4034. whether the container is registered with systemd-machined or
  4035. not. This is useful for containers that do not run full
  4036. OS images, but only specific apps.
  4037. * systemd-nspawn gained a new --keep-unit which may be used
  4038. when invoked as the only program from a service unit, and
  4039. results in registration of the unit service itself in
  4040. systemd-machined, instead of a newly opened scope unit.
  4041. * systemd-nspawn gained a new --network-interface= switch for
  4042. moving arbitrary interfaces to the container. The new
  4043. --network-veth switch creates a virtual Ethernet connection
  4044. between host and container. The new --network-bridge=
  4045. switch then allows assigning the host side of this virtual
  4046. Ethernet connection to a bridge device.
  4047. * systemd-nspawn gained a new --personality= switch for
  4048. setting the kernel personality for the container. This is
  4049. useful when running a 32-bit container on a 64-bit host. A
  4050. similar option Personality= is now also available for service
  4051. units to use.
  4052. * logind will now also track a "Desktop" identifier for each
  4053. session which encodes the desktop environment of it. This is
  4054. useful for desktop environments that want to identify
  4055. multiple running sessions of itself easily.
  4056. * A new SELinuxContext= setting for service units has been
  4057. added that allows setting a specific SELinux execution
  4058. context for a service.
  4059. * Most systemd client tools will now honour $SYSTEMD_LESS for
  4060. settings of the "less" pager. By default, these tools will
  4061. override $LESS to allow certain operations to work, such as
  4062. jump-to-the-end. With $SYSTEMD_LESS, it is possible to
  4063. influence this logic.
  4064. * systemd's "seccomp" hook-up has been changed to make use of
  4065. the libseccomp library instead of using its own
  4066. implementation. This has benefits for portability among
  4067. other things.
  4068. * For usage together with SystemCallFilter=, a new
  4069. SystemCallErrorNumber= setting has been introduced that
  4070. allows configuration of a system error number to be returned
  4071. on filtered system calls, instead of immediately killing the
  4072. process. Also, SystemCallArchitectures= has been added to
  4073. limit access to system calls of a particular architecture
  4074. (in order to turn off support for unused secondary
  4075. architectures). There is also a global
  4076. SystemCallArchitectures= setting in system.conf now to turn
  4077. off support for non-native system calls system-wide.
  4078. * systemd requires a kernel with a working name_to_handle_at(),
  4079. please see the kernel config requirements in the README file.
  4080. Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
  4081. Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
  4082. Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
  4083. Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
  4084. Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
  4085. David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
  4086. Elia Pinto, Florian Weimer, George McCollister, Goffredo
  4087. Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
  4088. Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
  4089. Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
  4090. Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
  4091. Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
  4092. Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
  4093. Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
  4094. Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
  4095. Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
  4096. Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
  4097. Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
  4098. Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
  4099. Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
  4100. Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
  4101. Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
  4102. Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
  4103. Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
  4104. — Berlin, 2014-02-20
  4105. CHANGES WITH 208:
  4106. * logind has gained support for facilitating privileged input
  4107. and drm device access for unprivileged clients. This work is
  4108. useful to allow Wayland display servers (and similar
  4109. programs, such as kmscon) to run under the user's ID and
  4110. access input and drm devices which are normally
  4111. protected. When this is used (and the kernel is new enough)
  4112. logind will "mute" IO on the file descriptors passed to
  4113. Wayland as long as it is in the background and "unmute" it
  4114. if it returns into the foreground. This allows secure
  4115. session switching without allowing background sessions to
  4116. eavesdrop on input and display data. This also introduces
  4117. session switching support if VT support is turned off in the
  4118. kernel, and on seats that are not seat0.
  4119. * A new kernel command line option luks.options= is understood
  4120. now which allows specifying LUKS options for usage for LUKS
  4121. encrypted partitions specified with luks.uuid=.
  4122. * tmpfiles.d(5) snippets may now use specifier expansion in
  4123. path names. More specifically %m, %b, %H, %v, are now
  4124. replaced by the local machine id, boot id, hostname, and
  4125. kernel version number.
  4126. * A new tmpfiles.d(5) command "m" has been introduced which
  4127. may be used to change the owner/group/access mode of a file
  4128. or directory if it exists, but do nothing if it does not.
  4129. * This release removes high-level support for the
  4130. MemorySoftLimit= cgroup setting. The underlying kernel
  4131. cgroup attribute memory.soft_limit= is currently badly
  4132. designed and likely to be removed from the kernel API in its
  4133. current form, hence we should not expose it for now.
  4134. * The memory.use_hierarchy cgroup attribute is now enabled for
  4135. all cgroups systemd creates in the memory cgroup
  4136. hierarchy. This option is likely to be come the built-in
  4137. default in the kernel anyway, and the non-hierarchical mode
  4138. never made much sense in the intrinsically hierarchical
  4139. cgroup system.
  4140. * A new field _SYSTEMD_SLICE= is logged along with all journal
  4141. messages containing the slice a message was generated
  4142. from. This is useful to allow easy per-customer filtering of
  4143. logs among other things.
  4144. * systemd-journald will no longer adjust the group of journal
  4145. files it creates to the "systemd-journal" group. Instead we
  4146. rely on the journal directory to be owned by the
  4147. "systemd-journal" group, and its setgid bit set, so that the
  4148. kernel file system layer will automatically enforce that
  4149. journal files inherit this group assignment. The reason for
  4150. this change is that we cannot allow NSS look-ups from
  4151. journald which would be necessary to resolve
  4152. "systemd-journal" to a numeric GID, because this might
  4153. create deadlocks if NSS involves synchronous queries to
  4154. other daemons (such as nscd, or sssd) which in turn are
  4155. logging clients of journald and might block on it, which
  4156. would then dead lock. A tmpfiles.d(5) snippet included in
  4157. systemd will make sure the setgid bit and group are
  4158. properly set on the journal directory if it exists on every
  4159. boot. However, we recommend adjusting it manually after
  4160. upgrades too (or from RPM scriptlets), so that the change is
  4161. not delayed until next reboot.
  4162. * Backlight and random seed files in /var/lib/ have moved into
  4163. the /var/lib/systemd/ directory, in order to centralize all
  4164. systemd generated files in one directory.
  4165. * Boot time performance measurements (as displayed by
  4166. "systemd-analyze" for example) will now read ACPI 5.0 FPDT
  4167. performance information if that's available to determine how
  4168. much time BIOS and boot loader initialization required. With
  4169. a sufficiently new BIOS you hence no longer need to boot
  4170. with Gummiboot to get access to such information.
  4171. Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
  4172. Cristian Rodríguez, Dave Reisner, David Herrmann, David
  4173. Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
  4174. feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
  4175. Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
  4176. Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
  4177. Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
  4178. — Berlin, 2013-10-02
  4179. CHANGES WITH 207:
  4180. * The Restart= option for services now understands a new
  4181. on-watchdog setting, which will restart the service
  4182. automatically if the service stops sending out watchdog keep
  4183. alive messages (as configured with WatchdogSec=).
  4184. * The getty generator (which is responsible for bringing up a
  4185. getty on configured serial consoles) will no longer only
  4186. start a getty on the primary kernel console but on all
  4187. others, too. This makes the order in which console= is
  4188. specified on the kernel command line less important.
  4189. * libsystemd-logind gained a new sd_session_get_vt() call to
  4190. retrieve the VT number of a session.
  4191. * If the option "tries=0" is set for an entry of /etc/crypttab
  4192. its passphrase is queried indefinitely instead of any
  4193. maximum number of tries.
  4194. * If a service with a configure PID file terminates its PID
  4195. file will now be removed automatically if it still exists
  4196. afterwards. This should put an end to stale PID files.
  4197. * systemd-run will now also take relative binary path names
  4198. for execution and no longer insists on absolute paths.
  4199. * InaccessibleDirectories= and ReadOnlyDirectories= now take
  4200. paths that are optionally prefixed with "-" to indicate that
  4201. it should not be considered a failure if they do not exist.
  4202. * journalctl -o (and similar commands) now understands a new
  4203. output mode "short-precise", it is similar to "short" but
  4204. shows timestamps with usec accuracy.
  4205. * The option "discard" (as known from Debian) is now
  4206. synonymous to "allow-discards" in /etc/crypttab. In fact,
  4207. "discard" is preferred now (since it is easier to remember
  4208. and type).
  4209. * Some licensing clean-ups were made, so that more code is now
  4210. LGPL-2.1 licensed than before.
  4211. * A minimal tool to save/restore the display backlight
  4212. brightness across reboots has been added. It will store the
  4213. backlight setting as late as possible at shutdown, and
  4214. restore it as early as possible during reboot.
  4215. * A logic to automatically discover and enable home and swap
  4216. partitions on GPT disks has been added. With this in place
  4217. /etc/fstab becomes optional for many setups as systemd can
  4218. discover certain partitions located on the root disk
  4219. automatically. Home partitions are recognized under their
  4220. GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
  4221. partitions are recognized under their GPT type ID
  4222. 0657fd6da4ab43c484e50933c84b4f4f.
  4223. * systemd will no longer pass any environment from the kernel
  4224. or initrd to system services. If you want to set an
  4225. environment for all services, do so via the kernel command
  4226. line systemd.setenv= assignment.
  4227. * The systemd-sysctl tool no longer natively reads the file
  4228. /etc/sysctl.conf. If desired, the file should be symlinked
  4229. from /etc/sysctl.d/99-sysctl.conf. Apart from providing
  4230. legacy support by a symlink rather than built-in code, it
  4231. also makes the otherwise hidden order of application of the
  4232. different files visible. (Note that this partly reverts to a
  4233. pre-198 application order of sysctl knobs!)
  4234. * The "systemctl set-log-level" and "systemctl dump" commands
  4235. have been moved to systemd-analyze.
  4236. * systemd-run learned the new --remain-after-exit switch,
  4237. which causes the scope unit not to be cleaned up
  4238. automatically after the process terminated.
  4239. * tmpfiles learned a new --exclude-prefix= switch to exclude
  4240. certain paths from operation.
  4241. * journald will now automatically flush all messages to disk
  4242. as soon as a message at the log level CRIT, ALERT or EMERG
  4243. is received.
  4244. Contributions from: Andrew Cook, Brandon Philips, Christian
  4245. Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
  4246. Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
  4247. McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
  4248. Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
  4249. Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
  4250. Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
  4251. Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
  4252. Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
  4253. Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
  4254. Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
  4255. Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
  4256. William Giokas, Zbigniew Jędrzejewski-Szmek
  4257. — Berlin, 2013-09-13
  4258. CHANGES WITH 206:
  4259. * The documentation has been updated to cover the various new
  4260. concepts introduced with 205.
  4261. * Unit files now understand the new %v specifier which
  4262. resolves to the kernel version string as returned by "uname
  4263. -r".
  4264. * systemctl now supports filtering the unit list output by
  4265. load state, active state and sub state, using the new
  4266. --state= parameter.
  4267. * "systemctl status" will now show the results of the
  4268. condition checks (like ConditionPathExists= and similar) of
  4269. the last start attempts of the unit. They are also logged to
  4270. the journal.
  4271. * "journalctl -b" may now be used to look for boot output of a
  4272. specific boot. Try "journalctl -b -1" for the previous boot,
  4273. but the syntax is substantially more powerful.
  4274. * "journalctl --show-cursor" has been added which prints the
  4275. cursor string the last shown log line. This may then be used
  4276. with the new "journalctl --after-cursor=" switch to continue
  4277. browsing logs from that point on.
  4278. * "journalctl --force" may now be used to force regeneration
  4279. of an FSS key.
  4280. * Creation of "dead" device nodes has been moved from udev
  4281. into kmod and tmpfiles. Previously, udev would read the kmod
  4282. databases to pre-generate dead device nodes based on meta
  4283. information contained in kernel modules, so that these would
  4284. be auto-loaded on access rather then at boot. As this
  4285. does not really have much to do with the exposing actual
  4286. kernel devices to userspace this has always been slightly
  4287. alien in the udev codebase. Following the new scheme kmod
  4288. will now generate a runtime snippet for tmpfiles from the
  4289. module meta information and it now is tmpfiles' job to the
  4290. create the nodes. This also allows overriding access and
  4291. other parameters for the nodes using the usual tmpfiles
  4292. facilities. As side effect this allows us to remove the
  4293. CAP_SYS_MKNOD capability bit from udevd entirely.
  4294. * logind's device ACLs may now be applied to these "dead"
  4295. devices nodes too, thus finally allowing managed access to
  4296. devices such as /dev/snd/sequencer without loading the
  4297. backing module right-away.
  4298. * A new RPM macro has been added that may be used to apply
  4299. tmpfiles configuration during package installation.
  4300. * systemd-detect-virt and ConditionVirtualization= now can
  4301. detect User-Mode-Linux machines (UML).
  4302. * journald will now implicitly log the effective capabilities
  4303. set of processes in the message metadata.
  4304. * systemd-cryptsetup has gained support for TrueCrypt volumes.
  4305. * The initrd interface has been simplified (more specifically,
  4306. support for passing performance data via environment
  4307. variables and fsck results via files in /run has been
  4308. removed). These features were non-essential, and are
  4309. nowadays available in a much nicer way by having systemd in
  4310. the initrd serialize its state and have the hosts systemd
  4311. deserialize it again.
  4312. * The udev "keymap" data files and tools to apply keyboard
  4313. specific mappings of scan to key codes, and force-release
  4314. scan code lists have been entirely replaced by a udev
  4315. "keyboard" builtin and a hwdb data file.
  4316. * systemd will now honour the kernel's "quiet" command line
  4317. argument also during late shutdown, resulting in a
  4318. completely silent shutdown when used.
  4319. * There's now an option to control the SO_REUSEPORT socket
  4320. option in .socket units.
  4321. * Instance units will now automatically get a per-template
  4322. subslice of system.slice unless something else is explicitly
  4323. configured. For example, instances of sshd@.service will now
  4324. implicitly be placed in system-sshd.slice rather than
  4325. system.slice as before.
  4326. * Test coverage support may now be enabled at build time.
  4327. Contributions from: Dave Reisner, Frederic Crozat, Harald
  4328. Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
  4329. Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
  4330. Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
  4331. Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
  4332. Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
  4333. Giokas, Zbigniew Jędrzejewski-Szmek
  4334. — Berlin, 2013-07-23
  4335. CHANGES WITH 205:
  4336. * Two new unit types have been introduced:
  4337. Scope units are very similar to service units, however, are
  4338. created out of pre-existing processes — instead of PID 1
  4339. forking off the processes. By using scope units it is
  4340. possible for system services and applications to group their
  4341. own child processes (worker processes) in a powerful way
  4342. which then maybe used to organize them, or kill them
  4343. together, or apply resource limits on them.
  4344. Slice units may be used to partition system resources in an
  4345. hierarchical fashion and then assign other units to them. By
  4346. default there are now three slices: system.slice (for all
  4347. system services), user.slice (for all user sessions),
  4348. machine.slice (for VMs and containers).
  4349. Slices and scopes have been introduced primarily in
  4350. context of the work to move cgroup handling to a
  4351. single-writer scheme, where only PID 1
  4352. creates/removes/manages cgroups.
  4353. * There's a new concept of "transient" units. In contrast to
  4354. normal units these units are created via an API at runtime,
  4355. not from configuration from disk. More specifically this
  4356. means it is now possible to run arbitrary programs as
  4357. independent services, with all execution parameters passed
  4358. in via bus APIs rather than read from disk. Transient units
  4359. make systemd substantially more dynamic then it ever was,
  4360. and useful as a general batch manager.
  4361. * logind has been updated to make use of scope and slice units
  4362. for managing user sessions. As a user logs in he will get
  4363. his own private slice unit, to which all sessions are added
  4364. as scope units. We also added support for automatically
  4365. adding an instance of user@.service for the user into the
  4366. slice. Effectively logind will no longer create cgroup
  4367. hierarchies on its own now, it will defer entirely to PID 1
  4368. for this by means of scope, service and slice units. Since
  4369. user sessions this way become entities managed by PID 1
  4370. the output of "systemctl" is now a lot more comprehensive.
  4371. * A new mini-daemon "systemd-machined" has been added which
  4372. may be used by virtualization managers to register local
  4373. VMs/containers. nspawn has been updated accordingly, and
  4374. libvirt will be updated shortly. machined will collect a bit
  4375. of meta information about the VMs/containers, and assign
  4376. them their own scope unit (see above). The collected
  4377. meta-data is then made available via the "machinectl" tool,
  4378. and exposed in "ps" and similar tools. machined/machinectl
  4379. is compile-time optional.
  4380. * As discussed earlier, the low-level cgroup configuration
  4381. options ControlGroup=, ControlGroupModify=,
  4382. ControlGroupPersistent=, ControlGroupAttribute= have been
  4383. removed. Please use high-level attribute settings instead as
  4384. well as slice units.
  4385. * A new bus call SetUnitProperties() has been added to alter
  4386. various runtime parameters of a unit. This is primarily
  4387. useful to alter cgroup parameters dynamically in a nice way,
  4388. but will be extended later on to make more properties
  4389. modifiable at runtime. systemctl gained a new set-properties
  4390. command that wraps this call.
  4391. * A new tool "systemd-run" has been added which can be used to
  4392. run arbitrary command lines as transient services or scopes,
  4393. while configuring a number of settings via the command
  4394. line. This tool is currently very basic, however already
  4395. very useful. We plan to extend this tool to even allow
  4396. queuing of execution jobs with time triggers from the
  4397. command line, similar in fashion to "at".
  4398. * nspawn will now inform the user explicitly that kernels with
  4399. audit enabled break containers, and suggest the user to turn
  4400. off audit.
  4401. * Support for detecting the IMA and AppArmor security
  4402. frameworks with ConditionSecurity= has been added.
  4403. * journalctl gained a new "-k" switch for showing only kernel
  4404. messages, mimicking dmesg output; in addition to "--user"
  4405. and "--system" switches for showing only user's own logs
  4406. and system logs.
  4407. * systemd-delta can now show information about drop-in
  4408. snippets extending unit files.
  4409. * libsystemd-bus has been substantially updated but is still
  4410. not available as public API.
  4411. * systemd will now look for the "debug" argument on the kernel
  4412. command line and enable debug logging, similar to what
  4413. "systemd.log_level=debug" already did before.
  4414. * "systemctl set-default", "systemctl get-default" has been
  4415. added to configure the symlink, which
  4416. controls what to boot into by default.
  4417. * "systemctl set-log-level" has been added as a convenient
  4418. way to raise and lower systemd logging threshold.
  4419. * "systemd-analyze plot" will now show the time the various
  4420. generators needed for execution, as well as information
  4421. about the unit file loading.
  4422. * libsystemd-journal gained a new sd_journal_open_files() call
  4423. for opening specific journal files. journactl also gained a
  4424. new switch to expose this new functionality. Previously we
  4425. only supported opening all files from a directory, or all
  4426. files from the system, as opening individual files only is
  4427. racy due to journal file rotation.
  4428. * systemd gained the new DefaultEnvironment= setting in
  4429. /etc/systemd/system.conf to set environment variables for
  4430. all services.
  4431. * If a privileged process logs a journal message with the
  4432. OBJECT_PID= field set, then journald will automatically
  4433. augment this with additional OBJECT_UID=, OBJECT_GID=,
  4434. OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
  4435. system services want to log events about specific client
  4436. processes. journactl/systemctl has been updated to make use
  4437. of this information if all log messages regarding a specific
  4438. unit is requested.
  4439. Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
  4440. Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
  4441. Reisner, David Coppa, David King, David Strauss, Eelco
  4442. Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
  4443. Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
  4444. Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
  4445. Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
  4446. Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
  4447. Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
  4448. Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
  4449. Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
  4450. Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
  4451. Łukasz Stelmach, 장동준
  4452. CHANGES WITH 204:
  4453. * The Python bindings gained some minimal support for the APIs
  4454. exposed by libsystemd-logind.
  4455. * ConditionSecurity= gained support for detecting SMACK. Since
  4456. this condition already supports SELinux and AppArmor we only
  4457. miss IMA for this. Patches welcome!
  4458. Contributions from: Karol Lewandowski, Lennart Poettering,
  4459. Zbigniew Jędrzejewski-Szmek
  4460. CHANGES WITH 203:
  4461. * systemd-nspawn will now create /etc/resolv.conf if
  4462. necessary, before bind-mounting the host's file onto it.
  4463. * systemd-nspawn will now store meta information about a
  4464. container on the container's cgroup as extended attribute
  4465. fields, including the root directory.
  4466. * The cgroup hierarchy has been reworked in many ways. All
  4467. objects any of the components systemd creates in the cgroup
  4468. tree are now suffixed. More specifically, user sessions are
  4469. now placed in cgroups suffixed with ".session", users in
  4470. cgroups suffixed with ".user", and nspawn containers in
  4471. cgroups suffixed with ".nspawn". Furthermore, all cgroup
  4472. names are now escaped in a simple scheme to avoid collision
  4473. of userspace object names with kernel filenames. This work
  4474. is preparation for making these objects relocatable in the
  4475. cgroup tree, in order to allow easy resource partitioning of
  4476. these objects without causing naming conflicts.
  4477. * systemctl list-dependencies gained the new switches
  4478. --plain, --reverse, --after and --before.
  4479. * systemd-inhibit now shows the process name of processes that
  4480. have taken an inhibitor lock.
  4481. * nss-myhostname will now also resolve "localhost"
  4482. implicitly. This makes /etc/hosts an optional file and
  4483. nicely handles that on IPv6 ::1 maps to both "localhost" and
  4484. the local hostname.
  4485. * gained a new call
  4486. sd_get_machine_names() to enumerate running containers and
  4487. VMs (currently only supported by very new libvirt and
  4488. nspawn). sd_login_monitor can now be used to watch
  4489. VMs/containers coming and going.
  4490. * .include is not allowed recursively anymore, and only in
  4491. unit files. Usually it is better to use drop-in snippets in
  4492. .d/*.conf anyway, as introduced with systemd 198.
  4493. * systemd-analyze gained a new "critical-chain" command that
  4494. determines the slowest chain of units run during system
  4495. boot-up. It is very useful for tracking down where
  4496. optimizing boot time is the most beneficial.
  4497. * systemd will no longer allow manipulating service paths in
  4498. the name=systemd:/system cgroup tree using ControlGroup= in
  4499. units. (But is still fine with it in all other dirs.)
  4500. * There's a new systemd-nspawn@.service service file that may
  4501. be used to easily run nspawn containers as system
  4502. services. With the container's root directory in
  4503. /var/lib/container/foobar it is now sufficient to run
  4504. "systemctl start systemd-nspawn@foobar.service" to boot it.
  4505. * systemd-cgls gained a new parameter "--machine" to list only
  4506. the processes within a certain container.
  4507. * ConditionSecurity= now can check for "apparmor". We still
  4508. are lacking checks for SMACK and IMA for this condition
  4509. check though. Patches welcome!
  4510. * A new configuration file /etc/systemd/sleep.conf has been
  4511. added that may be used to configure which kernel operation
  4512. systemd is supposed to execute when "suspend", "hibernate"
  4513. or "hybrid-sleep" is requested. This makes the new kernel
  4514. "freeze" state accessible to the user.
  4515. * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
  4516. the passed argument if applicable.
  4517. Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
  4518. Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
  4519. Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
  4520. Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
  4521. MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
  4522. Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
  4523. Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
  4524. Jędrzejewski-Szmek
  4525. CHANGES WITH 202:
  4526. * The output of 'systemctl list-jobs' got some polishing. The
  4527. '--type=' argument may now be passed more than once. A new
  4528. command 'systemctl list-sockets' has been added which shows
  4529. a list of kernel sockets systemd is listening on with the
  4530. socket units they belong to, plus the units these socket
  4531. units activate.
  4532. * The experimental libsystemd-bus library got substantial
  4533. updates to work in conjunction with the (also experimental)
  4534. kdbus kernel project. It works well enough to exchange
  4535. messages with some sophistication. Note that kdbus is not
  4536. ready yet, and the library is mostly an elaborate test case
  4537. for now, and not installable.
  4538. * systemd gained a new unit 'systemd-static-nodes.service'
  4539. that generates static device nodes earlier during boot, and
  4540. can run in conjunction with udev.
  4541. * libsystemd-login gained a new call sd_pid_get_user_unit()
  4542. to retrieve the user systemd unit a process is running
  4543. in. This is useful for systems where systemd is used as
  4544. session manager.
  4545. * systemd-nspawn now places all containers in the new /machine
  4546. top-level cgroup directory in the name=systemd
  4547. hierarchy. libvirt will soon do the same, so that we get a
  4548. uniform separation of /system, /user and /machine for system
  4549. services, user processes and containers/virtual
  4550. machines. This new cgroup hierarchy is also useful to stick
  4551. stable names to specific container instances, which can be
  4552. recognized later this way (this name may be controlled
  4553. via systemd-nspawn's new -M switch). libsystemd-login also
  4554. gained a new call sd_pid_get_machine_name() to retrieve the
  4555. name of the container/VM a specific process belongs to.
  4556. * bootchart can now store its data in the journal.
  4557. * libsystemd-journal gained a new call
  4558. sd_journal_add_conjunction() for AND expressions to the
  4559. matching logic. This can be used to express more complex
  4560. logical expressions.
  4561. * journactl can now take multiple --unit= and --user-unit=
  4562. switches.
  4563. * The cryptsetup logic now understands the "luks.key=" kernel
  4564. command line switch for specifying a file to read the
  4565. decryption key from. Also, if a configured key file is not
  4566. found the tool will now automatically fall back to prompting
  4567. the user.
  4568. * Python systemd.journal module was updated to wrap recently
  4569. added functions from libsystemd-journal. The interface was
  4570. changed to bring the low level interface in s.j._Reader
  4571. closer to the C API, and the high level interface in
  4572. s.j.Reader was updated to wrap and convert all data about
  4573. an entry.
  4574. Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
  4575. Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
  4576. Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
  4577. Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
  4578. Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
  4579. Tom Gundersen, Zbigniew Jędrzejewski-Szmek
  4580. CHANGES WITH 201:
  4581. * journalctl --update-catalog now understands a new --root=
  4582. option to operate on catalogs found in a different root
  4583. directory.
  4584. * During shutdown after systemd has terminated all running
  4585. services a final killing loop kills all remaining left-over
  4586. processes. We will now print the name of these processes
  4587. when we send SIGKILL to them, since this usually indicates a
  4588. problem.
  4589. * If /etc/crypttab refers to password files stored on
  4590. configured mount points automatic dependencies will now be
  4591. generated to ensure the specific mount is established first
  4592. before the key file is attempted to be read.
  4593. * 'systemctl status' will now show information about the
  4594. network sockets a socket unit is listening on.
  4595. * 'systemctl status' will also shown information about any
  4596. drop-in configuration file for units. (Drop-In configuration
  4597. files in this context are files such as
  4598. /etc/systemd/systemd/foobar.service.d/*.conf)
  4599. * systemd-cgtop now optionally shows summed up CPU times of
  4600. cgroups. Press '%' while running cgtop to switch between
  4601. percentage and absolute mode. This is useful to determine
  4602. which cgroups use up the most CPU time over the entire
  4603. runtime of the system. systemd-cgtop has also been updated
  4604. to be 'pipeable' for processing with further shell tools.
  4605. * 'hostnamectl set-hostname' will now allow setting of FQDN
  4606. hostnames.
  4607. * The formatting and parsing of time span values has been
  4608. changed. The parser now understands fractional expressions
  4609. such as "5.5h". The formatter will now output fractional
  4610. expressions for all time spans under 1min, i.e. "5.123456s"
  4611. rather than "5s 123ms 456us". For time spans under 1s
  4612. millisecond values are shown, for those under 1ms
  4613. microsecond values are shown. This should greatly improve
  4614. all time-related output of systemd.
  4615. * libsystemd-login and libsystemd-journal gained new
  4616. functions for querying the poll() events mask and poll()
  4617. timeout value for integration into arbitrary event
  4618. loops.
  4619. * localectl gained the ability to list available X11 keymaps
  4620. (models, layouts, variants, options).
  4621. * 'systemd-analyze dot' gained the ability to filter for
  4622. specific units via shell-style globs, to create smaller,
  4623. more useful graphs. I.e. it is now possible to create simple
  4624. graphs of all the dependencies between only target units, or
  4625. of all units that Avahi has dependencies with.
  4626. Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
  4627. Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
  4628. Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
  4629. Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
  4630. Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
  4631. Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
  4632. Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
  4633. CHANGES WITH 200:
  4634. * The boot-time readahead implementation for rotating media
  4635. will now read the read-ahead data in multiple passes which
  4636. consist of all read requests made in equidistant time
  4637. intervals. This means instead of strictly reading read-ahead
  4638. data in its physical order on disk we now try to find a
  4639. middle ground between physical and access time order.
  4640. * /etc/os-release files gained a new BUILD_ID= field for usage
  4641. on operating systems that provide continuous builds of OS
  4642. images.
  4643. Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
  4644. Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
  4645. William Douglas, Zbigniew Jędrzejewski-Szmek
  4646. CHANGES WITH 199:
  4647. * systemd-python gained an API exposing libsystemd-daemon.
  4648. * The SMACK setup logic gained support for uploading CIPSO
  4649. security policy.
  4650. * Behaviour of PrivateTmp=, ReadWriteDirectories=,
  4651. ReadOnlyDirectories= and InaccessibleDirectories= has
  4652. changed. The private /tmp and /var/tmp directories are now
  4653. shared by all processes of a service (which means
  4654. ExecStartPre= may now leave data in /tmp that ExecStart= of
  4655. the same service can still access). When a service is
  4656. stopped its temporary directories are immediately deleted
  4657. (normal clean-up with tmpfiles is still done in addition to
  4658. this though).
  4659. * By default, systemd will now set a couple of sysctl
  4660. variables in the kernel: the safe sysrq options are turned
  4661. on, IP route verification is turned on, and source routing
  4662. disabled. The recently added hardlink and softlink
  4663. protection of the kernel is turned on. These settings should
  4664. be reasonably safe, and good defaults for all new systems.
  4665. * The predictable network naming logic may now be turned off
  4666. with a new kernel command line switch: net.ifnames=0.
  4667. * A new libsystemd-bus module has been added that implements a
  4668. pretty complete D-Bus client library. For details see:
  4670. * journald will now explicitly flush the journal files to disk
  4671. at the latest 5min after each write. The file will then also
  4672. be marked offline until the next write. This should increase
  4673. reliability in case of a crash. The synchronization delay
  4674. can be configured via SyncIntervalSec= in journald.conf.
  4675. * There's a new unit that can be used
  4676. to pull in specific services when at least one remote file
  4677. system is to be mounted.
  4678. * There are new targets and as
  4679. canonical targets to pull user timer and path units in
  4680. from. This complements with a similar
  4681. purpose for socket units.
  4682. * libudev gained a new call udev_device_set_attribute_value()
  4683. to set sysfs attributes of a device.
  4684. * The udev daemon now sets the default number of worker
  4685. processes executed in parallel based on the number of available
  4686. CPUs instead of the amount of available RAM. This is supposed
  4687. to provide a more reliable default and limit a too aggressive
  4688. parallelism for setups with 1000s of devices connected.
  4689. Contributions from: Auke Kok, Colin Walters, Cristian
  4690. Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
  4691. Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
  4692. Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
  4693. Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
  4694. Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
  4695. Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
  4696. Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
  4697. Zbigniew Jędrzejewski-Szmek
  4698. CHANGES WITH 198:
  4699. * Configuration of unit files may now be extended via drop-in
  4700. files without having to edit/override the unit files
  4701. themselves. More specifically, if the administrator wants to
  4702. change one value for a service file foobar.service he can
  4703. now do so by dropping in a configuration snippet into
  4704. /etc/systemd/system/foobar.service.d/*.conf. The unit logic
  4705. will load all these snippets and apply them on top of the
  4706. main unit configuration file, possibly extending or
  4707. overriding its settings. Using these drop-in snippets is
  4708. generally nicer than the two earlier options for changing
  4709. unit files locally: copying the files from
  4710. /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
  4711. them there; or creating a new file in /etc/systemd/system/
  4712. that incorporates the original one via ".include". Drop-in
  4713. snippets into these .d/ directories can be placed in any
  4714. directory systemd looks for units in, and the usual
  4715. overriding semantics between /usr/lib, /etc and /run apply
  4716. for them too.
  4717. * Most unit file settings which take lists of items can now be
  4718. reset by assigning the empty string to them. For example,
  4719. normally, settings such as Environment=FOO=BAR append a new
  4720. environment variable assignment to the environment block,
  4721. each time they are used. By assigning Environment= the empty
  4722. string the environment block can be reset to empty. This is
  4723. particularly useful with the .d/*.conf drop-in snippets
  4724. mentioned above, since this adds the ability to reset list
  4725. settings from vendor unit files via these drop-ins.
  4726. * systemctl gained a new "list-dependencies" command for
  4727. listing the dependencies of a unit recursively.
  4728. * Inhibitors are now honored and listed by "systemctl
  4729. suspend", "systemctl poweroff" (and similar) too, not only
  4730. GNOME. These commands will also list active sessions by
  4731. other users.
  4732. * Resource limits (as exposed by the various control group
  4733. controllers) can now be controlled dynamically at runtime
  4734. for all units. More specifically, you can now use a command
  4735. like "systemctl set-cgroup-attr foobar.service cpu.shares
  4736. 2000" to alter the CPU shares a specific service gets. These
  4737. settings are stored persistently on disk, and thus allow the
  4738. administrator to easily adjust the resource usage of
  4739. services with a few simple commands. This dynamic resource
  4740. management logic is also available to other programs via the
  4741. bus. Almost any kernel cgroup attribute and controller is
  4742. supported.
  4743. * systemd-vconsole-setup will now copy all font settings to
  4744. all allocated VTs, where it previously applied them only to
  4745. the foreground VT.
  4746. * libsystemd-login gained the new sd_session_get_tty() API
  4747. call.
  4748. * This release drops support for a few legacy or
  4749. distribution-specific LSB facility names when parsing init
  4750. scripts: $x-display-manager, $mail-transfer-agent,
  4751. $mail-transport-agent, $mail-transfer-agent, $smtp,
  4752. $null. Also, the unit backing
  4753. this has been removed. Distributions which want to retain
  4754. compatibility with this should carry the burden for
  4755. supporting this themselves and patch support for these back
  4756. in, if they really need to. Also, the facilities $syslog and
  4757. $local_fs are now ignored, since systemd does not support
  4758. early-boot LSB init scripts anymore, and these facilities
  4759. are implied anyway for normal services. has
  4760. also been removed.
  4761. * There are new bus calls on PID1's Manager object for
  4762. cancelling jobs, and removing snapshot units. Previously,
  4763. both calls were only available on the Job and Snapshot
  4764. objects themselves.
  4765. * systemd-journal-gatewayd gained SSL support.
  4766. * The various "environment" files, such as /etc/locale.conf
  4767. now support continuation lines with a backslash ("\") as
  4768. last character in the line, similarly in style (but different)
  4769. to how this is supported in shells.
  4770. * For normal user processes the _SYSTEMD_USER_UNIT= field is
  4771. now implicitly appended to every log entry logged. systemctl
  4772. has been updated to filter by this field when operating on a
  4773. user systemd instance.
  4774. * nspawn will now implicitly add the CAP_AUDIT_WRITE and
  4775. CAP_AUDIT_CONTROL capabilities to the capabilities set for
  4776. the container. This makes it easier to boot unmodified
  4777. Fedora systems in a container, which however still requires
  4778. audit=0 to be passed on the kernel command line. Auditing in
  4779. kernel and userspace is unfortunately still too broken in
  4780. context of containers, hence we recommend compiling it out
  4781. of the kernel or using audit=0. Hopefully this will be fixed
  4782. one day for good in the kernel.
  4783. * nspawn gained the new --bind= and --bind-ro= parameters to
  4784. bind mount specific directories from the host into the
  4785. container.
  4786. * nspawn will now mount its own devpts file system instance
  4787. into the container, in order not to leak pty devices from
  4788. the host into the container.
  4789. * systemd will now read the firmware boot time performance
  4790. information from the EFI variables, if the used boot loader
  4791. supports this, and takes it into account for boot performance
  4792. analysis via "systemd-analyze". This is currently supported
  4793. only in conjunction with Gummiboot, but could be supported
  4794. by other boot loaders too. For details see:
  4796. * A new generator has been added that automatically mounts the
  4797. EFI System Partition (ESP) to /boot, if that directory
  4798. exists, is empty, and no other file system has been
  4799. configured to be mounted there.
  4800. * logind will now send out PrepareForSleep(false) out
  4801. unconditionally, after coming back from suspend. This may be
  4802. used by applications as asynchronous notification for
  4803. system resume events.
  4804. * "systemctl unlock-sessions" has been added, that allows
  4805. unlocking the screens of all user sessions at once, similar
  4806. to how "systemctl lock-sessions" already locked all users
  4807. sessions. This is backed by a new D-Bus call UnlockSessions().
  4808. * "loginctl seat-status" will now show the master device of a
  4809. seat. (i.e. the device of a seat that needs to be around for
  4810. the seat to be considered available, usually the graphics
  4811. card).
  4812. * tmpfiles gained a new "X" line type, that allows
  4813. configuration of files and directories (with wildcards) that
  4814. shall be excluded from automatic cleanup ("aging").
  4815. * udev default rules set the device node permissions now only
  4816. at "add" events, and do not change them any longer with a
  4817. later "change" event.
  4818. * The log messages for lid events and power/sleep keypresses
  4819. now carry a message ID.
  4820. * We now have a substantially larger unit test suite, but this
  4821. continues to be work in progress.
  4822. * udevadm hwdb gained a new --root= parameter to change the
  4823. root directory to operate relative to.
  4824. * logind will now issue a background sync() request to the kernel
  4825. early at shutdown, so that dirty buffers are flushed to disk early
  4826. instead of at the last moment, in order to optimize shutdown
  4827. times a little.
  4828. * A new bootctl tool has been added that is an interface for
  4829. certain boot loader operations. This is currently a preview
  4830. and is likely to be extended into a small mechanism daemon
  4831. like timedated, localed, hostnamed, and can be used by
  4832. graphical UIs to enumerate available boot options, and
  4833. request boot into firmware operations.
  4834. * systemd-bootchart has been relicensed to LGPLv2.1+ to match
  4835. the rest of the package. It also has been updated to work
  4836. correctly in initrds.
  4837. * Policykit previously has been runtime optional, and is now
  4838. also compile time optional via a configure switch.
  4839. * systemd-analyze has been reimplemented in C. Also "systemctl
  4840. dot" has moved into systemd-analyze.
  4841. * "systemctl status" with no further parameters will now print
  4842. the status of all active or failed units.
  4843. * Operations such as "systemctl start" can now be executed
  4844. with a new mode "--irreversible" which may be used to queue
  4845. operations that cannot accidentally be reversed by a later
  4846. job queuing. This is by default used to make shutdown
  4847. requests more robust.
  4848. * The Python API of systemd now gained a new module for
  4849. reading journal files.
  4850. * A new tool kernel-install has been added that can install
  4851. kernel images according to the Boot Loader Specification:
  4853. * Boot time console output has been improved to provide
  4854. animated boot time output for hanging jobs.
  4855. * A new tool systemd-activate has been added which can be used
  4856. to test socket activation with, directly from the command
  4857. line. This should make it much easier to test and debug
  4858. socket activation in daemons.
  4859. * journalctl gained a new "--reverse" (or -r) option to show
  4860. journal output in reverse order (i.e. newest line first).
  4861. * journalctl gained a new "--pager-end" (or -e) option to jump
  4862. to immediately jump to the end of the journal in the
  4863. pager. This is only supported in conjunction with "less".
  4864. * journalctl gained a new "--user-unit=" option, that works
  4865. similarly to "--unit=" but filters for user units rather than
  4866. system units.
  4867. * A number of unit files to ease adoption of systemd in
  4868. initrds has been added. This moves some minimal logic from
  4869. the various initrd implementations into systemd proper.
  4870. * The journal files are now owned by a new group
  4871. "systemd-journal", which exists specifically to allow access
  4872. to the journal, and nothing else. Previously, we used the
  4873. "adm" group for that, which however possibly covers more
  4874. than just journal/log file access. This new group is now
  4875. already used by systemd-journal-gatewayd to ensure this
  4876. daemon gets access to the journal files and as little else
  4877. as possible. Note that "make install" will also set FS ACLs
  4878. up for /var/log/journal to give "adm" and "wheel" read
  4879. access to it, in addition to "systemd-journal" which owns
  4880. the journal files. We recommend that packaging scripts also
  4881. add read access to "adm" + "wheel" to /var/log/journal, and
  4882. all existing/future journal files. To normal users and
  4883. administrators little changes, however packagers need to
  4884. ensure to create the "systemd-journal" system group at
  4885. package installation time.
  4886. * The systemd-journal-gatewayd now runs as unprivileged user
  4887. systemd-journal-gateway:systemd-journal-gateway. Packaging
  4888. scripts need to create these system user/group at
  4889. installation time.
  4890. * timedated now exposes a new boolean property CanNTP that
  4891. indicates whether a local NTP service is available or not.
  4892. * systemd-detect-virt will now also detect xen PVs
  4893. * The pstore file system is now mounted by default, if it is
  4894. available.
  4895. * In addition to the SELinux and IMA policies we will now also
  4896. load SMACK policies at early boot.
  4897. Contributions from: Adel Gadllah, Aleksander Morgado, Auke
  4898. Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
  4899. Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
  4900. Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
  4901. Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
  4902. Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
  4903. Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
  4904. Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
  4905. Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
  4906. Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
  4907. Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
  4908. Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
  4909. Gundersen, Umut Tezduyar, William Giokas, Zbigniew
  4910. Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
  4911. CHANGES WITH 197:
  4912. * Timer units now support calendar time events in addition to
  4913. monotonic time events. That means you can now trigger a unit
  4914. based on a calendar time specification such as "Thu,Fri
  4915. 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
  4916. or fifth day of any month of the year 2013, given that it is
  4917. a thursday or friday. This brings timer event support
  4918. considerably closer to cron's capabilities. For details on
  4919. the supported calendar time specification language see
  4920. systemd.time(7).
  4921. * udev now supports a number of different naming policies for
  4922. network interfaces for predictable names, and a combination
  4923. of these policies is now the default. Please see this wiki
  4924. document for details:
  4926. * Auke Kok's bootchart implementation has been added to the
  4927. systemd tree. It is an optional component that can graph the
  4928. boot in quite some detail. It is one of the best bootchart
  4929. implementations around and minimal in its code and
  4930. dependencies.
  4931. * nss-myhostname has been integrated into the systemd source
  4932. tree. nss-myhostname guarantees that the local hostname
  4933. always stays resolvable via NSS. It has been a weak
  4934. requirement of systemd-hostnamed since a long time, and
  4935. since its code is actually trivial we decided to just
  4936. include it in systemd's source tree. It can be turned off
  4937. with a configure switch.
  4938. * The read-ahead logic is now capable of properly detecting
  4939. whether a btrfs file system is on SSD or rotating media, in
  4940. order to optimize the read-ahead scheme. Previously, it was
  4941. only capable of detecting this on traditional file systems
  4942. such as ext4.
  4943. * In udev, additional device properties are now read from the
  4944. IAB in addition to the OUI database. Also, Bluetooth company
  4945. identities are attached to the devices as well.
  4946. * In service files %U may be used as specifier that is
  4947. replaced by the configured user name of the service.
  4948. * nspawn may now be invoked without a controlling TTY. This
  4949. makes it suitable for invocation as its own service. This
  4950. may be used to set up a simple containerized server system
  4951. using only core OS tools.
  4952. * systemd and nspawn can now accept socket file descriptors
  4953. when they are started for socket activation. This enables
  4954. implementation of socket activated nspawn
  4955. containers. i.e. think about autospawning an entire OS image
  4956. when the first SSH or HTTP connection is received. We expect
  4957. that similar functionality will also be added to libvirt-lxc
  4958. eventually.
  4959. * journalctl will now suppress ANSI color codes when
  4960. presenting log data.
  4961. * systemctl will no longer show control group information for
  4962. a unit if the control group is empty anyway.
  4963. * logind can now automatically suspend/hibernate/shutdown the
  4964. system on idle.
  4965. * /etc/machine-info and hostnamed now also expose the chassis
  4966. type of the system. This can be used to determine whether
  4967. the local system is a laptop, desktop, handset or
  4968. tablet. This information may either be configured by the
  4969. user/vendor or is automatically determined from ACPI and DMI
  4970. information if possible.
  4971. * A number of PolicyKit actions are now bound together with
  4972. "imply" rules. This should simplify creating UIs because
  4973. many actions will now authenticate similar ones as well.
  4974. * Unit files learnt a new condition ConditionACPower= which
  4975. may be used to conditionalize a unit depending on whether an
  4976. AC power source is connected or not, of whether the system
  4977. is running on battery power.
  4978. * systemctl gained a new "is-failed" verb that may be used in
  4979. shell scripts and suchlike to check whether a specific unit
  4980. is in the "failed" state.
  4981. * The EnvironmentFile= setting in unit files now supports file
  4982. globbing, and can hence be used to easily read a number of
  4983. environment files at once.
  4984. * systemd will no longer detect and recognize specific
  4985. distributions. All distribution-specific #ifdeffery has been
  4986. removed, systemd is now fully generic and
  4987. distribution-agnostic. Effectively, not too much is lost as
  4988. a lot of the code is still accessible via explicit configure
  4989. switches. However, support for some distribution specific
  4990. legacy configuration file formats has been dropped. We
  4991. recommend distributions to simply adopt the configuration
  4992. files everybody else uses now and convert the old
  4993. configuration from packaging scripts. Most distributions
  4994. already did that. If that's not possible or desirable,
  4995. distributions are welcome to forward port the specific
  4996. pieces of code locally from the git history.
  4997. * When logging a message about a unit systemd will now always
  4998. log the unit name in the message meta data.
  4999. * localectl will now also discover system locale data that is
  5000. not stored in locale archives, but directly unpacked.
  5001. * logind will no longer unconditionally use framebuffer
  5002. devices as seat masters, i.e. as devices that are required
  5003. to be existing before a seat is considered preset. Instead,
  5004. it will now look for all devices that are tagged as
  5005. "seat-master" in udev. By default, framebuffer devices will
  5006. be marked as such, but depending on local systems, other
  5007. devices might be marked as well. This may be used to
  5008. integrate graphics cards using closed source drivers (such
  5009. as NVidia ones) more nicely into logind. Note however, that
  5010. we recommend using the open source NVidia drivers instead,
  5011. and no udev rules for the closed-source drivers will be
  5012. shipped from us upstream.
  5013. Contributions from: Adam Williamson, Alessandro Crismani, Auke
  5014. Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
  5015. Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
  5016. Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
  5017. Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
  5018. Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
  5019. Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
  5020. Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
  5021. Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
  5022. Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
  5023. Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
  5024. Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
  5025. Jędrzejewski-Szmek
  5026. CHANGES WITH 196:
  5027. * udev gained support for loading additional device properties
  5028. from an indexed database that is keyed by vendor/product IDs
  5029. and similar device identifiers. For the beginning this
  5030. "hwdb" is populated with data from the well-known PCI and
  5031. USB database, but also includes PNP, ACPI and OID data. In
  5032. the longer run this indexed database shall grow into
  5033. becoming the one central database for non-essential
  5034. userspace device metadata. Previously, data from the PCI/USB
  5035. database was only attached to select devices, since the
  5036. lookup was a relatively expensive operation due to O(n) time
  5037. complexity (with n being the number of entries in the
  5038. database). Since this is now O(1), we decided to add in this
  5039. data for all devices where this is available, by
  5040. default. Note that the indexed database needs to be rebuilt
  5041. when new data files are installed. To achieve this you need
  5042. to update your packaging scripts to invoke "udevadm hwdb
  5043. --update" after installation of hwdb data files. For
  5044. RPM-based distributions we introduced the new
  5045. %udev_hwdb_update macro for this purpose.
  5046. * The Journal gained support for the "Message Catalog", an
  5047. indexed database to link up additional information with
  5048. journal entries. For further details please check:
  5050. The indexed message catalog database also needs to be
  5051. rebuilt after installation of message catalog files. Use
  5052. "journalctl --update-catalog" for this. For RPM-based
  5053. distributions we introduced the %journal_catalog_update
  5054. macro for this purpose.
  5055. * The Python Journal bindings gained support for the standard
  5056. Python logging framework.
  5057. * The Journal API gained new functions for checking whether
  5058. the underlying file system of a journal file is capable of
  5059. properly reporting file change notifications, or whether
  5060. applications that want to reflect journal changes "live"
  5061. need to recheck journal files continuously in appropriate
  5062. time intervals.
  5063. * It is now possible to set the "age" field for tmpfiles
  5064. entries to 0, indicating that files matching this entry
  5065. shall always be removed when the directories are cleaned up.
  5066. * coredumpctl gained a new "gdb" verb which invokes gdb
  5067. right-away on the selected coredump.
  5068. * There's now support for "hybrid sleep" on kernels that
  5069. support this, in addition to "suspend" and "hibernate". Use
  5070. "systemctl hybrid-sleep" to make use of this.
  5071. * logind's HandleSuspendKey= setting (and related settings)
  5072. now gained support for a new "lock" setting to simply
  5073. request the screen lock on all local sessions, instead of
  5074. actually executing a suspend or hibernation.
  5075. * systemd will now mount the EFI variables file system by
  5076. default.
  5077. * Socket units now gained support for configuration of the
  5078. SMACK security label.
  5079. * timedatectl will now output the time of the last and next
  5080. daylight saving change.
  5081. * We dropped support for various legacy and distro-specific
  5082. concepts, such as insserv, early-boot SysV services
  5083. (i.e. those for non-standard runlevels such as 'b' or 'S')
  5084. or ArchLinux /etc/rc.conf support. We recommend the
  5085. distributions who still need support this to either continue
  5086. to maintain the necessary patches downstream, or find a
  5087. different solution. (Talk to us if you have questions!)
  5088. * Various systemd components will now bypass PolicyKit checks
  5089. for root and otherwise handle properly if PolicyKit is not
  5090. found to be around. This should fix most issues for
  5091. PolicyKit-less systems. Quite frankly this should have been
  5092. this way since day one. It is absolutely our intention to
  5093. make systemd work fine on PolicyKit-less systems, and we
  5094. consider it a bug if something does not work as it should if
  5095. PolicyKit is not around.
  5096. * For embedded systems it is now possible to build udev and
  5097. systemd without blkid and/or kmod support.
  5098. * "systemctl switch-root" is now capable of switching root
  5099. more than once. I.e. in addition to transitions from the
  5100. initrd to the host OS it is now possible to transition to
  5101. further OS images from the host. This is useful to implement
  5102. offline updating tools.
  5103. * Various other additions have been made to the RPM macros
  5104. shipped with systemd. Use %udev_rules_update() after
  5105. installing new udev rules files. %_udevhwdbdir,
  5106. %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
  5107. %_sysctldir are now available which resolve to the right
  5108. directories for packages to place various data files in.
  5109. * journalctl gained the new --full switch (in addition to
  5110. --all, to disable ellipsation for long messages.
  5111. Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
  5112. Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
  5113. Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
  5114. Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
  5115. Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
  5116. Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
  5117. Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
  5118. Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
  5119. Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
  5120. CHANGES WITH 195:
  5121. * journalctl gained new --since= and --until= switches to
  5122. filter by time. It also now supports nice filtering for
  5123. units via --unit=/-u.
  5124. * Type=oneshot services may use ExecReload= and do the
  5125. right thing.
  5126. * The journal daemon now supports time-based rotation and
  5127. vacuuming, in addition to the usual disk-space based
  5128. rotation.
  5129. * The journal will now index the available field values for
  5130. each field name. This enables clients to show pretty drop
  5131. downs of available match values when filtering. The bash
  5132. completion of journalctl has been updated
  5133. accordingly. journalctl gained a new switch -F to list all
  5134. values a certain field takes in the journal database.
  5135. * More service events are now written as structured messages
  5136. to the journal, and made recognizable via message IDs.
  5137. * The timedated, localed and hostnamed mini-services which
  5138. previously only provided support for changing time, locale
  5139. and hostname settings from graphical DEs such as GNOME now
  5140. also have a minimal (but very useful) text-based client
  5141. utility each. This is probably the nicest way to changing
  5142. these settings from the command line now, especially since
  5143. it lists available options and is fully integrated with bash
  5144. completion.
  5145. * There's now a new tool "systemd-coredumpctl" to list and
  5146. extract coredumps from the journal.
  5147. * We now install a README each in /var/log/ and
  5148. /etc/rc.d/init.d explaining where the system logs and init
  5149. scripts went. This hopefully should help folks who go to
  5150. that dirs and look into the otherwise now empty void and
  5151. scratch their heads.
  5152. * When user-services are invoked (by systemd --user) the
  5153. $MANAGERPID env var is set to the PID of systemd.
  5154. * SIGRTMIN+24 when sent to a --user instance will now result
  5155. in immediate termination of systemd.
  5156. * gatewayd received numerous feature additions such as a
  5157. "follow" mode, for live syncing and filtering.
  5158. * browse.html now allows filtering and showing detailed
  5159. information on specific entries. Keyboard navigation and
  5160. mouse screen support has been added.
  5161. * gatewayd/journalctl now supports HTML5/JSON
  5162. Server-Sent-Events as output.
  5163. * The SysV init script compatibility logic will now
  5164. heuristically determine whether a script supports the
  5165. "reload" verb, and only then make this available as
  5166. "systemctl reload".
  5167. * "systemctl status --follow" has been removed, use "journalctl
  5168. -u" instead.
  5169. * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
  5170. have been removed since they are hardly useful to be
  5171. configured.
  5172. * And I'd like to take the opportunity to specifically mention
  5173. Zbigniew for his great contributions. Zbigniew, you rock!
  5174. Contributions from: Andrew Eikum, Christian Hesse, Colin
  5175. Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
  5176. Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
  5177. Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
  5178. Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
  5179. Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
  5180. Jędrzejewski-Szmek, Сковорода Никита Андреевич
  5181. CHANGES WITH 194:
  5182. * If /etc/vconsole.conf is non-existent or empty we will no
  5183. longer load any console font or key map at boot by
  5184. default. Instead the kernel defaults will be left
  5185. intact. This is definitely the right thing to do, as no
  5186. configuration should mean no configuration, and hard-coding
  5187. font names that are different on all archs is probably a bad
  5188. idea. Also, the kernel default key map and font should be
  5189. good enough for most cases anyway, and mostly identical to
  5190. the userspace fonts/key maps we previously overloaded them
  5191. with. If distributions want to continue to default to a
  5192. non-kernel font or key map they should ship a default
  5193. /etc/vconsole.conf with the appropriate contents.
  5194. Contributions from: Colin Walters, Daniel J Walsh, Dave
  5195. Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
  5196. Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
  5197. CHANGES WITH 193:
  5198. * journalctl gained a new --cursor= switch to show entries
  5199. starting from the specified location in the journal.
  5200. * We now enforce a size limit on journal entry fields exported
  5201. with "-o json" in journalctl. Fields larger than 4K will be
  5202. assigned null. This can be turned off with --all.
  5203. * An (optional) journal gateway daemon is now available as
  5204. "systemd-journal-gatewayd.service". This service provides
  5205. access to the journal via HTTP and JSON. This functionality
  5206. will be used to implement live log synchronization in both
  5207. pull and push modes, but has various other users too, such
  5208. as easy log access for debugging of embedded devices. Right
  5209. now it is already useful to retrieve the journal via HTTP:
  5210. # systemctl start systemd-journal-gatewayd.service
  5211. # wget http://localhost:19531/entries
  5212. This will download the journal contents in a
  5213. /var/log/messages compatible format. The same as JSON:
  5214. # curl -H"Accept: application/json" http://localhost:19531/entries
  5215. This service is also accessible via a web browser where a
  5216. single static HTML5 app is served that uses the JSON logic
  5217. to enable the user to do some basic browsing of the
  5218. journal. This will be extended later on. Here's an example
  5219. screenshot of this app in its current state:
  5221. Contributions from: Kay Sievers, Lennart Poettering, Robert
  5222. Milasan, Tom Gundersen
  5223. CHANGES WITH 192:
  5224. * The bash completion logic is now available for journalctl
  5225. too.
  5226. * We do not mount the "cpuset" controller anymore together with
  5227. "cpu" and "cpuacct", as "cpuset" groups generally cannot be
  5228. started if no parameters are assigned to it. "cpuset" hence
  5229. broke code that assumed it could create "cpu" groups and
  5230. just start them.
  5231. * journalctl -f will now subscribe to terminal size changes,
  5232. and line break accordingly.
  5233. Contributions from: Dave Reisner, Kay Sievers, Lennart
  5234. Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
  5235. CHANGES WITH 191:
  5236. * nspawn will now create a symlink /etc/localtime in the
  5237. container environment, copying the host's timezone
  5238. setting. Previously this has been done via a bind mount, but
  5239. since symlinks cannot be bind mounted this has now been
  5240. changed to create/update the appropriate symlink.
  5241. * journalctl -n's line number argument is now optional, and
  5242. will default to 10 if omitted.
  5243. * journald will now log the maximum size the journal files may
  5244. take up on disk. This is particularly useful if the default
  5245. built-in logic of determining this parameter from the file
  5246. system size is used. Use "systemctl status
  5247. systemd-journald.service" to see this information.
  5248. * The multi-seat X wrapper tool has been stripped down. As X
  5249. is now capable of enumerating graphics devices via udev in a
  5250. seat-aware way the wrapper is not strictly necessary
  5251. anymore. A stripped down temporary stop-gap is still shipped
  5252. until the upstream display managers have been updated to
  5253. fully support the new X logic. Expect this wrapper to be
  5254. removed entirely in one of the next releases.
  5255. * HandleSleepKey= in logind.conf has been split up into
  5256. HandleSuspendKey= and HandleHibernateKey=. The old setting
  5257. is not available anymore. X11 and the kernel are
  5258. distinguishing between these keys and we should too. This
  5259. also means the inhibition lock for these keys has been split
  5260. into two.
  5261. Contributions from: Dave Airlie, Eelco Dolstra, Lennart
  5262. Poettering, Lukas Nykryn, Václav Pavlín
  5263. CHANGES WITH 190:
  5264. * Whenever a unit changes state we will now log this to the
  5265. journal and show along the unit's own log output in
  5266. "systemctl status".
  5267. * ConditionPathIsMountPoint= can now properly detect bind
  5268. mount points too. (Previously, a bind mount of one file
  5269. system to another place in the same file system could not be
  5270. detected as mount, since they shared struct stat's st_dev
  5271. field.)
  5272. * We will now mount the cgroup controllers cpu, cpuacct,
  5273. cpuset and the controllers net_cls, net_prio together by
  5274. default.
  5275. * nspawn containers will now have a virtualized boot
  5276. ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
  5277. over with a randomized ID at container initialization). This
  5278. has the effect of making "journalctl -b" do the right thing
  5279. in a container.
  5280. * The JSON output journal serialization has been updated not
  5281. to generate "endless" list objects anymore, but rather one
  5282. JSON object per line. This is more in line how most JSON
  5283. parsers expect JSON objects. The new output mode
  5284. "json-pretty" has been added to provide similar output, but
  5285. neatly aligned for readability by humans.
  5286. * We dropped all explicit sync() invocations in the shutdown
  5287. code. The kernel does this implicitly anyway in the kernel
  5288. reboot() syscall. halt(8)'s -n option is now a compatibility
  5289. no-op.
  5290. * We now support virtualized reboot() in containers, as
  5291. supported by newer kernels. We will fall back to exit() if
  5292. CAP_SYS_REBOOT is not available to the container. Also,
  5293. nspawn makes use of this now and will actually reboot the
  5294. container if the containerized OS asks for that.
  5295. * journalctl will only show local log output by default
  5296. now. Use --merge (-m) to show remote log output, too.
  5297. * libsystemd-journal gained the new sd_journal_get_usage()
  5298. call to determine the current disk usage of all journal
  5299. files. This is exposed in the new "journalctl --disk-usage"
  5300. command.
  5301. * journald gained a new configuration setting SplitMode= in
  5302. journald.conf which may be used to control how user journals
  5303. are split off. See journald.conf(5) for details.
  5304. * A new condition type ConditionFileNotEmpty= has been added.
  5305. * tmpfiles' "w" lines now support file globbing, to write
  5306. multiple files at once.
  5307. * We added Python bindings for the journal submission
  5308. APIs. More Python APIs for a number of selected APIs will
  5309. likely follow. Note that we intend to add native bindings
  5310. only for the Python language, as we consider it common
  5311. enough to deserve bindings shipped within systemd. There are
  5312. various projects outside of systemd that provide bindings
  5313. for languages such as PHP or Lua.
  5314. * Many conditions will now resolve specifiers such as %i. In
  5315. addition, PathChanged= and related directives of .path units
  5316. now support specifiers as well.
  5317. * There's now a new RPM macro definition for the system preset
  5318. dir: %_presetdir.
  5319. * journald will now warn if it ca not forward a message to the
  5320. syslog daemon because its socket is full.
  5321. * timedated will no longer write or process /etc/timezone,
  5322. except on Debian. As we do not support late mounted /usr
  5323. anymore /etc/localtime always being a symlink is now safe,
  5324. and hence the information in /etc/timezone is not necessary
  5325. anymore.
  5326. * logind will now always reserve one VT for a text getty (VT6
  5327. by default). Previously if more than 6 X sessions where
  5328. started they took up all the VTs with auto-spawned gettys,
  5329. so that no text gettys were available anymore.
  5330. * udev will now automatically inform the btrfs kernel logic
  5331. about btrfs RAID components showing up. This should make
  5332. simple hotplug based btrfs RAID assembly work.
  5333. * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
  5334. (but not for its children which will stay at the kernel
  5335. default). This should allow setups with a lot more listening
  5336. sockets.
  5337. * systemd will now always pass the configured timezone to the
  5338. kernel at boot. timedated will do the same when the timezone
  5339. is changed.
  5340. * logind's inhibition logic has been updated. By default,
  5341. logind will now handle the lid switch, the power and sleep
  5342. keys all the time, even in graphical sessions. If DEs want
  5343. to handle these events on their own they should take the new
  5344. handle-power-key, handle-sleep-key and handle-lid-switch
  5345. inhibitors during their runtime. A simple way to achieve
  5346. that is to invoke the DE wrapped in an invocation of:
  5347. systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
  5348. * Access to unit operations is now checked via SELinux taking
  5349. the unit file label and client process label into account.
  5350. * systemd will now notify the administrator in the journal
  5351. when he over-mounts a non-empty directory.
  5352. * There are new specifiers that are resolved in unit files,
  5353. for the host name (%H), the machine ID (%m) and the boot ID
  5354. (%b).
  5355. Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
  5356. Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
  5357. Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
  5358. Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
  5359. Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
  5360. Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
  5361. Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
  5362. CHANGES WITH 189:
  5363. * Support for reading structured kernel messages from
  5364. /dev/kmsg has now been added and is enabled by default.
  5365. * Support for reading kernel messages from /proc/kmsg has now
  5366. been removed. If you want kernel messages in the journal
  5367. make sure to run a recent kernel (>= 3.5) that supports
  5368. reading structured messages from /dev/kmsg (see
  5369. above). /proc/kmsg is now exclusive property of classic
  5370. syslog daemons again.
  5371. * The libudev API gained the new
  5372. udev_device_new_from_device_id() call.
  5373. * The logic for file system namespace (ReadOnlyDirectory=,
  5374. ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
  5375. require pivot_root() anymore. This means fewer temporary
  5376. directories are created below /tmp for this feature.
  5377. * nspawn containers will now see and receive all submounts
  5378. made on the host OS below the root file system of the
  5379. container.
  5380. * Forward Secure Sealing is now supported for Journal files,
  5381. which provide cryptographical sealing of journal files so
  5382. that attackers cannot alter log history anymore without this
  5383. being detectable. Lennart will soon post a blog story about
  5384. this explaining it in more detail.
  5385. * There are two new service settings RestartPreventExitStatus=
  5386. and SuccessExitStatus= which allow configuration of exit
  5387. status (exit code or signal) which will be excepted from the
  5388. restart logic, resp. consider successful.
  5389. * journalctl gained the new --verify switch that can be used
  5390. to check the integrity of the structure of journal files and
  5391. (if Forward Secure Sealing is enabled) the contents of
  5392. journal files.
  5393. * nspawn containers will now be run with /dev/stdin, /dev/fd/
  5394. and similar symlinks pre-created. This makes running shells
  5395. as container init process a lot more fun.
  5396. * The fstab support can now handle PARTUUID= and PARTLABEL=
  5397. entries.
  5398. * A new ConditionHost= condition has been added to match
  5399. against the hostname (with globs) and machine ID. This is
  5400. useful for clusters where a single OS image is used to
  5401. provision a large number of hosts which shall run slightly
  5402. different sets of services.
  5403. * Services which hit the restart limit will now be placed in a
  5404. failure state.
  5405. Contributions from: Bertram Poettering, Dave Reisner, Huang
  5406. Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
  5407. Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
  5408. CHANGES WITH 188:
  5409. * When running in --user mode systemd will now become a
  5410. subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
  5411. tree a lot more organized.
  5412. * A new PartOf= unit dependency type has been introduced that
  5413. may be used to group services in a natural way.
  5414. * "systemctl enable" may now be used to enable instances of
  5415. services.
  5416. * journalctl now prints error log levels in red, and
  5417. warning/notice log levels in bright white. It also supports
  5418. filtering by log level now.
  5419. * cgtop gained a new -n switch (similar to top), to configure
  5420. the maximum number of iterations to run for. It also gained
  5421. -b, to run in batch mode (accepting no input).
  5422. * The suffix ".service" may now be omitted on most systemctl
  5423. command lines involving service unit names.
  5424. * There's a new bus call in logind to lock all sessions, as
  5425. well as a loginctl verb for it "lock-sessions".
  5426. * gained a new call sd_journal_perror()
  5427. that works similar to libc perror() but logs to the journal
  5428. and encodes structured information about the error number.
  5429. * /etc/crypttab entries now understand the new keyfile-size=
  5430. option.
  5431. * shutdown(8) now can send a (configurable) wall message when
  5432. a shutdown is cancelled.
  5433. * The mount propagation mode for the root file system will now
  5434. default to "shared", which is useful to make containers work
  5435. nicely out-of-the-box so that they receive new mounts from
  5436. the host. This can be undone locally by running "mount
  5437. --make-rprivate /" if needed.
  5438. * The prefdm.service file has been removed. Distributions
  5439. should maintain this unit downstream if they intend to keep
  5440. it around. However, we recommend writing normal unit files
  5441. for display managers instead.
  5442. * Since systemd is a crucial part of the OS we will now
  5443. default to a number of compiler switches that improve
  5444. security (hardening) such as read-only relocations, stack
  5445. protection, and suchlike.
  5446. * The TimeoutSec= setting for services is now split into
  5447. TimeoutStartSec= and TimeoutStopSec= to allow configuration
  5448. of individual time outs for the start and the stop phase of
  5449. the service.
  5450. Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
  5451. Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
  5452. Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
  5453. Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
  5454. Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
  5455. Gundersen, Zbigniew Jędrzejewski-Szmek
  5456. CHANGES WITH 187:
  5457. * The journal and id128 C APIs are now fully documented as man
  5458. pages.
  5459. * Extra safety checks have been added when transitioning from
  5460. the initial RAM disk to the main system to avoid accidental
  5461. data loss.
  5462. * /etc/crypttab entries now understand the new keyfile-offset=
  5463. option.
  5464. * systemctl -t can now be used to filter by unit load state.
  5465. * The journal C API gained the new sd_journal_wait() call to
  5466. make writing synchronous journal clients easier.
  5467. * journalctl gained the new -D switch to show journals from a
  5468. specific directory.
  5469. * journalctl now displays a special marker between log
  5470. messages of two different boots.
  5471. * The journal is now explicitly flushed to /var via a service
  5472. systemd-journal-flush.service, rather than implicitly simply
  5473. by seeing /var/log/journal to be writable.
  5474. * journalctl (and the journal C APIs) can now match for much
  5475. more complex expressions, with alternatives and
  5476. disjunctions.
  5477. * When transitioning from the initial RAM disk to the main
  5478. system we will now kill all processes in a killing spree to
  5479. ensure no processes stay around by accident.
  5480. * Three new specifiers may be used in unit files: %u, %h, %s
  5481. resolve to the user name, user home directory resp. user
  5482. shell. This is useful for running systemd user instances.
  5483. * We now automatically rotate journal files if their data
  5484. object hash table gets a fill level > 75%. We also size the
  5485. hash table based on the configured maximum file size. This
  5486. together should lower hash collisions drastically and thus
  5487. speed things up a bit.
  5488. * journalctl gained the new "--header" switch to introspect
  5489. header data of journal files.
  5490. * A new setting SystemCallFilters= has been added to services
  5491. which may be used to apply blacklists or whitelists to
  5492. system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
  5493. * nspawn gained a new --link-journal= switch (and quicker: -j)
  5494. to link the container journal with the host. This makes it
  5495. very easy to centralize log viewing on the host for all
  5496. guests while still keeping the journal files separated.
  5497. * Many bugfixes and optimizations
  5498. Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
  5499. Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
  5500. Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
  5501. Jędrzejewski-Szmek
  5502. CHANGES WITH 186:
  5503. * Several tools now understand kernel command line arguments,
  5504. which are only read when run in an initial RAM disk. They
  5505. usually follow closely their normal counterparts, but are
  5506. prefixed with rd.
  5507. * There's a new tool to analyze the readahead files that are
  5508. automatically generated at boot. Use:
  5509. /usr/lib/systemd/systemd-readahead analyze /.readahead
  5510. * We now provide an early debug shell on tty9 if this enabled. Use:
  5511. systemctl enable debug-shell.service
  5512. * All plymouth related units have been moved into the Plymouth
  5513. package. Please make sure to upgrade your Plymouth version
  5514. as well.
  5515. * systemd-tmpfiles now supports getting passed the basename of
  5516. a configuration file only, in which case it will look for it
  5517. in all appropriate directories automatically.
  5518. * udevadm info now takes a /dev or /sys path as argument, and
  5519. does the right thing. Example:
  5520. udevadm info /dev/sda
  5521. udevadm info /sys/class/block/sda
  5522. * systemctl now prints a warning if a unit is stopped but a
  5523. unit that might trigger it continues to run. Example: a
  5524. service is stopped but the socket that activates it is left
  5525. running.
  5526. * "systemctl status" will now mention if the log output was
  5527. shortened due to rotation since a service has been started.
  5528. * The journal API now exposes functions to determine the
  5529. "cutoff" times due to rotation.
  5530. * journald now understands SIGUSR1 and SIGUSR2 for triggering
  5531. immediately flushing of runtime logs to /var if possible,
  5532. resp. for triggering immediate rotation of the journal
  5533. files.
  5534. * It is now considered an error if a service is attempted to
  5535. be stopped that is not loaded.
  5536. * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
  5537. * systemd-analyze now supports Python 3
  5538. * tmpfiles now supports cleaning up directories via aging
  5539. where the first level dirs are always kept around but
  5540. directories beneath it automatically aged. This is enabled
  5541. by prefixing the age field with '~'.
  5542. * Seat objects now expose CanGraphical, CanTTY properties
  5543. which is required to deal with very fast bootups where the
  5544. display manager might be running before the graphics drivers
  5545. completed initialization.
  5546. * Seat objects now expose a State property.
  5547. * We now include RPM macros for service enabling/disabling
  5548. based on the preset logic. We recommend RPM based
  5549. distributions to make use of these macros if possible. This
  5550. makes it simpler to reuse RPM spec files across
  5551. distributions.
  5552. * We now make sure that the collected systemd unit name is
  5553. always valid when services log to the journal via
  5555. * There's a new man page kernel-command-line(7) detailing all
  5556. command line options we understand.
  5557. * The fstab generator may now be disabled at boot by passing
  5558. fstab=0 on the kernel command line.
  5559. * A new kernel command line option modules-load= is now understood
  5560. to load a specific kernel module statically, early at boot.
  5561. * Unit names specified on the systemctl command line are now
  5562. automatically escaped as needed. Also, if file system or
  5563. device paths are specified they are automatically turned
  5564. into the appropriate mount or device unit names. Example:
  5565. systemctl status /home
  5566. systemctl status /dev/sda
  5567. * The SysVConsole= configuration option has been removed from
  5568. system.conf parsing.
  5569. * The SysV search path is no longer exported on the D-Bus
  5570. Manager object.
  5571. * The Names= option has been removed from unit file parsing.
  5572. * There's a new man page bootup(7) detailing the boot process.
  5573. * Every unit and every generator we ship with systemd now
  5574. comes with full documentation. The self-explanatory boot is
  5575. complete.
  5576. * A couple of services gained "systemd-" prefixes in their
  5577. name if they wrap systemd code, rather than only external
  5578. code. Among them fsck@.service which is now
  5579. systemd-fsck@.service.
  5580. * The HaveWatchdog property has been removed from the D-Bus
  5581. Manager object.
  5582. * systemd.confirm_spawn= on the kernel command line should now
  5583. work sensibly.
  5584. * There's a new man page crypttab(5) which details all options
  5585. we actually understand.
  5586. * systemd-nspawn gained a new --capability= switch to pass
  5587. additional capabilities to the container.
  5588. * timedated will now read known NTP implementation unit names
  5589. from /usr/lib/systemd/ntp-units.d/*.list,
  5590. has been removed.
  5591. * journalctl gained a new switch "-b" that lists log data of
  5592. the current boot only.
  5593. * The notify socket is in the abstract namespace again, in
  5594. order to support daemons which chroot() at start-up.
  5595. * There is a new Storage= configuration option for journald
  5596. which allows configuration of where log data should go. This
  5597. also provides a way to disable journal logging entirely, so
  5598. that data collected is only forwarded to the console, the
  5599. kernel log buffer or another syslog implementation.
  5600. * Many bugfixes and optimizations
  5601. Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
  5602. David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
  5603. Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
  5604. Shawn Landden, Tom Gundersen
  5605. CHANGES WITH 185:
  5606. * "systemctl help <unit>" now shows the man page if one is
  5607. available.
  5608. * Several new man pages have been added.
  5609. * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
  5610. MaxLevelConsole= can now be specified in
  5611. journald.conf. These options allow reducing the amount of
  5612. data stored on disk or forwarded by the log level.
  5613. * TimerSlackNSec= can now be specified in system.conf for
  5614. PID1. This allows system-wide power savings.
  5615. Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
  5616. Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
  5617. Matthias Clasen
  5618. CHANGES WITH 184:
  5619. * logind is now capable of (optionally) handling power and
  5620. sleep keys as well as the lid switch.
  5621. * journalctl now understands the syntax "journalctl
  5622. /usr/bin/avahi-daemon" to get all log output of a specific
  5623. daemon.
  5624. * CapabilityBoundingSet= in system.conf now also influences
  5625. the capability bound set of usermode helpers of the kernel.
  5626. Contributions from: Daniel Drake, Daniel J. Walsh, Gert
  5627. Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
  5628. Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
  5629. Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
  5630. CHANGES WITH 183:
  5631. * Note that we skipped 139 releases here in order to set the
  5632. new version to something that is greater than both udev's
  5633. and systemd's most recent version number.
  5634. * udev: all udev sources are merged into the systemd source tree now.
  5635. All future udev development will happen in the systemd tree. It
  5636. is still fully supported to use the udev daemon and tools without
  5637. systemd running, like in initramfs or other init systems. Building
  5638. udev though, will require the *build* of the systemd tree, but
  5639. udev can be properly *run* without systemd.
  5640. * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
  5641. should be used to create dead device nodes as workarounds for broken
  5642. subsystems.
  5643. * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
  5644. no longer supported. udev_monitor_new_from_netlink() needs to be
  5645. used to subscribe to events.
  5646. * udev: when udevd is started by systemd, processes which are left
  5647. behind by forking them off of udev rules, are unconditionally cleaned
  5648. up and killed now after the event handling has finished. Services or
  5649. daemons must be started as systemd services. Services can be
  5650. pulled-in by udev to get started, but they can no longer be directly
  5651. forked by udev rules.
  5652. * udev: the daemon binary is called systemd-udevd now and installed
  5653. in /usr/lib/systemd/. Standalone builds or non-systemd systems need
  5654. to adapt to that, create symlink, or rename the binary after building
  5655. it.
  5656. * libudev no longer provides these symbols:
  5657. udev_monitor_from_socket()
  5658. udev_queue_get_failed_list_entry()
  5659. udev_get_{dev,sys,run}_path()
  5660. The versions number was bumped and symbol versioning introduced.
  5661. * systemd-loginctl and systemd-journalctl have been renamed
  5662. to loginctl and journalctl to match systemctl.
  5663. * The config files: /etc/systemd/systemd-logind.conf and
  5664. /etc/systemd/systemd-journald.conf have been renamed to
  5665. logind.conf and journald.conf. Package updates should rename
  5666. the files to the new names on upgrade.
  5667. * For almost all files the license is now LGPL2.1+, changed
  5668. from the previous GPL2.0+. Exceptions are some minor stuff
  5669. of udev (which will be changed to LGPL2.1 eventually, too),
  5670. and the MIT licensed sd-daemon.[ch] library that is suitable
  5671. to be used as drop-in files.
  5672. * systemd and logind now handle system sleep states, in
  5673. particular suspending and hibernating.
  5674. * logind now implements a sleep/shutdown/idle inhibiting logic
  5675. suitable for a variety of uses. Soonishly Lennart will blog
  5676. about this in more detail.
  5677. * var-run.mount and var-lock.mount are no longer provided
  5678. (which previously bind mounted these directories to their new
  5679. places). Distributions which have not converted these
  5680. directories to symlinks should consider stealing these files
  5681. from git history and add them downstream.
  5682. * We introduced the Documentation= field for units and added
  5683. this to all our shipped units. This is useful to make it
  5684. easier to explore the boot and the purpose of the various
  5685. units.
  5686. * All smaller setup units (such as
  5687. systemd-vconsole-setup.service) now detect properly if they
  5688. are run in a container and are skipped when
  5689. appropriate. This guarantees an entirely noise-free boot in
  5690. Linux container environments such as systemd-nspawn.
  5691. * A framework for implementing offline system updates is now
  5692. integrated, for details see:
  5694. * A new service type Type=idle is available now which helps us
  5695. avoiding ugly interleaving of getty output and boot status
  5696. messages.
  5697. * There's now a system-wide CapabilityBoundingSet= option to
  5698. globally reduce the set of capabilities for the
  5699. system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
  5701. even CAP_NET_ADMIN system-wide for secure systems.
  5702. * There are now system-wide DefaultLimitXXX= options to
  5703. globally change the defaults of the various resource limits
  5704. for all units started by PID 1.
  5705. * Harald Hoyer's systemd test suite has been integrated into
  5706. systemd which allows easy testing of systemd builds in qemu
  5707. and nspawn. (This is really awesome! Ask us for details!)
  5708. * The fstab parser is now implemented as generator, not inside
  5709. of PID 1 anymore.
  5710. * systemctl will now warn you if .mount units generated from
  5711. /etc/fstab are out of date due to changes in fstab that
  5712. have not been read by systemd yet.
  5713. * systemd is now suitable for usage in initrds. Dracut has
  5714. already been updated to make use of this. With this in place
  5715. initrds get a slight bit faster but primarily are much
  5716. easier to introspect and debug since "systemctl status" in
  5717. the host system can be used to introspect initrd services,
  5718. and the journal from the initrd is kept around too.
  5719. * systemd-delta has been added, a tool to explore differences
  5720. between user/admin configuration and vendor defaults.
  5721. * PrivateTmp= now affects both /tmp and /var/tmp.
  5722. * Boot time status messages are now much prettier and feature
  5723. proper english language. Booting up systemd has never been
  5724. so sexy.
  5725. * Read-ahead pack files now include the inode number of all
  5726. files to pre-cache. When the inode changes the pre-caching
  5727. is not attempted. This should be nicer to deal with updated
  5728. packages which might result in changes of read-ahead
  5729. patterns.
  5730. * We now temporaritly lower the kernel's read_ahead_kb variable
  5731. when collecting read-ahead data to ensure the kernel's
  5732. built-in read-ahead does not add noise to our measurements
  5733. of necessary blocks to pre-cache.
  5734. * There's now RequiresMountsFor= to add automatic dependencies
  5735. for all mounts necessary for a specific file system path.
  5736. * MountAuto= and SwapAuto= have been removed from
  5737. system.conf. Mounting file systems at boot has to take place
  5738. in systemd now.
  5739. * nspawn now learned a new switch --uuid= to set the machine
  5740. ID on the command line.
  5741. * nspawn now learned the -b switch to automatically search
  5742. for an init system.
  5743. * vt102 is now the default TERM for serial TTYs, upgraded from
  5744. vt100.
  5745. * systemd-logind now works on VT-less systems.
  5746. * The build tree has been reorganized. The individual
  5747. components now have directories of their own.
  5748. * A new condition type ConditionPathIsReadWrite= is now available.
  5749. * nspawn learned the new -C switch to create cgroups for the
  5750. container in other hierarchies.
  5751. * We now have support for hardware watchdogs, configurable in
  5752. system.conf.
  5753. * The scheduled shutdown logic now has a public API.
  5754. * We now mount /tmp as tmpfs by default, but this can be
  5755. masked and /etc/fstab can override it.
  5756. * Since udisks does not make use of /media anymore we are not
  5757. mounting a tmpfs on it anymore.
  5758. * journalctl gained a new --local switch to only interleave
  5759. locally generated journal files.
  5760. * We can now load the IMA policy at boot automatically.
  5761. * The GTK tools have been split off into a systemd-ui.
  5762. Contributions from: Andreas Schwab, Auke Kok, Ayan George,
  5763. Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
  5764. Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
  5765. Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
  5766. Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
  5767. Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
  5768. A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
  5769. Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
  5770. Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
  5771. Gundersen
  5772. CHANGES WITH 44:
  5773. * This is mostly a bugfix release
  5774. * Support optional initialization of the machine ID from the
  5775. KVM or container configured UUID.
  5776. * Support immediate reboots with "systemctl reboot -ff"
  5777. * Show /etc/os-release data in systemd-analyze output
  5778. * Many bugfixes for the journal, including endianness fixes and
  5779. ensuring that disk space enforcement works
  5780. * sd-login.h is C++ compatible again
  5781. * Extend the /etc/os-release format on request of the Debian
  5782. folks
  5783. * We now refuse non-UTF8 strings used in various configuration
  5784. and unit files. This is done to ensure we do not pass invalid
  5785. data over D-Bus or expose it elsewhere.
  5786. * Register Mimo USB Screens as suitable for automatic seat
  5787. configuration
  5788. * Read SELinux client context from journal clients in a race
  5789. free fashion
  5790. * Reorder configuration file lookup order. /etc now always
  5791. overrides /run in order to allow the administrator to always
  5792. and unconditionally override vendor-supplied or
  5793. automatically generated data.
  5794. * The various user visible bits of the journal now have man
  5795. pages. We still lack man pages for the journal API calls
  5796. however.
  5797. * We now ship all man pages in HTML format again in the
  5798. tarball.
  5799. Contributions from: Dave Reisner, Dirk Eibach, Frederic
  5800. Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
  5801. Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
  5802. Reding
  5803. CHANGES WITH 43:
  5804. * This is mostly a bugfix release
  5805. * systems lacking /etc/os-release are no longer supported.
  5806. * Various functionality updates to
  5807. * Track class of PAM logins to distinguish greeters from
  5808. normal user logins.
  5809. Contributions from: Kay Sievers, Lennart Poettering, Michael
  5810. Biebl
  5811. CHANGES WITH 42:
  5812. * This is an important bugfix release for v41.
  5813. * Building man pages is now optional which should be useful
  5814. for those building systemd from git but unwilling to install
  5815. xsltproc.
  5816. * Watchdog support for supervising services is now usable. In
  5817. a future release support for hardware watchdogs
  5818. (i.e. /dev/watchdog) will be added building on this.
  5819. * Service start rate limiting is now configurable and can be
  5820. turned off per service. When a start rate limit is hit a
  5821. reboot can automatically be triggered.
  5822. * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
  5823. Contributions from: Benjamin Franzke, Bill Nottingham,
  5824. Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
  5825. Schmidt, Michał Górny, Piotr Drąg
  5826. CHANGES WITH 41:
  5827. * The systemd binary is installed /usr/lib/systemd/systemd now;
  5828. An existing /sbin/init symlink needs to be adapted with the
  5829. package update.
  5830. * The code that loads kernel modules has been ported to invoke
  5831. libkmod directly, instead of modprobe. This means we do not
  5832. support systems with module-init-tools anymore.
  5833. * Watchdog support is now already useful, but still not
  5834. complete.
  5835. * A new kernel command line option systemd.setenv= is
  5836. understood to set system wide environment variables
  5837. dynamically at boot.
  5838. * We now limit the set of capabilities of systemd-journald.
  5839. * We now set SIGPIPE to ignore by default, since it only is
  5840. useful in shell pipelines, and has little use in general
  5841. code. This can be disabled with IgnoreSIPIPE=no in unit
  5842. files.
  5843. Contributions from: Benjamin Franzke, Kay Sievers, Lennart
  5844. Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
  5845. William Douglas
  5846. CHANGES WITH 40:
  5847. * This is mostly a bugfix release
  5848. * We now expose the reason why a service failed in the
  5849. "Result" D-Bus property.
  5850. * Rudimentary service watchdog support (will be completed over
  5851. the next few releases.)
  5852. * When systemd forks off in order execute some service we will
  5853. now immediately changes its argv[0] to reflect which process
  5854. it will execute. This is useful to minimize the time window
  5855. with a generic argv[0], which makes bootcharts more useful
  5856. Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
  5857. Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
  5858. Mike Kazantsev, Ray Strode
  5859. CHANGES WITH 39:
  5860. * This is mostly a test release, but incorporates many
  5861. bugfixes.
  5862. * New systemd-cgtop tool to show control groups by their
  5863. resource usage.
  5864. * Linking against libacl for ACLs is optional again. If
  5865. disabled, support tracking device access for active logins
  5866. goes becomes unavailable, and so does access to the user
  5867. journals by the respective users.
  5868. * If a group "adm" exists, journal files are automatically
  5869. owned by them, thus allow members of this group full access
  5870. to the system journal as well as all user journals.
  5871. * The journal now stores the SELinux context of the logging
  5872. client for all entries.
  5873. * Add C++ inclusion guards to all public headers
  5874. * New output mode "cat" in the journal to print only text
  5875. messages, without any meta data like date or time.
  5876. * Include tiny X server wrapper as a temporary stop-gap to
  5877. teach XOrg udev display enumeration. This is used by display
  5878. managers such as gdm, and will go away as soon as XOrg
  5879. learned native udev hotplugging for display devices.
  5880. * Add new systemd-cat tool for executing arbitrary programs
  5881. with STDERR/STDOUT connected to the journal. Can also act as
  5882. BSD logger replacement, and does so by default.
  5883. * Optionally store all locally generated coredumps in the
  5884. journal along with meta data.
  5885. * systemd-tmpfiles learnt four new commands: n, L, c, b, for
  5886. writing short strings to files (for usage for /sys), and for
  5887. creating symlinks, character and block device nodes.
  5888. * New unit file option ControlGroupPersistent= to make cgroups
  5889. persistent, following the mechanisms outlined in
  5891. * Support multiple local RTCs in a sane way
  5892. * No longer monopolize IO when replaying readahead data on
  5893. rotating disks, since we might starve non-file-system IO to
  5894. death, since fanotify() will not see accesses done by blkid,
  5895. or fsck.
  5896. * Do not show kernel threads in systemd-cgls anymore, unless
  5897. requested with new -k switch.
  5898. Contributions from: Dan Horák, Kay Sievers, Lennart
  5899. Poettering, Michal Schmidt
  5900. CHANGES WITH 38:
  5901. * This is mostly a test release, but incorporates many
  5902. bugfixes.
  5903. * The git repository moved to:
  5904. git://
  5905. ssh://
  5906. * First release with the journal
  5908. * The journal replaces both systemd-kmsg-syslogd and
  5909. systemd-stdout-bridge.
  5910. * New sd_pid_get_unit() API call in libsystemd-logind
  5911. * Many systemadm clean-ups
  5912. * Introduce which is ordered before all
  5913. remote mounts and may be used to start services before all
  5914. remote mounts.
  5915. * Added Mageia support
  5916. * Add bash completion for systemd-loginctl
  5917. * Actively monitor PID file creation for daemons which exit in
  5918. the parent process before having finished writing the PID
  5919. file in the daemon process. Daemons which do this need to be
  5920. fixed (i.e. PID file creation must have finished before the
  5921. parent exits), but we now react a bit more gracefully to them.
  5922. * Add colourful boot output, mimicking the well-known output
  5923. of existing distributions.
  5924. * New option PassCredentials= for socket units, for
  5925. compatibility with a recent kernel ABI breakage.
  5926. * /etc/rc.local is now hooked in via a generator binary, and
  5927. thus will no longer act as synchronization point during
  5928. boot.
  5929. * systemctl list-unit-files now supports --root=.
  5930. * systemd-tmpfiles now understands two new commands: z, Z for
  5931. relabelling files according to the SELinux database. This is
  5932. useful to apply SELinux labels to specific files in /sys,
  5933. among other things.
  5934. * Output of SysV services is now forwarded to both the console
  5935. and the journal by default, not only just the console.
  5936. * New man pages for all APIs from libsystemd-login.
  5937. * The build tree got reorganized and the build system is a
  5938. lot more modular allowing embedded setups to specifically
  5939. select the components of systemd they are interested in.
  5940. * Support for Linux systems lacking the kernel VT subsystem is
  5941. restored.
  5942. * configure's --with-rootdir= got renamed to
  5943. --with-rootprefix= to follow the naming used by udev and
  5944. kmod
  5945. * Unless specified otherwise we will now install to /usr instead
  5946. of /usr/local by default.
  5947. * Processes with '@' in argv[0][0] are now excluded from the
  5948. final shut-down killing spree, following the logic explained
  5949. in:
  5951. * All processes remaining in a service cgroup when we enter
  5952. the START or START_PRE states are now killed with
  5953. SIGKILL. That means it is no longer possible to spawn
  5954. background processes from ExecStart= lines (which was never
  5955. supported anyway, and bad style).
  5956. * New PropagateReloadTo=/PropagateReloadFrom= options to bind
  5957. reloading of units together.
  5958. Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
  5959. Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
  5960. Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
  5961. Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
  5962. Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek