"Fossies" - the Fresh Open Source Software Archive

Member "systemd-248/NEWS" (30 Mar 2021, 666758 Bytes) of package /linux/misc/systemd-248.tar.gz:

As a special service "Fossies" has tried to format the requested text file into HTML format (style: standard) with prefixed line numbers. Alternatively you can here view or download the uninterpreted source code file. See also the latest Fossies "Diffs" side-by-side code changes report for "NEWS": 247_vs_248.

    1 systemd System and Service Manager
    3 CHANGES WITH 248:
    5         * A concept of system extension images is introduced. Such images may
    6           be used to extend the /usr/ and /opt/ directory hierarchies at
    7           runtime with additional files (even if the file system is read-only).
    8           When a system extension image is activated, its /usr/ and /opt/
    9           hierarchies and os-release information are combined via overlayfs
   10           with the file system hierarchy of the host OS.
   12           A new systemd-sysext tool can be used to merge, unmerge, list, and
   13           refresh system extension hierarchies. See
   14           https://www.freedesktop.org/software/systemd/man/systemd-sysext.html.
   16           The systemd-sysext.service automatically merges installed system
   17           extensions during boot (before basic.target, but not in very early
   18           boot, since various file systems have to be mounted first).
   20           The SYSEXT_LEVEL= field in os-release(5) may be used to specify the
   21           supported system extension level.
   23         * A new ExtensionImages= unit setting can be used to apply the same
   24           system extension image concept from systemd-sysext to the namespaced
   25           file hierarchy of specific services, following the same rules and
   26           constraints.
   28         * Support for a new special "root=tmpfs" kernel command-line option has
   29           been added. When specified, a tmpfs is mounted on /, and mount.usr=
   30           should be used to point to the operating system implementation.
   32         * A new configuration file /etc/veritytab may be used to configure
   33           dm-verity integrity protection for block devices. Each line is in the
   34           format "volume-name data-device hash-device roothash options",
   35           similar to /etc/crypttab.
   37         * A new kernel command-line option systemd.verity.root_options= may be
   38           used to configure dm-verity behaviour for the root device.
   40         * The key file specified in /etc/crypttab (the third field) may now
   41           refer to an AF_UNIX/SOCK_STREAM socket in the file system. The key is
   42           acquired by connecting to that socket and reading from it. This
   43           allows the implementation of a service to provide key information
   44           dynamically, at the moment when it is needed.
   46         * When the hostname is set explicitly to "localhost", systemd-hostnamed
   47           will respect this. Previously such a setting would be mostly silently
   48           ignored. The goal is to honour configuration as specified by the
   49           user.
   51         * The fallback hostname that will be used by the system manager and
   52           systemd-hostnamed can now be configured in two new ways: by setting
   53           DEFAULT_HOSTNAME= in os-release(5), or by setting
   54           $SYSTEMD_DEFAULT_HOSTNAME in the environment block. As before, it can
   55           also be configured during compilation. The environment variable is
   56           intended for testing and local overrides, the os-release(5) field is
   57           intended to allow customization by different variants of a
   58           distribution that share the same compiled packages.
   60         * The environment block of the manager itself may be configured through
   61           a new ManagerEnvironment= setting in system.conf or user.conf. This
   62           complements existing ways to set the environment block (the kernel
   63           command line for the system manager, the inherited environment and
   64           user@.service unit file settings for the user manager).
   66         * systemd-hostnamed now exports the default hostname and the source of
   67           the configured hostname ("static", "transient", or "default") as
   68           D-Bus properties.
   70         * systemd-hostnamed now exports the "HardwareVendor" and
   71           "HardwareModel" D-Bus properties, which are supposed to contain a
   72           pair of cleaned up, human readable strings describing the system's
   73           vendor and model. It's typically sourced from the firmware's DMI
   74           tables, but may be augmented from a new hwdb database. hostnamectl
   75           shows this in the status output.
   77         * Support has been added to systemd-cryptsetup for extracting the
   78           PKCS#11 token URI and encrypted key from the LUKS2 JSON embedded
   79           metadata header. This allows the information how to open the
   80           encrypted device to be embedded directly in the device and obviates
   81           the need for configuration in an external file.
   83         * systemd-cryptsetup gained support for unlocking LUKS2 volumes using
   84           TPM2 hardware, as well as FIDO2 security tokens (in addition to the
   85           pre-existing support for PKCS#11 security tokens).
   87         * systemd-repart may enroll encrypted partitions using TPM2
   88           hardware. This may be useful for example to create an encrypted /var
   89           partition bound to the machine on first boot.
   91         * A new systemd-cryptenroll tool has been added to enroll TPM2, FIDO2
   92           and PKCS#11 security tokens to LUKS volumes, list and destroy
   93           them. See:
   95           http://0pointer.net/blog/unlocking-luks2-volumes-with-tpm2-fido2-pkcs11-security-hardware-on-systemd-248.html
   97           It also supports enrolling "recovery keys" and regular passphrases.
   99         * The libfido2 dependency is now based on dlopen(), so that the library
  100           is used at runtime when installed, but is not a hard runtime
  101           dependency.
  103         * systemd-cryptsetup gained support for two new options in
  104           /etc/crypttab: "no-write-workqueue" and "no-read-workqueue" which
  105           request synchronous processing of encryption/decryption IO.
  107         * The manager may be configured at compile time to use the fexecve()
  108           instead of the execve() system call when spawning processes. Using
  109           fexecve() closes a window between checking the security context of an
  110           executable and spawning it, but unfortunately the kernel displays
  111           stale information in the process' "comm" field, which impacts ps
  112           output and such.
  114         * The configuration option -Dcompat-gateway-hostname has been dropped.
  115           "_gateway" is now the only supported name.
  117         * The ConditionSecurity=tpm2 unit file setting may be used to check if
  118           the system has at least one TPM2 (tpmrm class) device.
  120         * A new ConditionCPUFeature= has been added that may be used to
  121           conditionalize units based on CPU features. For example,
  122           ConditionCPUFeature=rdrand will condition a unit so that it is only
  123           run when the system CPU supports the RDRAND opcode.
  125         * The existing ConditionControlGroupController= setting has been
  126           extended with two new values "v1" and "v2". "v2" means that the
  127           unified v2 cgroup hierarchy is used, and "v1" means that legacy v1
  128           hierarchy or the hybrid hierarchy are used.
  130         * A new PrivateIPC= setting on a unit file allows executed processes to
  131           be moved into a private IPC namespace, with separate System V IPC
  132           identifiers and POSIX message queues.
  134           A new IPCNamespacePath= allows the unit to be joined to an existing
  135           IPC namespace.
  137         * The tables of system calls in seccomp filters are now automatically
  138           generated from kernel lists exported on
  139           https://fedora.juszkiewicz.com.pl/syscalls.html.
  141           The following architectures should now have complete lists:
  142           alpha, arc, arm64, arm, i386, ia64, m68k, mips64n32, mips64, mipso32,
  143           powerpc, powerpc64, s390, s390x, tilegx, sparc, x86_64, x32.
  145         * The MountAPIVFS= service file setting now additionally mounts a tmpfs
  146           on /run/ if it is not already a mount point. A writable /run/ has
  147           always been a requirement for a functioning system, but this was not
  148           guaranteed when using a read-only image.
  150           Users can always specify BindPaths= or InaccessiblePaths= as
  151           overrides, and they will take precedence. If the host's root mount
  152           point is used, there is no change in behaviour.
  154         * New bind mounts and file system image mounts may be injected into the
  155           mount namespace of a service (without restarting it). This is exposed
  156           respectively as 'systemctl bind <unit> <path>…' and
  157           'systemctl mount-image <unit> <image>…'.
  159         * The StandardOutput= and StandardError= settings can now specify files
  160           to be truncated for output (as "truncate:<path>").
  162         * The ExecPaths= and NoExecPaths= settings may be used to specify
  163           noexec for parts of the file system.
  165         * sd-bus has a new function sd_bus_open_user_machine() to open a
  166           connection to the session bus of a specific user in a local container
  167           or on the local host. This is exposed in the existing -M switch to
  168           systemctl and similar tools:
  170               systemctl --user -M lennart@foobar start foo
  172           This will connect to the user bus of a user "lennart" in container
  173           "foobar". If no container name is specified, the specified user on
  174           the host itself is connected to
  176               systemctl --user -M lennart@ start quux
  178         * sd-bus also gained a convenience function sd_bus_message_send() to
  179           simplify invocations of sd_bus_send(), taking only a single
  180           parameter: the message to send.
  182         * sd-event allows rate limits to be set on event sources, for dealing
  183           with high-priority event sources that might starve out others. See
  184           the new man page sd_event_source_set_ratelimit(3) for details.
  186         * systemd.link files gained a [Link] Promiscuous= switch, which allows
  187           the device to be raised in promiscuous mode.
  189           New [Link] TransmitQueues= and ReceiveQueues= settings allow the
  190           number of TX and RX queues to be configured.
  192           New [Link] TransmitQueueLength= setting allows the size of the TX
  193           queue to be configured.
  195           New [Link] GenericSegmentOffloadMaxBytes= and
  196           GenericSegmentOffloadMaxSegments= allow capping the packet size and
  197           the number of segments accepted in Generic Segment Offload.
  199         * systemd-networkd gained support for the "B.A.T.M.A.N. advanced"
  200           wireless routing protocol that operates on ISO/OSI Layer 2 only and
  201           uses ethernet frames to route/bridge packets. This encompasses a new
  202           "batadv" netdev Type=, a new [BatmanAdvanced] section with a bunch of
  203           new settings in .netdev files, and a new BatmanAdvanced= setting in
  204           .network files.
  206         * systemd.network files gained a [Network] RouteTable= configuration
  207           switch to select the routing policy table.
  209           systemd.network files gained a [RoutingPolicyRule] Type=
  210           configuration switch (one of "blackhole, "unreachable", "prohibit").
  212           systemd.network files gained a [IPv6AcceptRA] RouteDenyList= and
  213           RouteAllowList= settings to ignore/accept route advertisements from
  214           routers matching specified prefixes. The DenyList= setting has been
  215           renamed to PrefixDenyList= and a new PrefixAllowList= option has been
  216           added.
  218           systemd.network files gained a [DHCPv6] UseAddress= setting to
  219           optionally ignore the address provided in the lease.
  221           systemd.network files gained a [DHCPv6PrefixDelegation]
  222           ManageTemporaryAddress= switch.
  224           systemd.network files gained a new ActivationPolicy= setting which
  225           allows configuring how the UP state of an interface shall be managed,
  226           i.e. whether the interface is always upped, always downed, or may be
  227           upped/downed by the user using "ip link set dev".
  229         * The default for the Broadcast= setting in .network files has slightly
  230           changed: the broadcast address will not be configured for wireguard
  231           devices.
  233         * systemd.netdev files gained a [VLAN] Protocol=, IngressQOSMaps=,
  234           EgressQOSMaps=, and [MACVLAN] BroadcastMulticastQueueLength=
  235           configuration options for VLAN packet handling.
  237         * udev rules may now set log_level= option. This allows debug logs to
  238           be enabled for select events, e.g. just for a specific subsystem or
  239           even a single device.
  241         * udev now exports the VOLUME_ID, LOGICAL_VOLUME_ID, VOLUME_SET_ID, and
  242           DATA_PREPARED_ID properties for block devices with ISO9660 file
  243           systems.
  245         * udev now exports decoded DMI information about installed memory slots
  246           as device properties under the /sys/class/dmi/id/ pseudo device.
  248         * /dev/ is not mounted noexec anymore. This didn't provide any
  249           significant security benefits and would conflict with the executable
  250           mappings used with /dev/sgx device nodes. The previous behaviour can
  251           be restored for individual services with NoExecPaths=/dev (or by allow-
  252           listing and excluding /dev from ExecPaths=).
  254         * Permissions for /dev/vsock are now set to 0o666, and /dev/vhost-vsock
  255           and /dev/vhost-net are owned by the kvm group.
  257         * The hardware database has been extended with a list of fingerprint
  258           readers that correctly support USB auto-suspend using data from
  259           libfprint.
  261         * systemd-resolved can now answer DNSSEC questions through the stub
  262           resolver interface in a way that allows local clients to do DNSSEC
  263           validation themselves. For a question with DO+CD set, it'll proxy the
  264           DNS query and respond with a mostly unmodified packet received from
  265           the upstream server.
  267         * systemd-resolved learnt a new boolean option CacheFromLocalhost= in
  268           resolved.conf. If true the service will provide caching even for DNS
  269           lookups made to an upstream DNS server on the
  270           addresses. By default (and when the option is false) systemd-resolved
  271           will not cache such lookups, in order to avoid duplicate local
  272           caching, under the assumption the local upstream server caches
  273           anyway.
  275         * systemd-resolved now implements RFC5001 NSID in its local DNS
  276           stub. This may be used by local clients to determine whether they are
  277           talking to the DNS resolver stub or a different DNS server.
  279         * When resolving host names and other records resolvectl will now
  280           report where the data was acquired from (i.e. the local cache, the
  281           network, locally synthesized, …) and whether the network traffic it
  282           effected was encrypted or not. Moreover the tool acquired a number of
  283           new options --cache=, --synthesize=, --network=, --zone=,
  284           --trust-anchor=, --validate= that take booleans and may be used to
  285           tweak a lookup, i.e. whether it may be answered from cached
  286           information, locally synthesized information, information acquired
  287           through the network, the local mDNS/LLMNR zone, the DNSSEC trust
  288           anchor, and whether DNSSEC validation shall be executed for the
  289           lookup.
  291         * systemd-nspawn gained a new --ambient-capability= setting
  292           (AmbientCapability= in .nspawn files) to configure ambient
  293           capabilities passed to the container payload.
  295         * systemd-nspawn gained the ability to configure the firewall using the
  296           nftables subsystem (in addition to the existing iptables
  297           support). Similarly, systemd-networkd's IPMasquerade= option now
  298           supports nftables as back-end, too. In both cases NAT on IPv6 is now
  299           supported too, in addition to IPv4 (the iptables back-end still is
  300           IPv4-only).
  302           "IPMasquerade=yes", which was the same as "IPMasquerade=ipv4" before,
  303           retains its meaning, but has been deprecated. Please switch to either
  304           "ivp4" or "both" (if covering IPv6 is desired).
  306         * systemd-importd will now download .verity and .roothash.p7s files
  307           along with the machine image (as exposed via machinectl pull-raw).
  309         * systemd-oomd now gained a new DefaultMemoryPressureDurationSec=
  310           setting to configure the time a unit's cgroup needs to exceed memory
  311           pressure limits before action will be taken, and a new
  312           ManagedOOMPreference=none|avoid|omit setting to avoid killing certain
  313           units.
  315           systemd-oomd is now considered fully supported (the usual
  316           backwards-compatiblity promises apply). Swap is not required for
  317           operation, but it is still recommended.
  319         * systemd-timesyncd gained a new ConnectionRetrySec= setting which
  320           configures the retry delay when trying to contact servers.
  322         * systemd-stdio-bridge gained --system/--user options to connect to the
  323           system bus (previous default) or the user session bus.
  325         * systemd-localed may now call locale-gen to generate missing locales
  326           on-demand (UTF-8-only). This improves integration with Debian-based
  327           distributions (Debian/Ubuntu/PureOS/Tanglu/...) and Arch Linux.
  329         * systemctl --check-inhibitors=true may now be used to obey inhibitors
  330           even when invoked non-interactively. The old --ignore-inhibitors
  331           switch is now deprecated and replaced by --check-inhibitors=false.
  333         * systemctl import-environment will now emit a warning when called
  334           without any arguments (i.e. to import the full environment block of
  335           the called program). This command will usually be invoked from a
  336           shell, which means that it'll inherit a bunch of variables which are
  337           specific to that shell, and usually to the TTY the shell is connected
  338           to, and don't have any meaning in the global context of the system or
  339           user service manager. Instead, only specific variables should be
  340           imported into the manager environment block.
  342           Similarly, programs which update the manager environment block by
  343           directly calling the D-Bus API of the manager, should also push
  344           specific variables, and not the full inherited environment.
  346         * systemctl's status output now shows unit state with a more careful
  347           choice of Unicode characters: units in maintenance show a "○" symbol
  348           instead of the usual "●", failed units show "×", and services being
  349           reloaded "↻".
  351         * coredumpctl gained a --debugger-arguments= switch to pass arguments
  352           to the debugger. It also gained support for showing coredump info in
  353           a simple JSON format.
  355         * systemctl/loginctl/machinectl's --signal= option now accept a special
  356           value "list", which may be used to show a brief table with known
  357           process signals and their numbers.
  359         * networkctl now shows the link activation policy in status.
  361         * Various tools gained --pager/--no-pager/--json= switches to
  362           enable/disable the pager and provide JSON output.
  364         * Various tools now accept two new values for the SYSTEMD_COLORS
  365           environment variable: "16" and "256", to configure how many terminal
  366           colors are used in output.
  368         * less 568 or newer is now required for the auto-paging logic of the
  369           various tools. Hyperlink ANSI sequences in terminal output are now
  370           used even if a pager is used, and older versions of less are not able
  371           to display these sequences correctly. SYSTEMD_URLIFY=0 may be used to
  372           disable this output again.
  374         * Builds with support for separate / and /usr/ hierarchies ("split-usr"
  375           builds, non-merged-usr builds) are now officially deprecated. A
  376           warning is emitted during build. Support is slated to be removed in
  377           about a year (when the Debian Bookworm release development starts).
  379         * Systems with the legacy cgroup v1 hierarchy are now marked as
  380           "tainted", to make it clearer that using the legacy hierarchy is not
  381           recommended.
  383         * systemd-localed will now refuse to configure a keymap which is not
  384           installed in the file system. This is intended as a bug fix, but
  385           could break cases where systemd-localed was used to configure the
  386           keymap in advanced of it being installed. It is necessary to install
  387           the keymap file first.
  389         * The main git development branch has been renamed to 'main'.
  391         * mmcblk[0-9]boot[0-9] devices will no longer be probed automatically
  392           for partitions, as in the vast majority of cases they contain none
  393           and are used internally by the bootloader (eg: uboot).
  395         * systemd will now set the $SYSTEMD_EXEC_PID environment variable for
  396           spawned processes to the PID of the process itself. This may be used
  397           by programs for detecting whether they were forked off by the service
  398           manager itself or are a process forked off further down the tree.
  400         * The sd-device API gained four new calls: sd_device_get_action() to
  401           determine the uevent add/remove/change/… action the device object has
  402           been seen for, sd_device_get_seqno() to determine the uevent sequence
  403           number, sd_device_new_from_stat_rdev() to allocate a new sd_device
  404           object from stat(2) data of a device node, and sd_device_trigger() to
  405           write to the 'uevent' attribute of a device.
  407         * For most tools the --no-legend= switch has been replaced by
  408           --legend=no and --legend=yes, to force whether tables are shown with
  409           headers/legends.
  411         * Units acquired a new property "Markers" that takes a list of zero,
  412           one or two of the following strings: "needs-reload" and
  413           "needs-restart". These markers may be set via "systemctl
  414           set-property". Once a marker is set, "systemctl reload-or-restart
  415           --marked" may be invoked to execute the operation the units are
  416           marked for. This is useful for package managers that want to mark
  417           units for restart/reload while updating, but effect the actual
  418           operations at a later step at once.
  420         * The sd_bus_message_read_strv() API call of sd-bus may now also be
  421           used to parse arrays of D-Bus signatures and D-Bus paths, in addition
  422           to regular strings.
  424         * bootctl will now report whether the UEFI firmware used a TPM2 device
  425           and measured the boot process into it.
  427         * systemd-tmpfiles learnt support for a new environment variable
  428           $SYSTEMD_TMPFILES_FORCE_SUBVOL which takes a boolean value. If true
  429           the v/q/Q lines in tmpfiles.d/ snippets will create btrfs subvolumes
  430           even if the root fs of the system is not itself a btrfs volume.
  432         * systemd-detect-virt/ConditionVirtualization= will now explicitly
  433           detect Docker/Podman environments where possible. Moreover, they
  434           should be able to generically detect any container manager as long as
  435           it assigns the container a cgroup.
  437         * portablectl gained a new "reattach" verb for detaching/reattaching a
  438           portable service image, useful for updating images on-the-fly.
  440         * Intel SGX enclave device nodes (which expose a security feature of
  441           newer Intel CPUs) will now be owned by a new system group "sgx".
  443         Contributions from: Adam Nielsen, Adrian Vovk, AJ Jordan, Alan Perry,
  444         Alastair Pharo, Alexander Batischev, Ali Abdallah, Andrew Balmos,
  445         Anita Zhang, Annika Wickert, Ansgar Burchardt, Antonio Terceiro,
  446         Antonius Frie, Ardy, Arian van Putten, Ariel Fermani, Arnaud T,
  447         A S Alam, Bastien Nocera, Benjamin Berg, Benjamin Robin, Björn Daase,
  448         caoxia, Carlo Wood, Charles Lee, ChopperRob, chri2, Christian Ehrhardt,
  449         Christian Hesse, Christopher Obbard, clayton craft, corvusnix, cprn,
  450         Daan De Meyer, Daniele Medri, Daniel Rusek, Dan Sanders, Dan Streetman,
  451         Darren Ng, David Edmundson, David Tardon, Deepak Rawat, Devon Pringle,
  452         Dmitry Borodaenko, dropsignal, Einsler Lee, Endre Szabo,
  453         Evgeny Vereshchagin, Fabian Affolter, Fangrui Song, Felipe Borges,
  454         feliperodriguesfr, Felix Stupp, Florian Hülsmann, Florian Klink,
  455         Florian Westphal, Franck Bui, Frantisek Sumsal, Gablegritule,
  456         Gaël PORTAY, Gaurav, Giedrius Statkevičius, Greg Depoire-Ferrer,
  457         Gustavo Costa, Hans de Goede, Hela Basa, heretoenhance, hide,
  458         Iago López Galeiras, igo95862, Ilya Dmitrichenko, Jameer Pathan,
  459         Jan Tojnar, Jiehong, Jinyuan Si, Joerg Behrmann, John Slade,
  460         Jonathan G. Underwood, Jonathan McDowell, Josh Triplett, Joshua Watt,
  461         Julia Cartwright, Julien Humbert, Kairui Song, Karel Zak,
  462         Kevin Backhouse, Kevin P. Fleming, Khem Raj, Konomi, krissgjeng,
  463         l4gfcm, Lajos Veres, Lennart Poettering, Lincoln Ramsay, Luca Boccassi,
  464         Luca BRUNO, Lucas Werkmeister, Luka Kudra, Luna Jernberg,
  465         Marc-André Lureau, Martin Wilck, Matthias Klumpp, Matt Turner,
  466         Michael Gisbers, Michael Marley, Michael Trapp, Michal Fabik,
  467         Michał Kopeć, Michal Koutný, Michal Sekletár, Michele Guerini Rocco,
  468         Mike Gilbert, milovlad, moson-mo, Nick, nihilix-melix, Oğuz Ersen,
  469         Ondrej Mosnacek, pali, Pavel Hrdina, Pavel Sapezhko, Perry Yuan,
  470         Peter Hutterer, Pierre Dubouilh, Piotr Drąg, Pjotr Vertaalt,
  471         Richard Laager, RussianNeuroMancer, Sam Lunt, Sebastiaan van Stijn,
  472         Sergey Bugaev, shenyangyang4, simmon, Simonas Kazlauskas,
  473         Slimane Selyan Amiri, Stefan Agner, Steve Ramage, Susant Sahani,
  474         Sven Mueller, Tad Fisher, Takashi Iwai, Thomas Haller, Tom Shield,
  475         Topi Miettinen, Torsten Hilbrich, tpgxyz, Tyler Hicks, ulf-f,
  476         Ulrich Ölmann, Vincent Pelletier, Vinnie Magro, Vito Caputo, Vlad,
  477         walbit-de, Whired Planck, wouter bolsterlee, Xℹ Ruoyao, Yangyang Shen,
  478         Yuri Chornoivan, Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek,
  479         Zmicer Turok, Дамјан Георгиевски
  481         — Berlin, 2021-03-30
  483 CHANGES WITH 247:
  485         * KERNEL API INCOMPATIBILITY: Linux 4.14 introduced two new uevents
  486           "bind" and "unbind" to the Linux device model. When this kernel
  487           change was made, systemd-udevd was only minimally updated to handle
  488           and propagate these new event types. The introduction of these new
  489           uevents (which are typically generated for USB devices and devices
  490           needing a firmware upload before being functional) resulted in a
  491           number of issues which we so far didn't address. We hoped the kernel
  492           maintainers would themselves address these issues in some form, but
  493           that did not happen. To handle them properly, many (if not most) udev
  494           rules files shipped in various packages need updating, and so do many
  495           programs that monitor or enumerate devices with libudev or sd-device,
  496           or otherwise process uevents. Please note that this incompatibility
  497           is not fault of systemd or udev, but caused by an incompatible kernel
  498           change that happened back in Linux 4.14, but is becoming more and
  499           more visible as the new uevents are generated by more kernel drivers.
  501           To minimize issues resulting from this kernel change (but not avoid
  502           them entirely) starting with systemd-udevd 247 the udev "tags"
  503           concept (which is a concept for marking and filtering devices during
  504           enumeration and monitoring) has been reworked: udev tags are now
  505           "sticky", meaning that once a tag is assigned to a device it will not
  506           be removed from the device again until the device itself is removed
  507           (i.e. unplugged). This makes sure that any application monitoring
  508           devices that match a specific tag is guaranteed to both see uevents
  509           where the device starts being relevant, and those where it stops
  510           being relevant (the latter now regularly happening due to the new
  511           "unbind" uevent type). The udev tags concept is hence now a concept
  512           tied to a *device* instead of a device *event* — unlike for example
  513           udev properties whose lifecycle (as before) is generally tied to a
  514           device event, meaning that the previously determined properties are
  515           forgotten whenever a new uevent is processed.
  517           With the newly redefined udev tags concept, sometimes it's necessary
  518           to determine which tags are the ones applied by the most recent
  519           uevent/database update, in order to discern them from those
  520           originating from earlier uevents/database updates of the same
  521           device. To accommodate for this a new automatic property CURRENT_TAGS
  522           has been added that works similar to the existing TAGS property but
  523           only lists tags set by the most recent uevent/database
  524           update. Similarly, the libudev/sd-device API has been updated with
  525           new functions to enumerate these 'current' tags, in addition to the
  526           existing APIs that now enumerate the 'sticky' ones.
  528           To properly handle "bind"/"unbind" on Linux 4.14 and newer it is
  529           essential that all udev rules files and applications are updated to
  530           handle the new events. Specifically:
  532           • All rule files that currently use a header guard similar to
  533             ACTION!="add|change",GOTO="xyz_end" should be updated to use
  534             ACTION=="remove",GOTO="xyz_end" instead, so that the
  535             properties/tags they add are also applied whenever "bind" (or
  536             "unbind") is seen. (This is most important for all physical device
  537             types — those for which "bind" and "unbind" are currently
  538             generated, for all other device types this change is still
  539             recommended but not as important — but certainly prepares for
  540             future kernel uevent type additions).
  542           • Similarly, all code monitoring devices that contains an 'if' branch
  543             discerning the "add" + "change" uevent actions from all other
  544             uevents actions (i.e. considering devices only relevant after "add"
  545             or "change", and irrelevant on all other events) should be reworked
  546             to instead negatively check for "remove" only (i.e. considering
  547             devices relevant after all event types, except for "remove", which
  548             invalidates the device). Note that this also means that devices
  549             should be considered relevant on "unbind", even though conceptually
  550             this — in some form — invalidates the device. Since the precise
  551             effect of "unbind" is not generically defined, devices should be
  552             considered relevant even after "unbind", however I/O errors
  553             accessing the device should then be handled gracefully.
  555           • Any code that uses device tags for deciding whether a device is
  556             relevant or not most likely needs to be updated to use the new
  557             udev_device_has_current_tag() API (or sd_device_has_current_tag()
  558             in case sd-device is used), to check whether the tag is set at the
  559             moment an uevent is seen (as opposed to the existing
  560             udev_device_has_tag() API which checks if the tag ever existed on
  561             the device, following the API concept redefinition explained
  562             above).
  564           We are very sorry for this breakage and the requirement to update
  565           packages using these interfaces. We'd again like to underline that
  566           this is not caused by systemd/udev changes, but result of a kernel
  567           behaviour change.
  569         * UPCOMING INCOMPATIBILITY: So far most downstream distribution
  570           packages have not retriggered devices once the udev package (or any
  571           auxiliary package installing additional udev rules) is updated. We
  572           intend to work with major distributions to change this, so that
  573           "udevadm trigger -a change" is issued on such upgrades, ensuring that
  574           the updated ruleset is applied to the devices already discovered, so
  575           that (asynchronously) after the upgrade completed the udev database
  576           is consistent with the updated rule set. This means udev rules must
  577           be ready to be retriggered with a "change" action any time, and
  578           result in correct and complete udev database entries. While the
  579           majority of udev rule files known to us currently get this right,
  580           some don't. Specifically, there are udev rules files included in
  581           various packages that only set udev properties on the "add" action,
  582           but do not handle the "change" action. If a device matching those
  583           rules is retriggered with the "change" action (as is intended here)
  584           it would suddenly lose the relevant properties. This always has been
  585           problematic, but as soon as all udev devices are triggered on relevant
  586           package upgrades this will become particularly so. It is strongly
  587           recommended to fix offending rules so that they can handle a "change"
  588           action at any time, and acquire all necessary udev properties even
  589           then. Or in other words: the header guard mentioned above
  590           (ACTION=="remove",GOTO="xyz_end") is the correct approach to handle
  591           this, as it makes sure rules are rerun on "change" correctly, and
  592           accumulate the correct and complete set of udev properties. udev rule
  593           definitions that cannot handle "change" events being triggered at
  594           arbitrary times should be considered buggy.
  596         * The MountAPIVFS= service file setting now defaults to on if
  597           RootImage= and RootDirectory= are used, which means that with those
  598           two settings /proc/, /sys/ and /dev/ are automatically properly set
  599           up for services. Previous behaviour may be restored by explicitly
  600           setting MountAPIVFS=off.
  602         * Since PAM 1.2.0 (2015) configuration snippets may be placed in
  603           /usr/lib/pam.d/ in addition to /etc/pam.d/. If a file exists in the
  604           latter it takes precedence over the former, similar to how most of
  605           systemd's own configuration is handled. Given that PAM stack
  606           definitions are primarily put together by OS vendors/distributions
  607           (though possibly overridden by users), this systemd release moves its
  608           own PAM stack configuration for the "systemd-user" PAM service (i.e.
  609           for the PAM session invoked by the per-user user@.service instance)
  610           from /etc/pam.d/ to /usr/lib/pam.d/. We recommend moving all
  611           packages' vendor versions of their PAM stack definitions from
  612           /etc/pam.d/ to /usr/lib/pam.d/, but if such OS-wide migration is not
  613           desired the location to which systemd installs its PAM stack
  614           configuration may be changed via the -Dpamconfdir Meson option.
  616         * The runtime dependencies on libqrencode, libpcre2, libidn/libidn2,
  617           libpwquality and libcryptsetup have been changed to be based on
  618           dlopen(): instead of regular dynamic library dependencies declared in
  619           the binary ELF headers, these libraries are now loaded on demand
  620           only, if they are available. If the libraries cannot be found the
  621           relevant operations will fail gracefully, or a suitable fallback
  622           logic is chosen. This is supposed to be useful for general purpose
  623           distributions, as it allows minimizing the list of dependencies the
  624           systemd packages pull in, permitting building of more minimal OS
  625           images, while still making use of these "weak" dependencies should
  626           they be installed. Since many package managers automatically
  627           synthesize package dependencies from ELF shared library dependencies,
  628           some additional manual packaging work has to be done now to replace
  629           those (slightly downgraded from "required" to "recommended" or
  630           whatever is conceptually suitable for the package manager). Note that
  631           this change does not alter build-time behaviour: as before the
  632           build-time dependencies have to be installed during build, even if
  633           they now are optional during runtime.
  635         * sd-event.h gained a new call sd_event_add_time_relative() for
  636           installing timers relative to the current time. This is mostly a
  637           convenience wrapper around the pre-existing sd_event_add_time() call
  638           which installs absolute timers.
  640         * sd-event event sources may now be placed in a new "exit-on-failure"
  641           mode, which may be controlled via the new
  642           sd_event_source_get_exit_on_failure() and
  643           sd_event_source_set_exit_on_failure() functions. If enabled, any
  644           failure returned by the event source handler functions will result in
  645           exiting the event loop (unlike the default behaviour of just
  646           disabling the event source but continuing with the event loop). This
  647           feature is useful to set for all event sources that define "primary"
  648           program behaviour (where failure should be fatal) in contrast to
  649           "auxiliary" behaviour (where failure should remain local).
  651         * Most event source types sd-event supports now accept a NULL handler
  652           function, in which case the event loop is exited once the event
  653           source is to be dispatched, using the userdata pointer — converted to
  654           a signed integer — as exit code of the event loop. Previously this
  655           was supported for IO and signal event sources already. Exit event
  656           sources still do not support this (simply because it makes little
  657           sense there, as the event loop is already exiting when they are
  658           dispatched).
  660         * A new per-unit setting RootImageOptions= has been added which allows
  661           tweaking the mount options for any file system mounted as effect of
  662           the RootImage= setting.
  664         * Another new per-unit setting MountImages= has been added, that allows
  665           mounting additional disk images into the file system tree accessible
  666           to the service.
  668         * Timer units gained a new FixedRandomDelay= boolean setting. If
  669           enabled, the random delay configured with RandomizedDelaySec= is
  670           selected in a way that is stable on a given system (though still
  671           different for different units).
  673         * Socket units gained a new setting Timestamping= that takes "us", "ns"
  674           or "off". This controls the SO_TIMESTAMP/SO_TIMESTAMPNS socket
  675           options.
  677         * systemd-repart now generates JSON output when requested with the new
  678           --json= switch.
  680         * systemd-machined's OpenMachineShell() bus call will now pass
  681           additional policy metadata data fields to the PolicyKit
  682           authentication request.
  684         * systemd-tmpfiles gained a new -E switch, which is equivalent to
  685           --exclude-prefix=/dev --exclude-prefix=/proc --exclude=/run
  686           --exclude=/sys. It's particularly useful in combination with --root=,
  687           when operating on OS trees that do not have any of these four runtime
  688           directories mounted, as this means no files below these subtrees are
  689           created or modified, since those mount points should probably remain
  690           empty.
  692         * systemd-tmpfiles gained a new --image= switch which is like --root=,
  693           but takes a disk image instead of a directory as argument. The
  694           specified disk image is mounted inside a temporary mount namespace
  695           and the tmpfiles.d/ drop-ins stored in the image are executed and
  696           applied to the image. systemd-sysusers similarly gained a new
  697           --image= switch, that allows the sysusers.d/ drop-ins stored in the
  698           image to be applied onto the image.
  700         * Similarly, the journalctl command also gained an --image= switch,
  701           which is a quick one-step solution to look at the log data included
  702           in OS disk images.
  704         * journalctl's --output=cat option (which outputs the log content
  705           without any metadata, just the pure text messages) will now make use
  706           of terminal colors when run on a suitable terminal, similarly to the
  707           other output modes.
  709         * JSON group records now support a "description" string that may be
  710           used to add a human-readable textual description to such groups. This
  711           is supposed to match the user's GECOS field which traditionally
  712           didn't have a counterpart for group records.
  714         * The "systemd-dissect" tool that may be used to inspect OS disk images
  715           and that was previously installed to /usr/lib/systemd/ has now been
  716           moved to /usr/bin/, reflecting its updated status of an officially
  717           supported tool with a stable interface. It gained support for a new
  718           --mkdir switch which when combined with --mount has the effect of
  719           creating the directory to mount the image to if it is missing
  720           first. It also gained two new commands --copy-from and --copy-to for
  721           copying files and directories in and out of an OS image without the
  722           need to manually mount it. It also acquired support for a new option
  723           --json= to generate JSON output when inspecting an OS image.
  725         * The cgroup2 file system is now mounted with the
  726           "memory_recursiveprot" mount option, supported since kernel 5.7. This
  727           means that the MemoryLow= and MemoryMin= unit file settings now apply
  728           recursively to whole subtrees.
  730         * systemd-homed now defaults to using the btrfs file system — if
  731           available — when creating home directories in LUKS volumes. This may
  732           be changed with the DefaultFileSystemType= setting in homed.conf.
  733           It's now the default file system in various major distributions and
  734           has the major benefit for homed that it can be grown and shrunk while
  735           mounted, unlike the other contenders ext4 and xfs, which can both be
  736           grown online, but not shrunk (in fact xfs is the technically most
  737           limited option here, as it cannot be shrunk at all).
  739         * JSON user records managed by systemd-homed gained support for
  740           "recovery keys". These are basically secondary passphrases that can
  741           unlock user accounts/home directories. They are computer-generated
  742           rather than user-chosen, and typically have greater entropy.
  743           homectl's --recovery-key= option may be used to add a recovery key to
  744           a user account. The generated recovery key is displayed as a QR code,
  745           so that it can be scanned to be kept in a safe place. This feature is
  746           particularly useful in combination with systemd-homed's support for
  747           FIDO2 or PKCS#11 authentication, as a secure fallback in case the
  748           security tokens are lost. Recovery keys may be entered wherever the
  749           system asks for a password.
  751         * systemd-homed now maintains a "dirty" flag for each LUKS encrypted
  752           home directory which indicates that a home directory has not been
  753           deactivated cleanly when offline. This flag is useful to identify
  754           home directories for which the offline discard logic did not run when
  755           offlining, and where it would be a good idea to log in again to catch
  756           up.
  758         * systemctl gained a new parameter --timestamp= which may be used to
  759           change the style in which timestamps are output, i.e. whether to show
  760           them in local timezone or UTC, or whether to show µs granularity.
  762         * Alibaba's "pouch" container manager is now detected by
  763           systemd-detect-virt, ConditionVirtualization= and similar
  764           constructs. Similar, they now also recognize IBM PowerVM machine
  765           virtualization.
  767         * systemd-nspawn has been reworked to use the /run/host/incoming/ as
  768           place to use for propagating external mounts into the
  769           container. Similarly /run/host/notify is now used as the socket path
  770           for container payloads to communicate with the container manager
  771           using sd_notify(). The container manager now uses the
  772           /run/host/inaccessible/ directory to place "inaccessible" file nodes
  773           of all relevant types which may be used by the container payload as
  774           bind mount source to over-mount inodes to make them inaccessible.
  775           /run/host/container-manager will now be initialized with the same
  776           string as the $container environment variable passed to the
  777           container's PID 1. /run/host/container-uuid will be initialized with
  778           the same string as $container_uuid. This means the /run/host/
  779           hierarchy is now the primary way to make host resources available to
  780           the container. The Container Interface documents these new files and
  781           directories:
  783           https://systemd.io/CONTAINER_INTERFACE
  785         * Support for the "ConditionNull=" unit file condition has been
  786           deprecated and undocumented for 6 years. systemd started to warn
  787           about its use 1.5 years ago. It has now been removed entirely.
  789         * sd-bus.h gained a new API call sd_bus_error_has_names(), which takes
  790           a sd_bus_error struct and a list of error names, and checks if the
  791           error matches one of these names. It's a convenience wrapper that is
  792           useful in cases where multiple errors shall be handled the same way.
  794         * A new system call filter list "@known" has been added, that contains
  795           all system calls known at the time systemd was built.
  797         * Behaviour of system call filter allow lists has changed slightly:
  798           system calls that are contained in @known will result in a EPERM by
  799           default, while those not contained in it result in ENOSYS. This
  800           should improve compatibility because known system calls will thus be
  801           communicated as prohibited, while unknown (and thus newer ones) will
  802           be communicated as not implemented, which hopefully has the greatest
  803           chance of triggering the right fallback code paths in client
  804           applications.
  806         * "systemd-analyze syscall-filter" will now show two separate sections
  807           at the bottom of the output: system calls known during systemd build
  808           time but not included in any of the filter groups shown above, and
  809           system calls defined on the local kernel but known during systemd
  810           build time.
  812         * If the $SYSTEMD_LOG_SECCOMP=1 environment variable is set for
  813           systemd-nspawn all system call filter violations will be logged by
  814           the kernel (audit). This is useful for tracking down system calls
  815           invoked by container payloads that are prohibited by the container's
  816           system call filter policy.
  818         * If the $SYSTEMD_SECCOMP=0 environment variable is set for
  819           systemd-nspawn (and other programs that use seccomp) all seccomp
  820           filtering is turned off.
  822         * Two new unit file settings ProtectProc= and ProcSubset= have been
  823           added that expose the hidepid= and subset= mount options of procfs.
  824           All processes of the unit will only see processes in /proc that are
  825           are owned by the unit's user. This is an important new sandboxing
  826           option that is recommended to be set on all system services. All
  827           long-running system services that are included in systemd itself set
  828           this option now. This option is only supported on kernel 5.8 and
  829           above, since the hidepid= option supported on older kernels was not a
  830           per-mount option but actually applied to the whole PID namespace.
  832         * Socket units gained a new boolean setting FlushPending=. If enabled
  833           all pending socket data/connections are flushed whenever the socket
  834           unit enters the "listening" state, i.e. after the associated service
  835           exited.
  837         * The unit file setting NUMAMask= gained a new "all" value: when used,
  838           all existing NUMA nodes are added to the NUMA mask.
  840         * A new "credentials" logic has been added to system services. This is
  841           a simple mechanism to pass privileged data to services in a safe and
  842           secure way. It's supposed to be used to pass per-service secret data
  843           such as passwords or cryptographic keys but also associated less
  844           private information such as user names, certificates, and similar to
  845           system services. Each credential is identified by a short user-chosen
  846           name and may contain arbitrary binary data. Two new unit file
  847           settings have been added: SetCredential= and LoadCredential=. The
  848           former allows setting a credential to a literal string, the latter
  849           sets a credential to the contents of a file (or data read from a
  850           user-chosen AF_UNIX stream socket). Credentials are passed to the
  851           service via a special credentials directory, one file for each
  852           credential. The path to the credentials directory is passed in a new
  853           $CREDENTIALS_DIRECTORY environment variable. Since the credentials
  854           are passed in the file system they may be easily referenced in
  855           ExecStart= command lines too, thus no explicit support for the
  856           credentials logic in daemons is required (though ideally daemons
  857           would look for the bits they need in $CREDENTIALS_DIRECTORY
  858           themselves automatically, if set). The $CREDENTIALS_DIRECTORY is
  859           backed by unswappable memory if privileges allow it, immutable if
  860           privileges allow it, is accessible only to the service's UID, and is
  861           automatically destroyed when the service stops.
  863         * systemd-nspawn supports the same credentials logic. It can both
  864           consume credentials passed to it via the aforementioned
  865           $CREDENTIALS_DIRECTORY protocol as well as pass these credentials on
  866           to its payload. The service manager/PID 1 has been updated to match
  867           this: it can also accept credentials from the container manager that
  868           invokes it (in fact: any process that invokes it), and passes them on
  869           to its services. Thus, credentials can be propagated recursively down
  870           the tree: from a system's service manager to a systemd-nspawn
  871           service, to the service manager that runs as container payload and to
  872           the service it runs below. Credentials may also be added on the
  873           systemd-nspawn command line, using new --set-credential= and
  874           --load-credential= command line switches that match the
  875           aforementioned service settings.
  877         * systemd-repart gained new settings Format=, Encrypt=, CopyFiles= in
  878           the partition drop-ins which may be used to format/LUKS
  879           encrypt/populate any created partitions. The partitions are
  880           encrypted/formatted/populated before they are registered in the
  881           partition table, so that they appear atomically: either the
  882           partitions do not exist yet or they exist fully encrypted, formatted,
  883           and populated — there is no time window where they are
  884           "half-initialized". Thus the system is robust to abrupt shutdown: if
  885           the tool is terminated half-way during its operations on next boot it
  886           will start from the beginning.
  888         * systemd-repart's --size= operation gained a new "auto" value. If
  889           specified, and operating on a loopback file it is automatically sized
  890           to the minimal size the size constraints permit. This is useful to
  891           use "systemd-repart" as an image builder for minimally sized images.
  893         * systemd-resolved now gained a third IPC interface for requesting name
  894           resolution: besides D-Bus and local DNS to a Varlink
  895           interface is now supported. The nss-resolve NSS module has been
  896           modified to use this new interface instead of D-Bus. Using Varlink
  897           has a major benefit over D-Bus: it works without a broker service,
  898           and thus already during earliest boot, before the dbus daemon has
  899           been started. This means name resolution via systemd-resolved now
  900           works at the same time systemd-networkd operates: from earliest boot
  901           on, including in the initrd.
  903         * systemd-resolved gained support for a new DNSStubListenerExtra=
  904           configuration file setting which may be used to specify additional IP
  905           addresses the built-in DNS stub shall listen on, in addition to the
  906           main one on
  908         * Name lookups issued via systemd-resolved's D-Bus and Varlink
  909           interfaces (and thus also via glibc NSS if nss-resolve is used) will
  910           now honour a trailing dot in the hostname: if specified the search
  911           path logic is turned off. Thus "resolvectl query foo." is now
  912           equivalent to "resolvectl query --search=off foo.".
  914         * systemd-resolved gained a new D-Bus property "ResolvConfMode" that
  915           exposes how /etc/resolv.conf is currently managed: by resolved (and
  916           in which mode if so) or another subsystem. "resolvctl" will display
  917           this property in its status output.
  919         * The resolv.conf snippets systemd-resolved provides will now set "."
  920           as the search domain if no other search domain is known. This turns
  921           off the derivation of an implicit search domain by nss-dns for the
  922           hostname, when the hostname is set to an FQDN. This change is done to
  923           make nss-dns using resolv.conf provided by systemd-resolved behave
  924           more similarly to nss-resolve.
  926         * systemd-tmpfiles' file "aging" logic (i.e. the automatic clean-up of
  927           /tmp/ and /var/tmp/ based on file timestamps) now looks at the
  928           "birth" time (btime) of a file in addition to the atime, mtime, and
  929           ctime.
  931         * systemd-analyze gained a new verb "capability" that lists all known
  932           capabilities by the systemd build and by the kernel.
  934         * If a file /usr/lib/clock-epoch exists, PID 1 will read its mtime and
  935           advance the system clock to it at boot if it is noticed to be before
  936           that time. Previously, PID 1 would only advance the time to an epoch
  937           time that is set during build-time. With this new file OS builders
  938           can change this epoch timestamp on individual OS images without
  939           having to rebuild systemd.
  941         * systemd-logind will now listen to the KEY_RESTART key from the Linux
  942           input layer and reboot the system if it is pressed, similarly to how
  943           it already handles KEY_POWER, KEY_SUSPEND or KEY_SLEEP. KEY_RESTART
  944           was originally defined in the Multimedia context (to restart playback
  945           of a song or film), but is now primarily used in various embedded
  946           devices for "Reboot" buttons. Accordingly, systemd-logind will now
  947           honour it as such. This may configured in more detail via the new
  948           HandleRebootKey= and RebootKeyIgnoreInhibited=.
  950         * systemd-nspawn/systemd-machined will now reconstruct hardlinks when
  951           copying OS trees, for example in "systemd-nspawn --ephemeral",
  952           "systemd-nspawn --template=", "machinectl clone" and similar. This is
  953           useful when operating with OSTree images, which use hardlinks heavily
  954           throughout, and where such copies previously resulting in "exploding"
  955           hardlinks.
  957         * systemd-nspawn's --console= setting gained support for a new
  958           "autopipe" value, which is identical to "interactive" when invoked on
  959           a TTY, and "pipe" otherwise.
  961         * systemd-networkd's .network files gained support for explicitly
  962           configuring the multicast membership entries of bridge devices in the
  963           [BridgeMDB] section. It also gained support for the PIE queuing
  964           discipline in the [FlowQueuePIE] sections.
  966         * systemd-networkd's .netdev files may now be used to create "BareUDP"
  967           tunnels, configured in the new [BareUDP] setting.
  969         * systemd-networkd's Gateway= setting in .network files now accepts the
  970           special values "_dhcp4" and "_ipv6ra" to configure additional,
  971           locally defined, explicit routes to the gateway acquired via DHCP or
  972           IPv6 Router Advertisements. The old setting "_dhcp" is deprecated,
  973           but still accepted for backwards compatibility.
  975         * systemd-networkd's [IPv6PrefixDelegation] section and
  976           IPv6PrefixDelegation= options have been renamed as [IPv6SendRA] and
  977           IPv6SendRA= (the old names are still accepted for backwards
  978           compatibility).
  980         * systemd-networkd's .network files gained the DHCPv6PrefixDelegation=
  981           boolean setting in [Network] section. If enabled, the delegated prefix
  982           gained by another link will be configured, and an address within the
  983           prefix will be assigned.
  985         * systemd-networkd's .network files gained the Announce= boolean setting
  986           in [DHCPv6PrefixDelegation] section. When enabled, the delegated
  987           prefix will be announced through IPv6 router advertisement (IPv6 RA).
  988           The setting is enabled by default.
  990         * VXLAN tunnels may now be marked as independent of any underlying
  991           network interface via the new Independent= boolean setting.
  993         * systemctl gained support for two new verbs: "service-log-level" and
  994           "service-log-target" may be used on services that implement the
  995           generic org.freedesktop.LogControl1 D-Bus interface to dynamically
  996           adjust the log level and target. All of systemd's long-running
  997           services support this now, but ideally all system services would
  998           implement this interface to make the system more uniformly
  999           debuggable.
 1001         * The SystemCallErrorNumber= unit file setting now accepts the new
 1002           "kill" and "log" actions, in addition to arbitrary error number
 1003           specifications as before. If "kill" the processes are killed on the
 1004           event, if "log" the offending system call is audit logged.
 1006         * A new SystemCallLog= unit file setting has been added that accepts a
 1007           list of system calls that shall be logged about (audit).
 1009         * The OS image dissection logic (as used by RootImage= in unit files or
 1010           systemd-nspawn's --image= switch) has gained support for identifying
 1011           and mounting explicit /usr/ partitions, which are now defined in the
 1012           discoverable partition specification. This should be useful for
 1013           environments where the root file system is
 1014           generated/formatted/populated dynamically on first boot and combined
 1015           with an immutable /usr/ tree that is supplied by the vendor.
 1017         * In the final phase of shutdown, within the systemd-shutdown binary
 1018           we'll now try to detach MD devices (i.e software RAID) in addition to
 1019           loopback block devices and DM devices as before. This is supposed to
 1020           be a safety net only, in order to increase robustness if things go
 1021           wrong. Storage subsystems are expected to properly detach their
 1022           storage volumes during regular shutdown already (or in case of
 1023           storage backing the root file system: in the initrd hook we return to
 1024           later).
 1026         * If the SYSTEMD_LOG_TID environment variable is set all systemd tools
 1027           will now log the thread ID in their log output. This is useful when
 1028           working with heavily threaded programs.
 1030         * If the SYSTEMD_RDRAND environment variable is set to "0", systemd will
 1031           not use the RDRAND CPU instruction. This is useful in environments
 1032           such as replay debuggers where non-deterministic behaviour is not
 1033           desirable.
 1035         * The autopaging logic in systemd's various tools (such as systemctl)
 1036           has been updated to turn on "secure" mode in "less"
 1037           (i.e. $LESSECURE=1) if execution in a "sudo" environment is
 1038           detected. This disables invoking external programs from the pager,
 1039           via the pipe logic. This behaviour may be overridden via the new
 1040           $SYSTEMD_PAGERSECURE environment variable.
 1042         * Units which have resource limits (.service, .mount, .swap, .slice,
 1043           .socket, and .slice) gained new configuration settings
 1044           ManagedOOMSwap=, ManagedOOMMemoryPressure=, and
 1045           ManagedOOMMemoryPressureLimitPercent= that specify resource pressure
 1046           limits and optional action taken by systemd-oomd.
 1048         * A new service systemd-oomd has been added. It monitors resource
 1049           contention for selected parts of the unit hierarchy using the PSI
 1050           information reported by the kernel, and kills processes when memory
 1051           or swap pressure is above configured limits. This service is only
 1052           enabled by default in developer mode (see below) and should be
 1053           considered a preview in this release. Behaviour details and option
 1054           names are subject to change without the usual backwards-compatibility
 1055           promises.
 1057         * A new helper oomctl has been added to introspect systemd-oomd state.
 1058           It is only enabled by default in developer mode and should be
 1059           considered a preview without the usual backwards-compatibility
 1060           promises.
 1062         * New meson option -Dcompat-mutable-uid-boundaries= has been added. If
 1063           enabled, systemd reads the system UID boundaries from /etc/login.defs
 1064           at runtime, instead of using the built-in values selected during
 1065           build. This is an option to improve compatibility for upgrades from
 1066           old systems. It's strongly recommended not to make use of this
 1067           functionality on new systems (or even enable it during build), as it
 1068           makes something runtime-configurable that is mostly an implementation
 1069           detail of the OS, and permits avoidable differences in deployments
 1070           that create all kinds of problems in the long run.
 1072         * New meson option '-Dmode=developer|release' has been added. When
 1073           'developer', additional checks and features are enabled that are
 1074           relevant during upstream development, e.g. verification that
 1075           semi-automatically-generated documentation has been properly updated
 1076           following API changes. Those checks are considered hints for
 1077           developers and are not actionable in downstream builds. In addition,
 1078           extra features that are not ready for general consumption may be
 1079           enabled in developer mode. It is thus recommended to set
 1080           '-Dmode=release' in end-user and distro builds.
 1082         * systemd-cryptsetup gained support for processing detached LUKS
 1083           headers specified on the kernel command line via the header=
 1084           parameter of the luks.options= kernel command line option. The same
 1085           device/path syntax as for key files is supported for header files
 1086           like this.
 1088         * The "net_id" built-in of udev has been updated to ignore ACPI _SUN
 1089           slot index data for devices that are connected through a PCI bridge
 1090           where the _SUN index is associated with the bridge instead of the
 1091           network device itself. Previously this would create ambiguous device
 1092           naming if multiple network interfaces were connected to the same PCI
 1093           bridge. Since this is a naming scheme incompatibility on systems that
 1094           possess hardware like this it has been introduced as new naming
 1095           scheme "v247". The previous scheme can be selected via the
 1096           "net.naming-scheme=v245" kernel command line parameter.
 1098         * ConditionFirstBoot= semantics have been modified to be safe towards
 1099           abnormal system power-off during first boot. Specifically, the
 1100           "systemd-machine-id-commit.service" service now acts as boot
 1101           milestone indicating when the first boot process is sufficiently
 1102           complete in order to not consider the next following boot also a
 1103           first boot. If the system is reset before this unit is reached the
 1104           first time, the next boot will still be considered a first boot; once
 1105           it has been reached, no further boots will be considered a first
 1106           boot. The "first-boot-complete.target" unit now acts as official hook
 1107           point to order against this. If a service shall be run on every boot
 1108           until the first boot fully succeeds it may thus be ordered before
 1109           this target unit (and pull it in) and carry ConditionFirstBoot=
 1110           appropriately.
 1112         * bootctl's set-default and set-oneshot commands now accept the three
 1113           special strings "@default", "@oneshot", "@current" in place of a boot
 1114           entry id. These strings are resolved to the current default and
 1115           oneshot boot loader entry, as well as the currently booted one. Thus
 1116           a command "bootctl set-default @current" may be used to make the
 1117           currently boot menu item the new default for all subsequent boots.
 1119         * "systemctl edit" has been updated to show the original effective unit
 1120           contents in commented form in the text editor.
 1122         * Units in user mode are now segregated into three new slices:
 1123           session.slice (units that form the core of graphical session),
 1124           app.slice ("normal" user applications), and background.slice
 1125           (low-priority tasks). Unless otherwise configured, user units are
 1126           placed in app.slice. The plan is to add resource limits and
 1127           protections for the different slices in the future.
 1129         * New GPT partition types for RISCV32/64 for the root and /usr
 1130           partitions, and their associated Verity partitions have been defined,
 1131           and are now understood by systemd-gpt-auto-generator, and the OS
 1132           image dissection logic.
 1134         Contributions from: Adolfo Jayme Barrientos, afg, Alec Moskvin, Alyssa
 1135         Ross, Amitanand Chikorde, Andrew Hangsleben, Anita Zhang, Ansgar
 1136         Burchardt, Arian van Putten, Aurelien Jarno, Axel Rasmussen, bauen1,
 1137         Beniamino Galvani, Benjamin Berg, Bjørn Mork, brainrom, Chandradeep
 1138         Dey, Charles Lee, Chris Down, Christian Göttsche, Christof Efkemann,
 1139         Christoph Ruegge, Clemens Gruber, Daan De Meyer, Daniele Medri, Daniel
 1140         Mack, Daniel Rusek, Dan Streetman, David Tardon, Dimitri John Ledkov,
 1141         Dmitry Borodaenko, Elias Probst, Elisei Roca, ErrantSpore, Etienne
 1142         Doms, Fabrice Fontaine, fangxiuning, Felix Riemann, Florian Klink,
 1143         Franck Bui, Frantisek Sumsal, fwSmit, George Rawlinson, germanztz,
 1144         Gibeom Gwon, Glen Whitney, Gogo Gogsi, Göran Uddeborg, Grant Mathews,
 1145         Hans de Goede, Hans Ulrich Niedermann, Haochen Tong, Harald Seiler,
 1146         huangyong, Hubert Kario, igo95862, Ikey Doherty, Insun Pyo, Jan Chren,
 1147         Jan Schlüter, Jérémy Nouhaud, Jian-Hong Pan, Joerg Behrmann, Jonathan
 1148         Lebon, Jörg Thalheim, Josh Brobst, Juergen Hoetzel, Julien Humbert,
 1149         Kai-Chuan Hsieh, Kairui Song, Kamil Dudka, Kir Kolyshkin, Kristijan
 1150         Gjoshev, Kyle Huey, Kyle Russell, Lee Whalen, Lennart Poettering,
 1151         lichangze, Luca Boccassi, Lucas Werkmeister, Luca Weiss, Marc
 1152         Kleine-Budde, Marco Wang, Martin Wilck, Marti Raudsepp, masmullin2000,
 1153         Máté Pozsgay, Matt Fenwick, Michael Biebl, Michael Scherer, Michal
 1154         Koutný, Michal Sekletár, Michal Suchanek, Mikael Szreder, Milo
 1155         Casagrande, mirabilos, Mitsuha_QuQ, mog422, Muhammet Kara, Nazar
 1156         Vinnichuk, Nicholas Narsing, Nicolas Fella, Njibhu, nl6720, Oğuz Ersen,
 1157         Olivier Le Moal, Ondrej Kozina, onlybugreports, Pass Automated Testing
 1158         Suite, Pat Coulthard, Pavel Sapezhko, Pedro Ruiz, perry_yuan, Peter
 1159         Hutterer, Phaedrus Leeds, PhoenixDiscord, Piotr Drąg, Plan C,
 1160         Purushottam choudhary, Rasmus Villemoes, Renaud Métrich, Robert Marko,
 1161         Roman Beranek, Ronan Pigott, Roy Chen (陳彥廷), RussianNeuroMancer,
 1162         Samanta Navarro, Samuel BF, scootergrisen, Sorin Ionescu, Steve Dodd,
 1163         Susant Sahani, Timo Rothenpieler, Tobias Hunger, Tobias Kaufmann, Topi
 1164         Miettinen, vanou, Vito Caputo, Weblate, Wen Yang, Whired Planck,
 1165         williamvds, Yu, Li-Yu, Yuri Chornoivan, Yu Watanabe, Zbigniew
 1166         Jędrzejewski-Szmek, Zmicer Turok, Дамјан Георгиевски
 1168         – Warsaw, 2020-11-26
 1170 CHANGES WITH 246:
 1172         * The service manager gained basic support for cgroup v2 freezer. Units
 1173           can now be suspended or resumed either using new systemctl verbs,
 1174           freeze and thaw respectively, or via D-Bus.
 1176         * PID 1 may now automatically load pre-compiled AppArmor policies from
 1177           /etc/apparmor/earlypolicy during early boot.
 1179         * The CPUAffinity= setting in service unit files now supports a new
 1180           special value "numa" that causes the CPU affinity masked to be set
 1181           based on the NUMA mask.
 1183         * systemd will now log about all left-over processes remaining in a
 1184           unit when the unit is stopped. It will now warn about services using
 1185           KillMode=none, as this is generally an unsafe thing to make use of.
 1187         * Two new unit file settings
 1188           ConditionPathIsEncrypted=/AssertPathIsEncrypted= have been
 1189           added. They may be used to check whether a specific file system path
 1190           resides on a block device that is encrypted on the block level
 1191           (i.e. using dm-crypt/LUKS).
 1193         * Another pair of new settings ConditionEnvironment=/AssertEnvironment=
 1194           has been added that may be used for simple environment checks. This
 1195           is particularly useful when passing in environment variables from a
 1196           container manager (or from PAM in case of the systemd --user
 1197           instance).
 1199         * .service unit files now accept a new setting CoredumpFilter= which
 1200           allows configuration of the memory sections coredumps of the
 1201           service's processes shall include.
 1203         * .mount units gained a new ReadWriteOnly= boolean option. If set
 1204           it will not be attempted to mount a file system read-only if mounting
 1205           in read-write mode doesn't succeed. An option x-systemd.rw-only is
 1206           available in /etc/fstab to control the same.
 1208         * .socket units gained a new boolean setting PassPacketInfo=. If
 1209           enabled, the kernel will attach additional per-packet metadata to all
 1210           packets read from the socket, as an ancillary message. This controls
 1211           the IP_PKTINFO, IPV6_RECVPKTINFO, NETLINK_PKTINFO socket options,
 1212           depending on socket type.
 1214         * .service units gained a new setting RootHash= which may be used to
 1215           specify the root hash for verity enabled disk images which are
 1216           specified in RootImage=. RootVerity= may be used to specify a path to
 1217           the Verity data matching a RootImage= file system. (The latter is
 1218           only useful for images that do not contain the Verity data embedded
 1219           into the same image that carries a GPT partition table following the
 1220           Discoverable Partition Specification). Similarly, systemd-nspawn
 1221           gained a new switch --verity-data= that takes a path to a file with
 1222           the verity data of the disk image supplied in --image=, if the image
 1223           doesn't contain the verity data itself.
 1225         * .service units gained a new setting RootHashSignature= which takes
 1226           either a base64 encoded PKCS#7 signature of the root hash specified
 1227           with RootHash=, or a path to a file to read the signature from. This
 1228           allows validation of the root hash against public keys available in
 1229           the kernel keyring, and is only supported on recent kernels
 1230           (>= 5.4)/libcryptsetup (>= 2.30). A similar switch has been added to
 1231           systemd-nspawn and systemd-dissect (--root-hash-sig=). Support for
 1232           this mechanism has also been added to systemd-veritysetup.
 1234         * .service unit files gained two new options
 1235           TimeoutStartFailureMode=/TimeoutStopFailureMode= that may be used to
 1236           tune behaviour if a start or stop timeout is hit, i.e. whether to
 1237           terminate the service with SIGTERM, SIGABRT or SIGKILL.
 1239         * Most options in systemd that accept hexadecimal values prefixed with
 1240           0x in additional to the usual decimal notation now also support octal
 1241           notation when the 0o prefix is used and binary notation if the 0b
 1242           prefix is used.
 1244         * Various command line parameters and configuration file settings that
 1245           configure key or certificate files now optionally take paths to
 1246           AF_UNIX sockets in the file system. If configured that way a stream
 1247           connection is made to the socket and the required data read from
 1248           it. This is a simple and natural extension to the existing regular
 1249           file logic, and permits other software to provide keys or
 1250           certificates via simple IPC services, for example when unencrypted
 1251           storage on disk is not desired. Specifically, systemd-networkd's
 1252           Wireguard and MACSEC key file settings as well as
 1253           systemd-journal-gatewayd's and systemd-journal-remote's PEM
 1254           key/certificate parameters support this now.
 1256         * Unit files, tmpfiles.d/ snippets, sysusers.d/ snippets and other
 1257           configuration files that support specifier expansion learnt six new
 1258           specifiers: %a resolves to the current architecture, %o/%w/%B/%W
 1259           resolve to the various ID fields from /etc/os-release, %l resolves to
 1260           the "short" hostname of the system, i.e. the hostname configured in
 1261           the kernel truncated at the first dot.
 1263         * Support for the .include syntax in unit files has been removed. The
 1264           concept has been obsolete for 6 years and we started warning about
 1265           its pending removal 2 years ago (also see NEWS file below). It's
 1266           finally gone now.
 1268         * StandardError= and StandardOutput= in unit files no longer support
 1269           the "syslog" and "syslog-console" switches. They were long removed
 1270           from the documentation, but will now result in warnings when used,
 1271           and be converted to "journal" and "journal+console" automatically.
 1273         * If the service setting User= is set to the "nobody" user, a warning
 1274           message is now written to the logs (but the value is nonetheless
 1275           accepted). Setting User=nobody is unsafe, since the primary purpose
 1276           of the "nobody" user is to own all files whose owner cannot be mapped
 1277           locally. It's in particular used by the NFS subsystem and in user
 1278           namespacing. By running a service under this user's UID it might get
 1279           read and even write access to all these otherwise unmappable files,
 1280           which is quite likely a major security problem.
 1282         * tmpfs mounts automatically created by systemd (/tmp, /run, /dev/shm,
 1283           and others) now have a size and inode limits applied (50% of RAM for
 1284           /tmp and /dev/shm, 10% of RAM for other mounts, etc.). Please note
 1285           that the implicit kernel default is 50% too, so there is no change
 1286           in the size limit for /tmp and /dev/shm.
 1288         * nss-mymachines lost support for resolution of users and groups, and
 1289           now only does resolution of hostnames. This functionality is now
 1290           provided by nss-systemd. Thus, the 'mymachines' entry should be
 1291           removed from the 'passwd:' and 'group:' lines in /etc/nsswitch.conf
 1292           (and 'systemd' added if it is not already there).
 1294         * A new kernel command line option systemd.hostname= has been added
 1295           that allows controlling the hostname that is initialized early during
 1296           boot.
 1298         * A kernel command line option "udev.blockdev_read_only" has been
 1299           added. If specified all hardware block devices that show up are
 1300           immediately marked as read-only by udev. This option is useful for
 1301           making sure that a specific boot under no circumstances modifies data
 1302           on disk. Use "blockdev --setrw" to undo the effect of this, per
 1303           device.
 1305         * A new boolean kernel command line option systemd.swap= has been
 1306           added, which may be used to turn off automatic activation of swap
 1307           devices listed in /etc/fstab.
 1309         * New kernel command line options systemd.condition-needs-update= and
 1310           systemd.condition-first-boot= have been added, which override the
 1311           result of the ConditionNeedsUpdate= and ConditionFirstBoot=
 1312           conditions.
 1314         * A new kernel command line option systemd.clock-usec= has been added
 1315           that allows setting the system clock to the specified time in µs
 1316           since Jan 1st, 1970 early during boot. This is in particular useful
 1317           in order to make test cases more reliable.
 1319         * The fs.suid_dumpable sysctl is set to 2 / "suidsafe". This allows
 1320           systemd-coredump to save core files for suid processes. When saving
 1321           the core file, systemd-coredump will use the effective uid and gid of
 1322           the process that faulted.
 1324         * The /sys/module/kernel/parameters/crash_kexec_post_notifiers file is
 1325           now automatically set to "Y" at boot, in order to enable pstore
 1326           generation for collection with systemd-pstore.
 1328         * We provide a set of udev rules to enable auto-suspend on PCI and USB
 1329           devices that were tested to correctly support it. Previously, this
 1330           was distributed as a set of udev rules, but has now been replaced by
 1331           by a set of hwdb entries (and a much shorter udev rule to take action
 1332           if the device modalias matches one of the new hwdb entries).
 1334           As before, entries are periodically imported from the database
 1335           maintained by the ChromiumOS project. If you have a device that
 1336           supports auto-suspend correctly and where it should be enabled by
 1337           default, please submit a patch that adds it to the database (see
 1338           /usr/lib/udev/hwdb.d/60-autosuspend.hwdb).
 1340         * systemd-udevd gained the new configuration option timeout_signal= as well
 1341           as a corresponding kernel command line option udev.timeout_signal=.
 1342           The option can be used to configure the UNIX signal that the main
 1343           daemon sends to the worker processes on timeout. Setting the signal
 1344           to SIGABRT is useful for debugging.
 1346         * .link files managed by systemd-udevd gained options RxFlowControl=,
 1347           TxFlowControl=, AutoNegotiationFlowControl= in the [Link] section, in
 1348           order to configure various flow control parameters. They also gained
 1349           RxMiniBufferSize= and RxJumboBufferSize= in order to configure jumbo
 1350           frame ring buffer sizes.
 1352         * networkd.conf gained a new boolean setting ManageForeignRoutes=. If
 1353           enabled systemd-networkd manages all routes configured by other tools.
 1355         * .network files managed by systemd-networkd gained a new section
 1356           [SR-IOV], in order to configure SR-IOV capable network devices.
 1358         * systemd-networkd's [IPv6Prefix] section in .network files gained a
 1359           new boolean setting Assign=. If enabled an address from the prefix is
 1360           automatically assigned to the interface.
 1362         * systemd-networkd gained a new section [DHCPv6PrefixDelegation] which
 1363           controls delegated prefixes assigned by DHCPv6 client. The section
 1364           has three settings: SubnetID=, Assign=, and Token=. The setting
 1365           SubnetID= allows explicit configuration of the preferred subnet that
 1366           systemd-networkd's Prefix Delegation logic assigns to interfaces. If
 1367           Assign= is enabled (which is the default) an address from any acquired
 1368           delegated prefix is automatically chosen and assigned to the
 1369           interface. The setting Token= specifies an optional address generation
 1370           mode for Assign=.
 1372         * systemd-networkd's [Network] section gained a new setting
 1373           IPv4AcceptLocal=. If enabled the interface accepts packets with local
 1374           source addresses.
 1376         * systemd-networkd gained support for configuring the HTB queuing
 1377           discipline in the [HierarchyTokenBucket] and
 1378           [HierarchyTokenBucketClass] sections. Similar the "pfifo" qdisc may
 1379           be configured in the [PFIFO] section, "GRED" in
 1380           [GenericRandomEarlyDetection], "SFB" in [StochasticFairBlue], "cake"
 1381           in [CAKE], "PIE" in [PIE], "DRR" in [DeficitRoundRobinScheduler] and
 1382           [DeficitRoundRobinSchedulerClass], "BFIFO" in [BFIFO],
 1383           "PFIFOHeadDrop" in [PFIFOHeadDrop], "PFIFOFast" in [PFIFOFast], "HHF"
 1384           in [HeavyHitterFilter], "ETS" in [EnhancedTransmissionSelection] and
 1385           "QFQ" in [QuickFairQueueing] and [QuickFairQueueingClass].
 1387         * systemd-networkd gained support for a new Termination= setting in the
 1388           [CAN] section for configuring the termination resistor. It also
 1389           gained a new ListenOnly= setting for controlling whether to only
 1390           listen on CAN interfaces, without interfering with traffic otherwise
 1391           (which is useful for debugging/monitoring CAN network
 1392           traffic). DataBitRate=, DataSamplePoint=, FDMode=, FDNonISO= have
 1393           been added to configure various CAN-FD aspects.
 1395         * systemd-networkd's [DHCPv6] section gained a new option WithoutRA=.
 1396           When enabled, DHCPv6 will be attempted right-away without requiring an
 1397           Router Advertisement packet suggesting it first (i.e. without the 'M'
 1398           or 'O' flags set). The [IPv6AcceptRA] section gained a boolean option
 1399           DHCPv6Client= that may be used to turn off the DHCPv6 client even if
 1400           the RA packets suggest it.
 1402         * systemd-networkd's [DHCPv4] section gained a new setting UseGateway=
 1403           which may be used to turn off use of the gateway information provided
 1404           by the DHCP lease. A new FallbackLeaseLifetimeSec= setting may be
 1405           used to configure how to process leases that lack a lifetime option.
 1407         * systemd-networkd's [DHCPv4] and [DHCPServer] sections gained a new
 1408           setting SendVendorOption= allowing configuration of additional vendor
 1409           options to send in the DHCP requests/responses. The [DHCPv6] section
 1410           gained a new SendOption= setting for sending arbitrary DHCP
 1411           options. RequestOptions= has been added to request arbitrary options
 1412           from the server. UserClass= has been added to set the DHCP user class
 1413           field.
 1415         * systemd-networkd's [DHCPServer] section gained a new set of options
 1416           EmitPOP3=/POP3=, EmitSMTP=/SMTP=, EmitLPR=/LPR= for including server
 1417           information about these three protocols in the DHCP lease. It also
 1418           gained support for including "MUD" URLs ("Manufacturer Usage
 1419           Description"). Support for "MUD" URLs was also added to the LLDP
 1420           stack, configurable in the [LLDP] section in .network files.
 1422         * The Mode= settings in [MACVLAN] and [MACVTAP] now support 'source'
 1423           mode. Also, the sections now support a new setting SourceMACAddress=.
 1425         * systemd-networkd's .netdev files now support a new setting
 1426           VLANProtocol= in the [Bridge] section that allows configuration of
 1427           the VLAN protocol to use.
 1429         * systemd-networkd supports a new Group= setting in the [Link] section
 1430           of the .network files, to control the link group.
 1432         * systemd-networkd's [Network] section gained a new
 1433           IPv6LinkLocalAddressGenerationMode= setting, which specifies how IPv6
 1434           link local address is generated.
 1436         * A new default .network file is now shipped that matches TUN/TAP
 1437           devices that begin with "vt-" in their name. Such interfaces will
 1438           have IP routing onto the host links set up automatically. This is
 1439           supposed to be used by VM managers to trivially acquire a network
 1440           interface which is fully set up for host communication, simply by
 1441           carefully picking an interface name to use.
 1443         * systemd-networkd's [DHCPv6] section gained a new setting RouteMetric=
 1444           which sets the route priority for routes specified by the DHCP server.
 1446         * systemd-networkd's [DHCPv6] section gained a new setting VendorClass=
 1447           which configures the vendor class information sent to DHCP server.
 1449         * The BlackList= settings in .network files' [DHCPv4] and
 1450           [IPv6AcceptRA] sections have been renamed DenyList=. The old names
 1451           are still understood to provide compatibility.
 1453         * networkctl gained the new "forcerenew" command for forcing all DHCP
 1454           server clients to renew their lease. The interface "status" output
 1455           will now show numerous additional fields of information about an
 1456           interface. There are new "up" and "down" commands to bring specific
 1457           interfaces up or down.
 1459         * systemd-resolved's DNS= configuration option now optionally accepts a
 1460           port number (after ":") and a host name (after "#"). When the host
 1461           name is specified, the DNS-over-TLS certificate is validated to match
 1462           the specified hostname. Additionally, in case of IPv6 addresses, an
 1463           interface may be specified (after "%").
 1465         * systemd-resolved may be configured to forward single-label DNS names.
 1466           This is not standard-conformant, but may make sense in setups where
 1467           public DNS servers are not used.
 1469         * systemd-resolved's DNS-over-TLS support gained SNI validation.
 1471         * systemd-nspawn's --resolv-conf= switch gained a number of new
 1472           supported values. Specifically, options starting with "replace-" are
 1473           like those prefixed "copy-" but replace any existing resolv.conf
 1474           file. And options ending in "-uplink" and "-stub" can now be used to
 1475           propagate other flavours of resolv.conf into the container (as
 1476           defined by systemd-resolved).
 1478         * The various programs included in systemd can now optionally output
 1479           their log messages on stderr prefixed with a timestamp, controlled by
 1480           the $SYSTEMD_LOG_TIME environment variable.
 1482         * systemctl gained a new "-P" switch that is a shortcut for "--value
 1483           --property=…".
 1485         * "systemctl list-units" and "systemctl list-machines" no longer hide
 1486           their first output column with --no-legend. To hide the first column,
 1487           use --plain.
 1489         * "systemctl reboot" takes the option "--reboot-argument=".
 1490           The optional positional argument to "systemctl reboot" is now
 1491           being deprecated in favor of this option.
 1493         * systemd-run gained a new switch --slice-inherit. If specified the
 1494           unit it generates is placed in the same slice as the systemd-run
 1495           process itself.
 1497         * systemd-journald gained support for zstd compression of large fields
 1498           in journal files. The hash tables in journal files have been hardened
 1499           against hash collisions. This is an incompatible change and means
 1500           that journal files created with new systemd versions are not readable
 1501           with old versions. If the $SYSTEMD_JOURNAL_KEYED_HASH boolean
 1502           environment variable for systemd-journald.service is set to 0 this
 1503           new hardening functionality may be turned off, so that generated
 1504           journal files remain compatible with older journalctl
 1505           implementations.
 1507         * journalctl will now include a clickable link in the default output for
 1508           each log message for which an URL with further documentation is
 1509           known. This is only supported on terminal emulators that support
 1510           clickable hyperlinks, and is turned off if a pager is used (since
 1511           "less" still doesn't support hyperlinks,
 1512           unfortunately). Documentation URLs may be included in log messages
 1513           either by including a DOCUMENTATION= journal field in it, or by
 1514           associating a journal message catalog entry with the log message's
 1515           MESSAGE_ID, which then carries a "Documentation:" tag.
 1517         * journald.conf gained a new boolean setting Audit= that may be used to
 1518           control whether systemd-journald will enable audit during
 1519           initialization.
 1521         * when systemd-journald's log stream is broken up into multiple lines
 1522           because the PID of the sender changed this is indicated in the
 1523           generated log records via the _LINE_BREAK=pid-change field.
 1525         * journalctl's "-o cat" output mode will now show one or more journal
 1526           fields specified with --output-fields= instead of unconditionally
 1527           MESSAGE=. This is useful to retrieve a very specific set of fields
 1528           without any decoration.
 1530         * The sd-journal.h API gained two new functions:
 1531           sd_journal_enumerate_available_unique() and
 1532           sd_journal_enumerate_available_data() that operate like their
 1533           counterparts that lack the _available_ in the name, but skip items
 1534           that cannot be read and processed by the local implementation
 1535           (i.e. are compressed in an unsupported format or such),
 1537         * coredumpctl gained a new --file= switch, matching the same one in
 1538           journalctl: a specific journal file may be specified to read the
 1539           coredump data from.
 1541         * coredumps collected by systemd-coredump may now be compressed using
 1542           the zstd algorithm.
 1544         * systemd-binfmt gained a new switch --unregister for unregistering all
 1545           registered entries at once. This is now invoked automatically at
 1546           shutdown, so that binary formats registered with the "F" flag will
 1547           not block clean file system unmounting.
 1549         * systemd-notify's --pid= switch gained new values: "parent", "self",
 1550           "auto" for controlling which PID to send to the service manager: the
 1551           systemd-notify process' PID, or the one of the process invoking it.
 1553         * systemd-logind's Session bus object learnt a new method call
 1554           SetType() for temporarily updating the session type of an already
 1555           allocated session. This is useful for upgrading tty sessions to
 1556           graphical ones once a compositor is invoked.
 1558         * systemd-socket-proxy gained a new switch --exit-idle-time= for
 1559           configuring an exit-on-idle time.
 1561         * systemd-repart's --empty= setting gained a new value "create". If
 1562           specified a new empty regular disk image file is created under the
 1563           specified name. Its size may be specified with the new --size=
 1564           option. The latter is also supported without the "create" mode, in
 1565           order to grow existing disk image files to the specified size. These
 1566           two new options are useful when creating or manipulating disk images
 1567           instead of operating on actual block devices.
 1569         * systemd-repart drop-ins now support a new UUID= setting to control
 1570           the UUID to assign to a newly created partition.
 1572         * systemd-repart's SizeMin= per-partition parameter now defaults to 10M
 1573           instead of 0.
 1575         * systemd-repart's Label= setting now support the usual, simple
 1576           specifier expansion.
 1578         * systemd-homed's LUKS backend gained the ability to discard empty file
 1579           system blocks automatically when the user logs out. This is enabled
 1580           by default to ensure that home directories take minimal space when
 1581           logged out but get full size guarantees when logged in. This may be
 1582           controlled with the new --luks-offline-discard= switch to homectl.
 1584         * If systemd-homed detects that /home/ is encrypted as a whole it will
 1585           now default to the directory or subvolume backends instead of the
 1586           LUKS backend, in order to avoid double encryption. The default
 1587           storage and file system may now be configured explicitly, too, via
 1588           the new /etc/systemd/homed.conf configuration file.
 1590         * systemd-homed now supports unlocking home directories with FIDO2
 1591           security tokens that support the 'hmac-secret' extension, in addition
 1592           to the existing support for PKCS#11 security token unlocking
 1593           support. Note that many recent hardware security tokens support both
 1594           interfaces. The FIDO2 support is accessible via homectl's
 1595           --fido2-device= option.
 1597         * homectl's --pkcs11-uri= setting now accepts two special parameters:
 1598           if "auto" is specified and only one suitable PKCS#11 security token
 1599           is plugged in, its URL is automatically determined and enrolled for
 1600           unlocking the home directory. If "list" is specified a brief table of
 1601           suitable PKCS#11 security tokens is shown. Similar, the new
 1602           --fido2-device= option also supports these two special values, for
 1603           automatically selecting and listing suitable FIDO2 devices.
 1605         * The /etc/crypttab tmp option now optionally takes an argument
 1606           selecting the file system to use. Moreover, the default is now
 1607           changed from ext2 to ext4.
 1609         * There's a new /etc/crypttab option "keyfile-erase". If specified the
 1610           key file listed in the same line is removed after use, regardless if
 1611           volume activation was successful or not. This is useful if the key
 1612           file is only acquired transiently at runtime and shall be erased
 1613           before the system continues to boot.
 1615         * There's also a new /etc/crypttab option "try-empty-password". If
 1616           specified, before asking the user for a password it is attempted to
 1617           unlock the volume with an empty password. This is useful for
 1618           installing encrypted images whose password shall be set on first boot
 1619           instead of at installation time.
 1621         * systemd-cryptsetup will now attempt to load the keys to unlock
 1622           volumes with automatically from files in
 1623           /etc/cryptsetup-keys.d/<volume>.key and
 1624           /run/cryptsetup-keys.d/<volume>.key, if any of these files exist.
 1626         * systemd-cryptsetup may now activate Microsoft BitLocker volumes via
 1627           /etc/crypttab, during boot.
 1629         * logind.conf gained a new RuntimeDirectoryInodesMax= setting to
 1630           control the inode limit for the per-user $XDG_RUNTIME_DIR tmpfs
 1631           instance.
 1633         * A new generator systemd-xdg-autostart-generator has been added. It
 1634           generates systemd unit files from XDG autostart .desktop files, and
 1635           may be used to let the systemd user instance manage services that are
 1636           started automatically as part of the desktop session.
 1638         * "bootctl" gained a new verb "reboot-to-firmware" that may be used
 1639           to query and change the firmware's 'reboot into firmware' setup flag.
 1641         * systemd-firstboot gained a new switch --kernel-command-line= that may
 1642           be used to initialize the /etc/kernel/cmdline file of the image. It
 1643           also gained a new switch --root-password-hashed= which is like
 1644           --root-password= but accepts a pre-hashed UNIX password as
 1645           argument. The new option --delete-root-password may be used to unset
 1646           any password for the root user (dangerous!). The --root-shell= switch
 1647           may be used to control the shell to use for the root account. A new
 1648           --force option may be used to override any already set settings with
 1649           the parameters specified on the command line (by default, the tool
 1650           will not override what has already been set before, i.e. is purely
 1651           incremental).
 1653         * systemd-firstboot gained support for a new --image= switch, which is
 1654           similar to --root= but accepts the path to a disk image file, on
 1655           which it then operates.
 1657         * A new sd-path.h API has been added to libsystemd. It provides a
 1658           simple API for retrieving various search paths and primary
 1659           directories for various resources.
 1661         * A new call sd_notify_barrier() has been added to the sd-daemon.h
 1662           API. The call will block until all previously sent sd_notify()
 1663           messages have been processed by the service manager. This is useful
 1664           to remove races caused by a process already having disappeared at the
 1665           time a notification message is processed by the service manager,
 1666           making correct attribution impossible. The systemd-notify tool will
 1667           now make use of this call implicitly, but this can be turned off again
 1668           via the new --no-block switch.
 1670         * When sending a file descriptor (fd) to the service manager to keep
 1671           track of, using the sd_notify() mechanism, a new parameter FDPOLL=0
 1672           may be specified. If passed the service manager will refrain from
 1673           poll()ing on the file descriptor. Traditionally (and when the
 1674           parameter is not specified), the service manager will poll it for
 1675           POLLHUP or POLLERR events, and immediately close the fds in that
 1676           case.
 1678         * The service manager (PID1) gained a new D-Bus method call
 1679           SetShowStatus() which may be used to control whether it shall show
 1680           boot-time status output on the console. This method has a similar
 1681           effect to sending SIGRTMIN+20/SIGRTMIN+21 to PID 1.
 1683         * The sd-bus API gained a number of convenience functions that take
 1684           va_list arguments rather than "...". For example, there's now
 1685           sd_bus_call_methodv() to match sd_bus_call_method(). Those calls make
 1686           it easier to build wrappers that accept variadic arguments and want
 1687           to pass a ready va_list structure to sd-bus.
 1689         * sd-bus vtable entries can have a new SD_BUS_VTABLE_ABSOLUTE_OFFSET
 1690           flag which alters how the userdata pointer to pass to the callbacks
 1691           is determined. When the flag is set, the offset field is converted
 1692           as-is into a pointer, without adding it to the object pointer the
 1693           vtable is associated with.
 1695         * sd-bus now exposes four new functions:
 1696           sd_bus_interface_name_is_valid() + sd_bus_service_name_is_valid() +
 1697           sd_bus_member_name_is_valid() + sd_bus_object_path_is_valid() will
 1698           validate strings to check if they qualify as various D-Bus concepts.
 1700         * The sd-bus API gained the SD_BUS_METHOD_WITH_ARGS(),
 1702           that simplify adding argument names to D-Bus methods and signals.
 1704         * The man pages for the sd-bus and sd-hwdb APIs have been completed.
 1706         * Various D-Bus APIs of systemd daemons now have man pages that
 1707           document the methods, signals and properties.
 1709         * The expectations on user/group name syntax are now documented in
 1710           detail; documentation on how classic home directories may be
 1711           converted into home directories managed by homed has been added;
 1712           documentation regarding integration of homed/userdb functionality in
 1713           desktops has been added:
 1715               https://systemd.io/USER_NAMES
 1716               https://systemd.io/CONVERTING_TO_HOMED
 1717               https://systemd.io/USERDB_AND_DESKTOPS
 1719         * Documentation for the on-disk Journal file format has been updated
 1720           and has now moved to:
 1722               https://systemd.io/JOURNAL_FILE_FORMAT
 1724         * The interface for containers (https://systemd.io/CONTAINER_INTERFACE)
 1725           has been extended by a set of environment variables that expose
 1726           select fields from the host's os-release file to the container
 1727           payload. Similarly, host's os-release files can be mounted into the
 1728           container underneath /run/host. Together, those mechanisms provide a
 1729           standardized way to expose information about the host to the
 1730           container payload. Both interfaces are implemented in systemd-nspawn.
 1732         * All D-Bus services shipped in systemd now implement the generic
 1733           LogControl1 D-Bus API which allows clients to change log level +
 1734           target of the service during runtime.
 1736         * Only relevant for developers: the mkosi.default symlink has been
 1737           dropped from version control. Please create a symlink to one of the
 1738           distribution-specific defaults in .mkosi/ based on your preference.
 1740         Contributions from: 24bisquitz, Adam Nielsen, Alan Perry, Alexander
 1741         Malafeev, Amitanand.Chikorde, Alin Popa, Alvin Šipraga, Amos Bird,
 1742         Andreas Rammhold, AndreRH, Andrew Doran, Anita Zhang, Ankit Jain,
 1743         antznin, Arnaud Ferraris, Arthur Moraes do Lago, Arusekk, Balaji
 1744         Punnuru, Balint Reczey, Bastien Nocera, bemarek, Benjamin Berg,
 1745         Benjamin Dahlhoff, Benjamin Robin, Chris Down, Chris Kerr, Christian
 1746         Göttsche, Christian Hesse, Christian Oder, Ciprian Hacman, Clinton Roy,
 1747         codicodi, Corey Hinshaw, Daan De Meyer, Dana Olson, Dan Callaghan,
 1748         Daniel Fullmer, Daniel Rusek, Dan Streetman, Dave Reisner, David
 1749         Edmundson, David Wood, Denis Pronin, Diego Escalante Urrelo, Dimitri
 1750         John Ledkov, dolphrundgren, duguxy, Einsler Lee, Elisei Roca, Emmanuel
 1751         Garette, Eric Anderson, Eric DeVolder, Evgeny Vereshchagin,
 1752         ExtinctFire, fangxiuning, Ferran Pallarès Roca, Filipe Brandenburger,
 1753         Filippo Falezza, Finn, Florian Klink, Florian Mayer, Franck Bui,
 1754         Frantisek Sumsal, gaurav, Georg Müller, Gergely Polonkai, Giedrius
 1755         Statkevičius, Gigadoc2, gogogogi, Gaurav Singh, gzjsgdsb, Hans de
 1756         Goede, Haochen Tong, ianhi, ignapk, Jakov Smolic, James T. Lee, Jan
 1757         Janssen, Jan Klötzke, Jan Palus, Jay Burger, Jeremy Cline, Jérémy
 1758         Rosen, Jian-Hong Pan, Jiri Slaby, Joel Shapiro, Joerg Behrmann, Jörg
 1759         Thalheim, Jouke Witteveen, Kai-Heng Feng, Kenny Levinsen, Kevin
 1760         Kuehler, Kumar Kartikeya Dwivedi, layderv, laydervus, Lénaïc Huard,
 1761         Lennart Poettering, Lidong Zhong, Luca Boccassi, Luca BRUNO, Lucas
 1762         Werkmeister, Lukas Klingsbo, Lukáš Nykrýn, Łukasz Stelmach, Maciej
 1763         S. Szmigiero, MadMcCrow, Marc-André Lureau, Marcel Holtmann, Marc
 1764         Kleine-Budde, Martin Hundebøll, Matthew Leeds, Matt Ranostay, Maxim
 1765         Fomin, MaxVerevkin, Michael Biebl, Michael Chapman, Michael Gubbels,
 1766         Michael Marley, Michał Bartoszkiewicz, Michal Koutný, Michal Sekletár,
 1767         Mike Gilbert, Mike Kazantsev, Mikhail Novosyolov, ml, Motiejus Jakštys,
 1768         nabijaczleweli, nerdopolis, Niccolò Maggioni, Niklas Hambüchen, Norbert
 1769         Lange, Paul Cercueil, pelzvieh, Peter Hutterer, Piero La Terza, Pieter
 1770         Lexis, Piotr Drąg, Rafael Fontenelle, Richard Petri, Ronan Pigott, Ross
 1771         Lagerwall, Rubens Figueiredo, satmandu, Sean-StarLabs, Sebastian
 1772         Jennen, sterlinghughes, Surhud More, Susant Sahani, szb512, Thomas
 1773         Haller, Tobias Hunger, Tom, Tomáš Pospíšek, Tomer Shechner, Tom Hughes,
 1774         Topi Miettinen, Tudor Roman, Uwe Kleine-König, Valery0xff, Vito Caputo,
 1775         Vladimir Panteleev, Vladyslav Tronko, Wen Yang, Yegor Vialov, Yigal
 1776         Korman, Yi Gao, YmrDtnJu, Yuri Chornoivan, Yu Watanabe, Zbigniew
 1777         Jędrzejewski-Szmek, Zhu Li, Дамјан Георгиевски, наб
 1779         – Warsaw, 2020-07-30
 1781 CHANGES WITH 245:
 1783         * A new tool "systemd-repart" has been added, that operates as an
 1784           idempotent declarative repartitioner for GPT partition tables.
 1785           Specifically, a set of partitions that must or may exist can be
 1786           configured via drop-in files, and during every boot the partition
 1787           table on disk is compared with these files, creating missing
 1788           partitions or growing existing ones based on configurable relative
 1789           and absolute size constraints. The tool is strictly incremental,
 1790           i.e. does not delete, shrink or move partitions, but only adds and
 1791           grows them. The primary use-case is OS images that ship in minimized
 1792           form, that on first boot are grown to the size of the underlying
 1793           block device or augmented with additional partitions. For example,
 1794           the root partition could be extended to cover the whole disk, or a
 1795           swap or /home partitions could be added on first boot. It can also be
 1796           used for systems that use an A/B update scheme but ship images with
 1797           just the A partition, with B added on first boot. The tool is
 1798           primarily intended to be run in the initrd, shortly before
 1799           transitioning into the host OS, but can also be run after the
 1800           transition took place. It automatically discovers the disk backing
 1801           the root file system, and should hence not require any additional
 1802           configuration besides the partition definition drop-ins. If no
 1803           configuration drop-ins are present, no action is taken.
 1805         * A new component "userdb" has been added, along with a small daemon
 1806           "systemd-userdbd.service" and a client tool "userdbctl". The framework
 1807           allows defining rich user and group records in a JSON format,
 1808           extending on the classic "struct passwd" and "struct group"
 1809           structures. Various components in systemd have been updated to
 1810           process records in this format, including systemd-logind and
 1811           pam-systemd. The user records are intended to be extensible, and
 1812           allow setting various resource management, security and runtime
 1813           parameters that shall be applied to processes and sessions of the
 1814           user as they log in. This facility is intended to allow associating
 1815           such metadata directly with user/group records so that they can be
 1816           produced, extended and consumed in unified form. We hope that
 1817           eventually frameworks such as sssd will generate records this way, so
 1818           that for the first time resource management and various other
 1819           per-user settings can be configured in LDAP directories and then
 1820           provided to systemd (specifically to systemd-logind and pam-system)
 1821           to apply on login. For further details see:
 1823           https://systemd.io/USER_RECORD
 1824           https://systemd.io/GROUP_RECORD
 1825           https://systemd.io/USER_GROUP_API
 1827         * A small new service systemd-homed.service has been added, that may be
 1828           used to securely manage home directories with built-in encryption.
 1829           The complete user record data is unified with the home directory,
 1830           thus making home directories naturally migratable. Its primary
 1831           back-end is based on LUKS volumes, but fscrypt, plain directories,
 1832           and other storage schemes are also supported. This solves a couple of
 1833           problems we saw with traditional ways to manage home directories, in
 1834           particular when it comes to encryption. For further discussion of
 1835           this, see the video of Lennart's talk at AllSystemsGo! 2019:
 1837           https://media.ccc.de/v/ASG2019-164-reinventing-home-directories
 1839           For further details about the format and expectations on home
 1840           directories this new daemon makes, see:
 1842           https://systemd.io/HOME_DIRECTORY
 1844         * systemd-journald is now multi-instantiable. In addition to the main
 1845           instance systemd-journald.service there's now a template unit
 1846           systemd-journald@.service, with each instance defining a new named
 1847           log 'namespace' (whose name is specified via the instance part of the
 1848           unit name). A new unit file setting LogNamespace= has been added,
 1849           taking such a namespace name, that assigns services to the specified
 1850           log namespaces. As each log namespace is serviced by its own
 1851           independent journal daemon, this functionality may be used to improve
 1852           performance and increase isolation of applications, at the price of
 1853           losing global message ordering. Each instance of journald has a
 1854           separate set of configuration files, with possibly different disk
 1855           usage limitations and other settings.
 1857           journalctl now takes a new option --namespace= to show logs from a
 1858           specific log namespace. The sd-journal.h API gained
 1859           sd_journal_open_namespace() for opening the log stream of a specific
 1860           log namespace. systemd-journald also gained the ability to exit on
 1861           idle, which is useful in the context of log namespaces, as this means
 1862           log daemons for log namespaces can be activated automatically on
 1863           demand and will stop automatically when no longer used, minimizing
 1864           resource usage.
 1866         * When systemd-tmpfiles copies a file tree using the 'C' line type it
 1867           will now label every copied file according to the SELinux database.
 1869         * When systemd/PID 1 detects it is used in the initrd it will now boot
 1870           into initrd.target rather than default.target by default. This should
 1871           make it simpler to build initrds with systemd as for many cases the
 1872           only difference between a host OS image and an initrd image now is
 1873           the presence of the /etc/initrd-release file.
 1875         * A new kernel command line option systemd.cpu_affinity= is now
 1876           understood. It's equivalent to the CPUAffinity= option in
 1877           /etc/systemd/system.conf and allows setting the CPU mask for PID 1
 1878           itself and the default for all other processes.
 1880         * When systemd/PID 1 is reloaded (with systemctl daemon-reload or
 1881           equivalent), the SELinux database is now reloaded, ensuring that
 1882           sockets and other file system objects are generated taking the new
 1883           database into account.
 1885         * systemd/PID 1 accepts a new "systemd.show-status=error" setting, and
 1886           "quiet" has been changed to imply that instead of
 1887           "systemd.show-status=auto". In this mode, only messages about errors
 1888           and significant delays in boot are shown on the console.
 1890         * The sd-event.h API gained native support for the new Linux "pidfd"
 1891           concept. This permits watching processes using file descriptors
 1892           instead of PID numbers, which fixes a number of races and makes
 1893           process supervision more robust and efficient. All of systemd's
 1894           components will now use pidfds if the kernel supports it for process
 1895           watching, with the exception of PID 1 itself, unfortunately. We hope
 1896           to move PID 1 to exclusively using pidfds too eventually, but this
 1897           requires some more kernel work first. (Background: PID 1 watches
 1898           processes using waitid() with the P_ALL flag, and that does not play
 1899           together nicely with pidfds yet.)
 1901         * Closely related to this, the sd-event.h API gained two new calls
 1902           sd_event_source_send_child_signal() (for sending a signal to a
 1903           watched process) and sd_event_source_get_child_process_own() (for
 1904           marking a process so that it is killed automatically whenever the
 1905           event source watching it is freed).
 1907         * systemd-networkd gained support for configuring Token Bucket Filter
 1908           (TBF) parameters in its qdisc configuration support. Similarly,
 1909           support for Stochastic Fairness Queuing (SFQ), Controlled-Delay
 1910           Active Queue Management (CoDel), and Fair Queue (FQ) has been added.
 1912         * systemd-networkd gained support for Intermediate Functional Block
 1913           (IFB) network devices.
 1915         * systemd-networkd gained support for configuring multi-path IP routes,
 1916           using the new MultiPathRoute= setting in the [Route] section.
 1918         * systemd-networkd's DHCPv4 client has been updated to support a new
 1919           SendDecline= option. If enabled, duplicate address detection is done
 1920           after a DHCP offer is received from the server. If a conflict is
 1921           detected, the address is declined. The DHCPv4 client also gained
 1922           support for a new RouteMTUBytes= setting that allows to configure the
 1923           MTU size to be used for routes generated from DHCPv4 leases.
 1925         * The PrefixRoute= setting in systemd-networkd's [Address] section of
 1926           .network files has been deprecated, and replaced by AddPrefixRoute=,
 1927           with its sense inverted.
 1929         * The Gateway= setting of [Route] sections of .network files gained
 1930           support for a special new value "_dhcp". If set, the configured
 1931           static route uses the gateway host configured via DHCP.
 1933         * New User= and SuppressPrefixLength= settings have been implemented
 1934           for the [RoutingPolicyRule] section of .network files to configure
 1935           source routing based on UID ranges and prefix length, respectively.
 1937         * sd-bus gained a new API call sd_bus_message_sensitive() that marks a
 1938           D-Bus message object as "sensitive". Those objects are erased from
 1939           memory when they are freed. This concept is intended to be used for
 1940           messages that contain security sensitive data. A new flag
 1941           SD_BUS_VTABLE_SENSITIVE has been introduced as well to mark methods
 1942           in sd-bus vtables, causing any incoming and outgoing messages of
 1943           those methods to be implicitly marked as "sensitive".
 1945         * sd-bus gained a new API call sd_bus_message_dump() for dumping the
 1946           contents of a message (or parts thereof) to standard output for
 1947           debugging purposes.
 1949         * systemd-sysusers gained support for creating users with the primary
 1950           group named differently than the user.
 1952         * systemd-growfs (i.e. the x-systemd.growfs mount option in /etc/fstab)
 1953           gained support for growing XFS partitions. Previously it supported
 1954           only ext4 and btrfs partitions.
 1956         * The support for /etc/crypttab gained a new x-initrd.attach option. If
 1957           set, the specified encrypted volume is unlocked already in the
 1958           initrd. This concept corresponds to the x-initrd.mount option in
 1959           /etc/fstab.
 1961         * systemd-cryptsetup gained native support for unlocking encrypted
 1962           volumes utilizing PKCS#11 smartcards, i.e. for example to bind
 1963           encryption of volumes to YubiKeys. This is exposed in the new
 1964           pkcs11-uri= option in /etc/crypttab.
 1966         * The /etc/fstab support in systemd now supports two new mount options
 1967           x-systemd.{required,wanted}-by=, for explicitly configuring the units
 1968           that the specified mount shall be pulled in by, in place of
 1969           the usual local-fs.target/remote-fs.target.
 1971         * The https://systemd.io/ web site has been relaunched, directly
 1972           populated with most of the documentation included in the systemd
 1973           repository. systemd also acquired a new logo, thanks to Tobias
 1974           Bernard.
 1976         * systemd-udevd gained support for managing "alternative" network
 1977           interface names, as supported by new Linux kernels. For the first
 1978           time this permits assigning multiple (and longer!) names to a network
 1979           interface. systemd-udevd will now by default assign the names
 1980           generated via all supported naming schemes to each interface. This
 1981           may be further tweaked with .link files and the AlternativeName= and
 1982           AlternativeNamesPolicy= settings. Other components of systemd have
 1983           been updated to support the new alternative names wherever
 1984           appropriate. For example, systemd-nspawn will now generate
 1985           alternative interface names for the host-facing side of container
 1986           veth links based on the full container name without truncation.
 1988         * systemd-nspawn interface naming logic has been updated in another way
 1989           too: if the main interface name (i.e. as opposed to new-style
 1990           "alternative" names) based on the container name is truncated, a
 1991           simple hashing scheme is used to give different interface names to
 1992           multiple containers whose names all begin with the same prefix. Since
 1993           this changes the primary interface names pointing to containers if
 1994           truncation happens, the old scheme may still be requested by
 1995           selecting an older naming scheme, via the net.naming-scheme= kernel
 1996           command line option.
 1998         * PrivateUsers= in service files now works in services run by the
 1999           systemd --user per-user instance of the service manager.
 2001         * A new per-service sandboxing option ProtectClock= has been added that
 2002           locks down write access to the system clock. It takes away device
 2003           node access to /dev/rtc as well as the system calls that set the
 2004           system clock and the CAP_SYS_TIME and CAP_WAKE_ALARM capabilities.
 2005           Note that this option does not affect access to auxiliary services
 2006           that allow changing the clock, for example access to
 2007           systemd-timedated.
 2009         * The systemd-id128 tool gained a new "show" verb for listing or
 2010           resolving a number of well-known UUIDs/128bit IDs, currently mostly
 2011           GPT partition table types.
 2013         * The Discoverable Partitions Specification has been updated to support
 2014           /var and /var/tmp partition discovery. Support for this has been
 2015           added to systemd-gpt-auto-generator. For details see:
 2017           https://systemd.io/DISCOVERABLE_PARTITIONS
 2019         * "systemctl list-unit-files" has been updated to show a new column
 2020           with the suggested enablement state based on the vendor preset files
 2021           for the respective units.
 2023         * "systemctl" gained a new option "--with-dependencies". If specified
 2024           commands such as "systemctl status" or "systemctl cat" will now show
 2025           all specified units along with all units they depend on.
 2027         * networkctl gained support for showing per-interface logs in its
 2028           "status" output.
 2030         * systemd-networkd-wait-online gained support for specifying the maximum
 2031           operational state to wait for, and to wait for interfaces to
 2032           disappear.
 2034         * The [Match] section of .link and .network files now supports a new
 2035           option PermanentMACAddress= which may be used to check against the
 2036           permanent MAC address of a network device even if a randomized MAC
 2037           address is used.
 2039         * The [TrafficControlQueueingDiscipline] section in .network files has
 2040           been renamed to [NetworkEmulator] with the "NetworkEmulator" prefix
 2041           dropped from the individual setting names.
 2043         * Any .link and .network files that have an empty [Match] section (this
 2044           also includes empty and commented-out files) will now be
 2045           rejected. systemd-udev and systemd-networkd started warning about
 2046           such files in version 243.
 2048         * systemd-logind will now validate access to the operation of changing
 2049           the virtual terminal via a polkit action. By default, only users
 2050           with at least one session on a local VT are granted permission.
 2052         * When systemd sets up PAM sessions that invoked service processes
 2053           shall run in, the pam_setcred() API is now invoked, thus permitting
 2054           PAM modules to set additional credentials for the processes.
 2056         * portablectl attach/detach verbs now accept --now and --enable options
 2057           to combine attachment with enablement and invocation, or detachment
 2058           with stopping and disablement.
 2060         * UPGRADE ISSUE: a bug where some jobs were trimmed as redundant was
 2061           fixed, which in turn exposed bugs in unit configuration of services
 2062           which have Type=oneshot and should only run once, but do not have
 2063           RemainAfterExit=yes set. Without RemainAfterExit=yes, a one-shot
 2064           service may be started again after exiting successfully, for example
 2065           as a dependency in another transaction. Affected services included
 2066           some internal systemd services (most notably
 2067           systemd-vconsole-setup.service, which was updated to have
 2068           RemainAfterExit=yes), and plymouth-start.service. Please ensure that
 2069           plymouth has been suitably updated or patched before upgrading to
 2070           this systemd release. See
 2071           https://bugzilla.redhat.com/show_bug.cgi?id=1807771 for some
 2072           additional discussion.
 2074         Contributions from: AJ Bagwell, Alin Popa, Andreas Rammhold, Anita
 2075         Zhang, Ansgar Burchardt, Antonio Russo, Arian van Putten, Ashley Davis,
 2076         Balint Reczey, Bart Willems, Bastien Nocera, Benjamin Dahlhoff, Charles
 2077         (Chas) Williams, cheese1, Chris Down, Chris Murphy, Christian Ehrhardt,
 2078         Christian Göttsche, cvoinf, Daan De Meyer, Daniele Medri, Daniel Rusek,
 2079         Daniel Shahaf, Dann Frazier, Dan Streetman, Dariusz Gadomski, David
 2080         Michael, Dimitri John Ledkov, Emmanuel Bourg, Evgeny Vereshchagin,
 2081         ezst036, Felipe Sateler, Filipe Brandenburger, Florian Klink, Franck
 2082         Bui, Fran Dieguez, Frantisek Sumsal, Greg "GothAck" Miell, Guilhem
 2083         Lettron, Guillaume Douézan-Grard, Hans de Goede, HATAYAMA Daisuke, Iain
 2084         Lane, James Buren, Jan Alexander Steffens (heftig), Jérémy Rosen, Jin
 2085         Park, Jun'ichi Nomura, Kai Krakow, Kevin Kuehler, Kevin P. Fleming,
 2086         Lennart Poettering, Leonid Bloch, Leonid Evdokimov, lothrond, Luca
 2087         Boccassi, Lukas K, Lynn Kirby, Mario Limonciello, Mark Deneen, Matthew
 2088         Leeds, Michael Biebl, Michal Koutný, Michal Sekletár, Mike Auty, Mike
 2089         Gilbert, mtron, nabijaczleweli, Naïm Favier, Nate Jones, Norbert Lange,
 2090         Oliver Giles, Paul Davey, Paul Menzel, Peter Hutterer, Piotr Drąg, Rafa
 2091         Couto, Raphael, rhn, Robert Scheck, Rocka, Romain Naour, Ryan Attard,
 2092         Sascha Dewald, Shengjing Zhu, Slava Kardakov, Spencer Michaels, Sylvain
 2093         Plantefeve, Stanislav Angelovič, Susant Sahani, Thomas Haller, Thomas
 2094         Schmitt, Timo Schlüßler, Timo Wilken, Tobias Bernard, Tobias Klauser,
 2095         Tobias Stoeckmann, Topi Miettinen, tsia, WataruMatsuoka, Wieland
 2096         Hoffmann, Wilhelm Schuster, Will Fleming, xduugu, Yong Cong Sin, Yuri
 2097         Chornoivan, Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek, Zeyu
 2098         DONG
 2100         – Warsaw, 2020-03-06
 2102 CHANGES WITH 244:
 2104         * Support for the cpuset cgroups v2 controller has been added.
 2105           Processes may be restricted to specific CPUs using the new
 2106           AllowedCPUs= setting, and to specific memory NUMA nodes using the new
 2107           AllowedMemoryNodes= setting.
 2109         * The signal used in restart jobs (as opposed to e.g. stop jobs) may
 2110           now be configured using a new RestartKillSignal= setting. This
 2111           allows units which signals to request termination to implement
 2112           different behaviour when stopping in preparation for a restart.
 2114         * "systemctl clean" may now be used also for socket, mount, and swap
 2115           units.
 2117         * systemd will also read configuration options from the EFI variable
 2118           SystemdOptions. This may be used to configure systemd behaviour when
 2119           modifying the kernel command line is inconvenient, but configuration
 2120           on disk is read too late, for example for the options related to
 2121           cgroup hierarchy setup. 'bootctl systemd-efi-options' may be used to
 2122           set the EFI variable.
 2124         * systemd will now disable printk ratelimits in early boot. This should
 2125           allow us to capture more logs from the early boot phase where normal
 2126           storage is not available and the kernel ring buffer is used for
 2127           logging. Configuration on the kernel command line has higher priority
 2128           and overrides the systemd setting.
 2130           systemd programs which log to /dev/kmsg directly use internal
 2131           ratelimits to prevent runaway logging. (Normally this is only used
 2132           during early boot, so in practice this change has very little
 2133           effect.)
 2135         * Unit files now support top level dropin directories of the form
 2136           <unit_type>.d/ (e.g. service.d/) that may be used to add configuration
 2137           that affects all corresponding unit files.
 2139         * systemctl gained support for 'stop --job-mode=triggering' which will
 2140           stop the specified unit and any units which could trigger it.
 2142         * Unit status display now includes units triggering and triggered by
 2143           the unit being shown.
 2145         * The RuntimeMaxSec= setting is now supported by scopes, not just
 2146           .service units. This is particularly useful for PAM sessions which
 2147           create a scope unit for the user login. systemd.runtime_max_sec=
 2148           setting may used with the pam_systemd module to limit the duration
 2149           of the PAM session, for example for time-limited logins.
 2151         * A new @pkey system call group is now defined to make it easier to
 2152           allow-list memory protection syscalls for containers and services
 2153           which need to use them.
 2155         * systemd-udevd: removed the 30s timeout for killing stale workers on
 2156           exit. systemd-udevd now waits for workers to finish. The hard-coded
 2157           exit timeout of 30s was too short for some large installations, where
 2158           driver initialization could be prematurely interrupted during initrd
 2159           processing if the root file system had been mounted and init was
 2160           preparing to switch root. If udevd is run without systemd and workers
 2161           are hanging while udevd receives an exit signal, udevd will now exit
 2162           when udev.event_timeout is reached for the last hanging worker. With
 2163           systemd, the exit timeout can additionally be configured using
 2164           TimeoutStopSec= in systemd-udevd.service.
 2166         * udev now provides a program (fido_id) that identifies FIDO CTAP1
 2167           ("U2F")/CTAP2 security tokens based on the usage declared in their
 2168           report and descriptor and outputs suitable environment variables.
 2169           This replaces the externally maintained allow lists of all known
 2170           security tokens that were used previously.
 2172         * Automatically generated autosuspend udev rules for allow-listed
 2173           devices have been imported from the Chromium OS project. This should
 2174           improve power saving with many more devices.
 2176         * udev gained a new "CONST{key}=value" setting that allows matching
 2177           against system-wide constants without forking a helper binary.
 2178           Currently "arch" and "virt" keys are supported.
 2180         * udev now opens CDROMs in non-exclusive mode when querying their
 2181           capabilities. This should fix issues where other programs trying to
 2182           use the CDROM cannot gain access to it, but carries a risk of
 2183           interfering with programs writing to the disk, if they did not open
 2184           the device in exclusive mode as they should.
 2186         * systemd-networkd does not create a default route for IPv4 link local
 2187           addressing anymore. The creation of the route was unexpected and was
 2188           breaking routing in various cases, but people who rely on it being
 2189           created implicitly will need to adjust. Such a route may be requested
 2190           with DefaultRouteOnDevice=yes.
 2192           Similarly, systemd-networkd will not assign a link-local IPv6 address
 2193           when IPv6 link-local routing is not enabled.
 2195         * Receive and transmit buffers may now be configured on links with
 2196           the new RxBufferSize= and TxBufferSize= settings.
 2198         * systemd-networkd may now advertise additional IPv6 routes. A new
 2199           [IPv6RoutePrefix] section with Route= and LifetimeSec= options is
 2200           now supported.
 2202         * systemd-networkd may now configure "next hop" routes using the
 2203           [NextHop] section and Gateway= and Id= settings.
 2205         * systemd-networkd will now retain DHCP config on restarts by default
 2206           (but this may be overridden using the KeepConfiguration= setting).
 2207           The default for SendRelease= has been changed to true.
 2209         * The DHCPv4 client now uses the OPTION_INFORMATION_REFRESH_TIME option
 2210           received from the server.
 2212           The client will use the received SIP server list if UseSIP=yes is
 2213           set.
 2215           The client may be configured to request specific options from the
 2216           server using a new RequestOptions= setting.
 2218           The client may be configured to send arbitrary options to the server
 2219           using a new SendOption= setting.
 2221           A new IPServiceType= setting has been added to configure the "IP
 2222           service type" value used by the client.
 2224         * The DHCPv6 client learnt a new PrefixDelegationHint= option to
 2225           request prefix hints in the DHCPv6 solicitation.
 2227         * The DHCPv4 server may be configured to send arbitrary options using
 2228           a new SendOption= setting.
 2230         * The DHCPv4 server may now be configured to emit SIP server list using
 2231           the new EmitSIP= and SIP= settings.
 2233         * systemd-networkd and networkctl may now renew DHCP leases on demand.
 2234           networkctl has a new 'networkctl renew' verb.
 2236         * systemd-networkd may now reconfigure links on demand. networkctl
 2237           gained two new verbs: "reload" will reload the configuration, and
 2238           "reconfigure DEVICE…" will reconfigure one or more devices.
 2240         * .network files may now match on SSID and BSSID of a wireless network,
 2241           i.e. the access point name and hardware address using the new SSID=
 2242           and BSSID= options. networkctl will display the current SSID and
 2243           BSSID for wireless links.
 2245           .network files may also match on the wireless network type using the
 2246           new WLANInterfaceType= option.
 2248         * systemd-networkd now includes default configuration that enables
 2249           link-local addressing when connected to an ad-hoc wireless network.
 2251         * systemd-networkd may configure the Traffic Control queueing
 2252           disciplines in the kernel using the new
 2253           [TrafficControlQueueingDiscipline] section and Parent=,
 2254           NetworkEmulatorDelaySec=, NetworkEmulatorDelayJitterSec=,
 2255           NetworkEmulatorPacketLimit=, NetworkEmulatorLossRate=,
 2256           NetworkEmulatorDuplicateRate= settings.
 2258         * systemd-tmpfiles gained a new w+ setting to append to files.
 2260         * systemd-analyze dump will now report when the memory configuration in
 2261           the kernel does not match what systemd has configured (usually,
 2262           because some external program has modified the kernel configuration
 2263           on its own).
 2265         * systemd-analyze gained a new --base-time= switch instructs the
 2266           'calendar' verb to resolve times relative to that timestamp instead
 2267           of the present time.
 2269         * journalctl --update-catalog now produces deterministic output (making
 2270           reproducible image builds easier).
 2272         * A new devicetree-overlay setting is now documented in the Boot Loader
 2273           Specification.
 2275         * The default value of the WatchdogSec= setting used in systemd
 2276           services (the ones bundled with the project itself) may be set at
 2277           configuration time using the -Dservice-watchdog= setting. If set to
 2278           empty, the watchdogs will be disabled.
 2280         * systemd-resolved validates IP addresses in certificates now when GnuTLS
 2281           is being used.
 2283         * libcryptsetup >= 2.0.1 is now required.
 2285         * A configuration option -Duser-path= may be used to override the $PATH
 2286           used by the user service manager. The default is again to use the same
 2287           path as the system manager.
 2289         * The systemd-id128 tool gained a new switch "-u" (or "--uuid") for
 2290           outputting the 128bit IDs in UUID format (i.e. in the "canonical
 2291           representation").
 2293         * Service units gained a new sandboxing option ProtectKernelLogs= which
 2294           makes sure the program cannot get direct access to the kernel log
 2295           buffer anymore, i.e. the syslog() system call (not to be confused
 2296           with the API of the same name in libc, which is not affected), the
 2297           /proc/kmsg and /dev/kmsg nodes and the CAP_SYSLOG capability are made
 2298           inaccessible to the service. It's recommended to enable this setting
 2299           for all services that should not be able to read from or write to the
 2300           kernel log buffer, which are probably almost all.
 2302         Contributions from: Aaron Plattner, Alcaro, Anita Zhang, Balint Reczey,
 2303         Bastien Nocera, Baybal Ni, Benjamin Bouvier, Benjamin Gilbert, Carlo
 2304         Teubner, cbzxt, Chen Qi, Chris Down, Christian Rebischke, Claudio
 2305         Zumbo, ClydeByrdIII, crashfistfight, Cyprien Laplace, Daniel Edgecumbe,
 2306         Daniel Gorbea, Daniel Rusek, Daniel Stuart, Dan Streetman, David
 2307         Pedersen, David Tardon, Dimitri John Ledkov, Dominique Martinet, Donald
 2308         A. Cupp Jr, Evgeny Vereshchagin, Fabian Henneke, Filipe Brandenburger,
 2309         Franck Bui, Frantisek Sumsal, Georg Müller, Hans de Goede, Haochen
 2310         Tong, HATAYAMA Daisuke, Iwan Timmer, Jan Janssen, Jan Kundrát, Jan
 2311         Synacek, Jan Tojnar, Jay Strict, Jérémy Rosen, Jóhann B. Guðmundsson,
 2312         Jonas Jelten, Jonas Thelemann, Justin Trudell, J. Xing, Kai-Heng Feng,
 2313         Kenneth D'souza, Kevin Becker, Kevin Kuehler, Lennart Poettering,
 2314         Léonard Gérard, Lorenz Bauer, Luca Boccassi, Maciej Stanczew, Mario
 2315         Limonciello, Marko Myllynen, Mark Stosberg, Martin Wilck, matthiasroos,
 2316         Michael Biebl, Michael Olbrich, Michael Tretter, Michal Sekletar,
 2317         Michal Sekletár, Michal Suchanek, Mike Gilbert, Mike Kazantsev, Nicolas
 2318         Douma, nikolas, Norbert Lange, pan93412, Pascal de Bruijn, Paul Menzel,
 2319         Pavel Hrdina, Peter Wu, Philip Withnall, Piotr Drąg, Rafael Fontenelle,
 2320         Renaud Métrich, Riccardo Schirone, RoadrunnerWMC, Ronan Pigott, Ryan
 2321         Attard, Sebastian Wick, Serge, Siddharth Chandrasekara, Steve Ramage,
 2322         Steve Traylen, Susant Sahani, Thibault Nélis, Tim Teichmann, Tom
 2323         Fitzhenry, Tommy J, Torsten Hilbrich, Vito Caputo, ypf791, Yu Watanabe,
 2324         Zach Smith, Zbigniew Jędrzejewski-Szmek
 2326         – Warsaw, 2019-11-29
 2328 CHANGES WITH 243:
 2330         * This release enables unprivileged programs (i.e. requiring neither
 2331           setuid nor file capabilities) to send ICMP Echo (i.e. ping) requests
 2332           by turning on the "net.ipv4.ping_group_range" sysctl of the Linux
 2333           kernel for the whole UNIX group range, i.e. all processes. This
 2334           change should be reasonably safe, as the kernel support for it was
 2335           specifically implemented to allow safe access to ICMP Echo for
 2336           processes lacking any privileges. If this is not desirable, it can be
 2337           disabled again by setting the parameter to "1 0".
 2339         * Previously, filters defined with SystemCallFilter= would have the
 2340           effect that any calling of an offending system call would terminate
 2341           the calling thread. This behaviour never made much sense, since
 2342           killing individual threads of unsuspecting processes is likely to
 2343           create more problems than it solves. With this release the default
 2344           action changed from killing the thread to killing the whole
 2345           process. For this to work correctly both a kernel version (>= 4.14)
 2346           and a libseccomp version (>= 2.4.0) supporting this new seccomp
 2347           action is required. If an older kernel or libseccomp is used the old
 2348           behaviour continues to be used. This change does not affect any
 2349           services that have no system call filters defined, or that use
 2350           SystemCallErrorNumber= (and thus see EPERM or another error instead
 2351           of being killed when calling an offending system call). Note that
 2352           systemd documentation always claimed that the whole process is
 2353           killed. With this change behaviour is thus adjusted to match the
 2354           documentation.
 2356         * On 64 bit systems, the "kernel.pid_max" sysctl is now bumped to
 2357           4194304 by default, i.e. the full 22bit range the kernel allows, up
 2358           from the old 16bit range. This should improve security and
 2359           robustness, as PID collisions are made less likely (though certainly
 2360           still possible). There are rumours this might create compatibility
 2361           problems, though at this moment no practical ones are known to
 2362           us. Downstream distributions are hence advised to undo this change in
 2363           their builds if they are concerned about maximum compatibility, but
 2364           for everybody else we recommend leaving the value bumped. Besides
 2365           improving security and robustness this should also simplify things as
 2366           the maximum number of allowed concurrent tasks was previously bounded
 2367           by both "kernel.pid_max" and "kernel.threads-max" and now effectively
 2368           only a single knob is left ("kernel.threads-max"). There have been
 2369           concerns that usability is affected by this change because larger PID
 2370           numbers are harder to type, but we believe the change from 5 digits
 2371           to 7 digits doesn't hamper usability.
 2373         * MemoryLow= and MemoryMin= gained hierarchy-aware counterparts,
 2374           DefaultMemoryLow= and DefaultMemoryMin=, which can be used to
 2375           hierarchically set default memory protection values for a particular
 2376           subtree of the unit hierarchy.
 2378         * Memory protection directives can now take a value of zero, allowing
 2379           explicit opting out of a default value propagated by an ancestor.
 2381         * systemd now defaults to the "unified" cgroup hierarchy setup during
 2382           build-time, i.e. -Ddefault-hierarchy=unified is now the build-time
 2383           default. Previously, -Ddefault-hierarchy=hybrid was the default. This
 2384           change reflects the fact that cgroupsv2 support has matured
 2385           substantially in both systemd and in the kernel, and is clearly the
 2386           way forward. Downstream production distributions might want to
 2387           continue to use -Ddefault-hierarchy=hybrid (or even =legacy) for
 2388           their builds as unfortunately the popular container managers have not
 2389           caught up with the kernel API changes.
 2391         * Man pages are not built by default anymore (html pages were already
 2392           disabled by default), to make development builds quicker. When
 2393           building systemd for a full installation with documentation, meson
 2394           should be called with -Dman=true and/or -Dhtml=true as appropriate.
 2395           The default was changed based on the assumption that quick one-off or
 2396           repeated development builds are much more common than full optimized
 2397           builds for installation, and people need to pass various other
 2398           options to when doing "proper" builds anyway, so the gain from making
 2399           development builds quicker is bigger than the one time disruption for
 2400           packagers.
 2402           Two scripts are created in the *build* directory to generate and
 2403           preview man and html pages on demand, e.g.:
 2405           build/man/man systemctl
 2406           build/man/html systemd.index
 2408         * libidn2 is used by default if both libidn2 and libidn are installed.
 2409           Please use -Dlibidn=true if libidn is preferred.
 2411         * The D-Bus "wire format" of the CPUAffinity= attribute is changed on
 2412           big-endian machines. Before, bytes were written and read in native
 2413           machine order as exposed by the native libc __cpu_mask interface.
 2414           Now, little-endian order is always used (CPUs 0–7 are described by
 2415           bits 0–7 in byte 0, CPUs 8–15 are described by byte 1, and so on).
 2416           This change fixes D-Bus calls that cross endianness boundary.
 2418           The presentation format used for CPUAffinity= by "systemctl show" and
 2419           "systemd-analyze dump" is changed to present CPU indices instead of
 2420           the raw __cpu_mask bitmask. For example, CPUAffinity=0-1 would be
 2421           shown as CPUAffinity=03000000000000000000000000000… (on
 2422           little-endian) or CPUAffinity=00000000000000300000000000000… (on
 2423           64-bit big-endian), and is now shown as CPUAffinity=0-1, matching the
 2424           input format. The maximum integer that will be printed in the new
 2425           format is 8191 (four digits), while the old format always used a very
 2426           long number (with the length varying by architecture), so they can be
 2427           unambiguously distinguished.
 2429         * /usr/sbin/halt.local is no longer supported. Implementation in
 2430           distributions was inconsistent and it seems this functionality was
 2431           very rarely used.
 2433           To replace this functionality, users should:
 2434           - either define a new unit and make it a dependency of final.target
 2435             (systemctl add-wants final.target my-halt-local.service)
 2436           - or move the shutdown script to /usr/lib/systemd/system-shutdown/
 2437             and ensure that it accepts "halt", "poweroff", "reboot", and
 2438             "kexec" as an argument, see the description in systemd-shutdown(8).
 2440         * When a [Match] section in .link or .network file is empty (contains
 2441           no match patterns), a warning will be emitted. Please add any "match
 2442           all" pattern instead, e.g. OriginalName=* or Name=* in case all
 2443           interfaces should really be matched.
 2445         * A new setting NUMAPolicy= may be used to set process memory
 2446           allocation policy. This setting can be specified in
 2447           /etc/systemd/system.conf and hence will set the default policy for
 2448           PID1. The default policy can be overridden on a per-service
 2449           basis. The related setting NUMAMask= is used to specify NUMA node
 2450           mask that should be associated with the selected policy.
 2452         * PID 1 will now listen to Out-Of-Memory (OOM) events the kernel
 2453           generates when processes it manages are reaching their memory limits,
 2454           and will place their units in a special state, and optionally kill or
 2455           stop the whole unit.
 2457         * The service manager will now expose bus properties for the IO
 2458           resources used by units. This information is also shown in "systemctl
 2459           status" now (for services that have IOAccounting=yes set). Moreover,
 2460           the IO accounting data is included in the resource log message
 2461           generated whenever a unit stops.
 2463         * Units may now configure an explicit timeout to wait for when killed
 2464           with SIGABRT, for example when a service watchdog is hit. Previously,
 2465           the regular TimeoutStopSec= timeout was applied in this case too —
 2466           now a separate timeout may be set using TimeoutAbortSec=.
 2468         * Services may now send a special WATCHDOG=trigger message with
 2469           sd_notify() to trigger an immediate "watchdog missed" event, and thus
 2470           trigger service termination. This is useful both for testing watchdog
 2471           handling, but also for defining error paths in services, that shall
 2472           be handled the same way as watchdog events.
 2474         * There are two new per-unit settings IPIngressFilterPath= and
 2475           IPEgressFilterPath= which allow configuration of a BPF program
 2476           (usually by specifying a path to a program uploaded to /sys/fs/bpf/)
 2477           to apply to the IP packet ingress/egress path of all processes of a
 2478           unit. This is useful to allow running systemd services with BPF
 2479           programs set up externally.
 2481         * systemctl gained a new "clean" verb for removing the state, cache,
 2482           runtime or logs directories of a service while it is terminated. The
 2483           new verb may also be used to remove the state maintained on disk for
 2484           timer units that have Persistent= configured.
 2486         * During the last phase of shutdown systemd will now automatically
 2487           increase the log level configured in the "kernel.printk" sysctl so
 2488           that any relevant loggable events happening during late shutdown are
 2489           made visible. Previously, loggable events happening so late during
 2490           shutdown were generally lost if the "kernel.printk" sysctl was set to
 2491           high thresholds, as regular logging daemons are terminated at that
 2492           time and thus nothing is written to disk.
 2494         * If processes terminated during the last phase of shutdown do not exit
 2495           quickly systemd will now show their names after a short time, to make
 2496           debugging easier. After a longer timeout they are forcibly killed,
 2497           as before.
 2499         * journalctl (and the other tools that display logs) will now highlight
 2500           warnings in yellow (previously, both LOG_NOTICE and LOG_WARNING where
 2501           shown in bright bold, now only LOG_NOTICE is). Moreover, audit logs
 2502           are now shown in blue color, to separate them visually from regular
 2503           logs. References to configuration files are now turned into clickable
 2504           links on terminals that support that.
 2506         * systemd-journald will now stop logging to /var/log/journal during
 2507           shutdown when /var/ is on a separate mount, so that it can be
 2508           unmounted safely during shutdown.
 2510         * systemd-resolved gained support for a new 'strict' DNS-over-TLS mode.
 2512         * systemd-resolved "Cache=" configuration option in resolved.conf has
 2513           been extended to also accept the 'no-negative' value. Previously,
 2514           only a boolean option was allowed (yes/no), having yes as the
 2515           default. If this option is set to 'no-negative', negative answers are
 2516           not cached while the old cache heuristics are used positive answers.
 2517           The default remains unchanged.
 2519         * The predictable naming scheme for network devices now supports
 2520           generating predictable names for "netdevsim" devices.
 2522           Moreover, the "en" prefix was dropped from the ID_NET_NAME_ONBOARD
 2523           udev property.
 2525           Those two changes form a new net.naming-policy-scheme= entry.
 2526           Distributions which want to preserve naming stability may want to set
 2527           the -Ddefault-net-naming-scheme= configuration option.
 2529         * systemd-networkd now supports MACsec, nlmon, IPVTAP and Xfrm
 2530           interfaces natively.
 2532         * systemd-networkd's bridge FDB support now allows configuration of a
 2533           destination address for each entry (Destination=), as well as the
 2534           VXLAN VNI (VNI=), as well as an option to declare what an entry is
 2535           associated with (AssociatedWith=).
 2537         * systemd-networkd's DHCPv4 support now understands a new MaxAttempts=
 2538           option for configuring the maximum number of DHCP lease requests.  It
 2539           also learnt a new BlackList= option for deny-listing DHCP servers (a
 2540           similar setting has also been added to the IPv6 RA client), as well
 2541           as a SendRelease= option for configuring whether to send a DHCP
 2542           RELEASE message when terminating.
 2544         * systemd-networkd's DHCPv4 and DHCPv6 stacks can now be configured
 2545           separately in the [DHCPv4] and [DHCPv6] sections.
 2547         * systemd-networkd's DHCP support will now optionally create an
 2548           implicit host route to the DNS server specified in the DHCP lease, in
 2549           addition to the routes listed explicitly in the lease. This should
 2550           ensure that in multi-homed systems DNS traffic leaves the systems on
 2551           the interface that acquired the DNS server information even if other
 2552           routes such as default routes exist. This behaviour may be turned on
 2553           with the new RoutesToDNS= option.
 2555         * systemd-networkd's VXLAN support gained a new option
 2556           GenericProtocolExtension= for enabling VXLAN Generic Protocol
 2557           Extension support, as well as IPDoNotFragment= for setting the IP
 2558           "Don't fragment" bit on outgoing packets. A similar option has been
 2559           added to the GENEVE support.
 2561         * In systemd-networkd's [Route] section you may now configure
 2562           FastOpenNoCookie= for configuring per-route TCP fast-open support, as
 2563           well as TTLPropagate= for configuring Label Switched Path (LSP) TTL
 2564           propagation. The Type= setting now supports local, broadcast,
 2565           anycast, multicast, any, xresolve routes, too.
 2567         * systemd-networkd's [Network] section learnt a new option
 2568           DefaultRouteOnDevice= for automatically configuring a default route
 2569           onto the network device.
 2571         * systemd-networkd's bridging support gained two new options ProxyARP=
 2572           and ProxyARPWifi= for configuring proxy ARP behaviour as well as
 2573           MulticastRouter= for configuring multicast routing behaviour. A new
 2574           option MulticastIGMPVersion= may be used to change bridge's multicast
 2575           Internet Group Management Protocol (IGMP) version.
 2577         * systemd-networkd's FooOverUDP support gained the ability to configure
 2578           local and peer IP addresses via Local= and Peer=. A new option
 2579           PeerPort= may be used to configure the peer's IP port.
 2581         * systemd-networkd's TUN support gained a new setting VnetHeader= for
 2582           tweaking Generic Segment Offload support.
 2584         * The address family for policy rules may be specified using the new
 2585           Family= option in the [RoutingPolicyRule] section.
 2587         * networkctl gained a new "delete" command for removing virtual network
 2588           devices, as well as a new "--stats" switch for showing device
 2589           statistics.
 2591         * networkd.conf gained a new setting SpeedMeter= and
 2592           SpeedMeterIntervalSec=, to measure bitrate of network interfaces. The
 2593           measured speed may be shown by 'networkctl status'.
 2595         * "networkctl status" now displays MTU and queue lengths, and more
 2596           detailed information about VXLAN and bridge devices.
 2598         * systemd-networkd's .network and .link files gained a new Property=
 2599           setting in the [Match] section, to match against devices with
 2600           specific udev properties.
 2602         * systemd-networkd's tunnel support gained a new option
 2603           AssignToLoopback= for selecting whether to use the loopback device
 2604           "lo" as underlying device.
 2606         * systemd-networkd's MACAddress= setting in the [Neighbor] section has
 2607           been renamed to LinkLayerAddress=, and it now allows configuration of
 2608           IP addresses, too.
 2610         * systemd-networkd's handling of the kernel's disable_ipv6 sysctl is
 2611           simplified: systemd-networkd will disable the sysctl (enable IPv6) if
 2612           IPv6 configuration (static or DHCPv6) was found for a given
 2613           interface. It will not touch the sysctl otherwise.
 2615         * The order of entries is $PATH used by the user manager instance was
 2616           changed to put bin/ entries before the corresponding sbin/ entries.
 2617           It is recommended to not rely on this order, and only ever have one
 2618           binary with a given name in the system paths under /usr.
 2620         * A new tool systemd-network-generator has been added that may generate
 2621           .network, .netdev and .link files from IP configuration specified on
 2622           the kernel command line in the format used by Dracut.
 2624         * The CriticalConnection= setting in .network files is now deprecated,
 2625           and replaced by a new KeepConfiguration= setting which allows more
 2626           detailed configuration of the IP configuration to keep in place.
 2628         * systemd-analyze gained a few new verbs:
 2630           - "systemd-analyze timestamp" parses and converts timestamps. This is
 2631             similar to the existing "systemd-analyze calendar" command which
 2632             does the same for recurring calendar events.
 2634           - "systemd-analyze timespan" parses and converts timespans (i.e.
 2635             durations as opposed to points in time).
 2637           - "systemd-analyze condition" will parse and test ConditionXYZ=
 2638             expressions.
 2640           - "systemd-analyze exit-status" will parse and convert exit status
 2641             codes to their names and back.
 2643           - "systemd-analyze unit-files" will print a list of all unit
 2644             file paths and unit aliases.
 2646         * SuccessExitStatus=, RestartPreventExitStatus=, and
 2647           RestartForceExitStatus= now accept exit status names (e.g. "DATAERR"
 2648           is equivalent to "65"). Those exit status name mappings may be
 2649           displayed with the systemd-analyze exit-status verb describe above.
 2651         * systemd-logind now exposes a per-session SetBrightness() bus call,
 2652           which may be used to securely change the brightness of a kernel
 2653           brightness device, if it belongs to the session's seat. By using this
 2654           call unprivileged clients can make changes to "backlight" and "leds"
 2655           devices securely with strict requirements on session membership.
 2656           Desktop environments may use this to generically make brightness
 2657           changes to such devices without shipping private SUID binaries or
 2658           udev rules for that purpose.
 2660         * "udevadm info" gained a --wait-for-initialization switch to wait for
 2661           a device to be initialized.
 2663         * systemd-hibernate-resume-generator will now look for resumeflags= on
 2664           the kernel command line, which is similar to rootflags= and may be
 2665           used to configure device timeout for the hibernation device.
 2667         * sd-event learnt a new API call sd_event_source_disable_unref() for
 2668           disabling and unref'ing an event source in a single function. A
 2669           related call sd_event_source_disable_unrefp() has been added for use
 2670           with gcc's cleanup extension.
 2672         * The sd-id128.h public API gained a new definition
 2673           SD_ID128_UUID_FORMAT_STR for formatting a 128bit ID in UUID format
 2674           with printf().
 2676         * "busctl introspect" gained a new switch --xml-interface for dumping
 2677           XML introspection data unmodified.
 2679         * PID 1 may now show the unit name instead of the unit description
 2680           string in its status output during boot. This may be configured in
 2681           the StatusUnitFormat= setting in /etc/systemd/system.conf or the
 2682           kernel command line option systemd.status_unit_format=.
 2684         * PID 1 now understands a new option KExecWatchdogSec= in
 2685           /etc/systemd/system.conf to set a watchdog timeout for kexec reboots.
 2686           Previously watchdog functionality was only available for regular
 2687           reboots. The new setting defaults to off, because we don't know in
 2688           the general case if the watchdog will be reset after kexec (some
 2689           drivers do reset it, but not all), and the new userspace might not be
 2690           configured to handle the watchdog.
 2692           Moreover, the old ShutdownWatchdogSec= setting has been renamed to
 2693           RebootWatchdogSec= to more clearly communicate what it is about. The
 2694           old name is still accepted for compatibility.
 2696         * The systemd.debug_shell kernel command line option now optionally
 2697           takes a tty name to spawn the debug shell on, which allows a
 2698           different tty to be selected than the built-in default.
 2700         * Service units gained a new ExecCondition= setting which will run
 2701           before ExecStartPre= and either continue execution of the unit (for
 2702           clean exit codes), stop execution without marking the unit failed
 2703           (for exit codes 1 through 254), or stop execution and fail the unit
 2704           (for exit code 255 or abnormal termination).
 2706         * A new service systemd-pstore.service has been added that pulls data
 2707           from /sys/fs/pstore/ and saves it to /var/lib/pstore for later
 2708           review.
 2710         * timedatectl gained new verbs for configuring per-interface NTP
 2711           service configuration for systemd-timesyncd.
 2713         * "localectl list-locales" won't list non-UTF-8 locales anymore. It's
 2714           2019. (You can set non-UTF-8 locales though, if you know their name.)
 2716         * If variable assignments in sysctl.d/ files are prefixed with "-" any
 2717           failures to apply them are now ignored.
 2719         * systemd-random-seed.service now optionally credits entropy when
 2720           applying the seed to the system. Set $SYSTEMD_RANDOM_SEED_CREDIT to
 2721           true for the service to enable this behaviour, but please consult the
 2722           documentation first, since this comes with a couple of caveats.
 2724         * systemd-random-seed.service is now a synchronization point for full
 2725           initialization of the kernel's entropy pool. Services that require
 2726           /dev/urandom to be correctly initialized should be ordered after this
 2727           service.
 2729         * The systemd-boot boot loader has been updated to optionally maintain
 2730           a random seed file in the EFI System Partition (ESP). During the boot
 2731           phase, this random seed is read and updated with a new seed
 2732           cryptographically derived from it. Another derived seed is passed to
 2733           the OS. The latter seed is then credited to the kernel's entropy pool
 2734           very early during userspace initialization (from PID 1). This allows
 2735           systems to boot up with a fully initialized kernel entropy pool from
 2736           earliest boot on, and thus entirely removes all entropy pool
 2737           initialization delays from systems using systemd-boot. Special care
 2738           is taken to ensure different seeds are derived on system images
 2739           replicated to multiple systems. "bootctl status" will show whether
 2740           a seed was received from the boot loader.
 2742         * bootctl gained two new verbs:
 2744           - "bootctl random-seed" will generate the file in ESP and an EFI
 2745             variable to allow a random seed to be passed to the OS as described
 2746             above.
 2748           - "bootctl is-installed" checks whether systemd-boot is currently
 2749             installed.
 2751         * bootctl will warn if it detects that boot entries are misconfigured
 2752           (for example if the kernel image was removed without purging the
 2753           bootloader entry).
 2755         * A new document has been added describing systemd's use and support
 2756           for the kernel's entropy pool subsystem:
 2758           https://systemd.io/RANDOM_SEEDS
 2760         * When the system is hibernated the swap device to write the
 2761           hibernation image to is now automatically picked from all available
 2762           swap devices, preferring the swap device with the highest configured
 2763           priority over all others, and picking the device with the most free
 2764           space if there are multiple devices with the highest priority.
 2766         * /etc/crypttab support has learnt a new keyfile-timeout= per-device
 2767           option that permits selecting the timeout how long to wait for a
 2768           device with an encryption key before asking for the password.
 2770         * IOWeight= has learnt to properly set the IO weight when using the
 2771           BFQ scheduler officially found in kernels 5.0+.
 2773         * A new mailing list has been created for reporting of security issues:
 2774           systemd-security@redhat.com. For mode details, see
 2775           https://systemd.io/CONTRIBUTING#security-vulnerability-reports.
 2777         Contributions from: Aaron Barany, Adrian Bunk, Alan Jenkins, Albrecht
 2778         Lohofener, Andrej Valek, Anita Zhang, Arian van Putten, Balint Reczey,
 2779         Bastien Nocera, Ben Boeckel, Benjamin Robin, camoz, Chen Qi, Chris
 2780         Chiu, Chris Down, Christian Göttsche, Christian Kellner, Clinton Roy,
 2781         Connor Reeder, Daniel Black, Daniel Lublin, Daniele Medri, Dan
 2782         Streetman, Dave Reisner, Dave Ross, David Art, David Tardon, Debarshi
 2783         Ray, Dimitri John Ledkov, Dominick Grift, Donald Buczek, Douglas
 2784         Christman, Eric DeVolder, EtherGraf, Evgeny Vereshchagin, Feldwor,
 2785         Felix Riemann, Florian Dollinger, Francesco Pennica, Franck Bui,
 2786         Frantisek Sumsal, Franz Pletz, frederik, Hans de Goede, Iago López
 2787         Galeiras, Insun Pyo, Ivan Shapovalov, Iwan Timmer, Jack, Jakob
 2788         Unterwurzacher, Jan Chren, Jan Klötzke, Jan Losinski, Jan Pokorný, Jan
 2789         Synacek, Jan-Michael Brummer, Jeka Pats, Jeremy Soller, Jérémy Rosen,
 2790         Jiri Pirko, Joe Lin, Joerg Behrmann, Joe Richey, Jóhann B. Guðmundsson,
 2791         Johannes Christ, Johannes Schmitz, Jonathan Rouleau, Jorge Niedbalski,
 2792         Jörg Thalheim, Kai Krakow, Kai Lüke, Karel Zak, Kashyap Chamarthy,
 2793         Krayushkin Konstantin, Lennart Poettering, Lubomir Rintel, Luca
 2794         Boccassi, Luís Ferreira, Marc-André Lureau, Markus Felten, Martin Pitt,
 2795         Matthew Leeds, Mattias Jernberg, Michael Biebl, Michael Olbrich,
 2796         Michael Prokop, Michael Stapelberg, Michael Zhivich, Michal Koutný,
 2797         Michal Sekletar, Mike Gilbert, Milan Broz, Miroslav Lichvar, mpe85,
 2798         Mr-Foo, Network Silence, Oliver Harley, pan93412, Paul Menzel, pEJipE,
 2799         Peter A. Bigot, Philip Withnall, Piotr Drąg, Rafael Fontenelle, Robert
 2800         Scheck, Roberto Santalla, Ronan Pigott, root, RussianNeuroMancer,
 2801         Sebastian Jennen, shinygold, Shreyas Behera, Simon Schricker, Susant
 2802         Sahani, Thadeu Lima de Souza Cascardo, Theo Ouzhinski, Thiebaud
 2803         Weksteen, Thomas Haller, Thomas Weißschuh, Tomas Mraz, Tommi Rantala,
 2804         Topi Miettinen, VD-Lycos, ven, Vladimir Yerilov, Wieland Hoffmann,
 2805         William A. Kennington III, William Wold, Xi Ruoyao, Yuri Chornoivan,
 2806         Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek, Zhang Xianwei
 2808         – Camerino, 2019-09-03
 2810 CHANGES WITH 242:
 2812         * In .link files, MACAddressPolicy=persistent (the default) is changed
 2813           to cover more devices. For devices like bridges, tun, tap, bond, and
 2814           similar interfaces that do not have other identifying information,
 2815           the interface name is used as the basis for persistent seed for MAC
 2816           and IPv4LL addresses. The way that devices that were handled
 2817           previously is not changed, and this change is about covering more
 2818           devices then previously by the "persistent" policy.
 2820           MACAddressPolicy=random may be used to force randomized MACs and
 2821           IPv4LL addresses for a device if desired.
 2823           Hint: the log output from udev (at debug level) was enhanced to
 2824           clarify what policy is followed and which attributes are used.
 2825           `SYSTEMD_LOG_LEVEL=debug udevadm test-builtin net_setup_link /sys/class/net/<name>`
 2826           may be used to view this.
 2828           Hint: if a bridge interface is created without any slaves, and gains
 2829           a slave later, then now the bridge does not inherit slave's MAC.
 2830           To inherit slave's MAC, for example, create the following file:
 2831           ```
 2832           # /etc/systemd/network/98-bridge-inherit-mac.link
 2833           [Match]
 2834           Type=bridge
 2836           [Link]
 2837           MACAddressPolicy=none
 2838           ```
 2840         * The .device units generated by systemd-fstab-generator and other
 2841           generators do not automatically pull in the corresponding .mount unit
 2842           as a Wants= dependency. This means that simply plugging in the device
 2843           will not cause the mount unit to be started automatically. But please
 2844           note that the mount unit may be started for other reasons, in
 2845           particular if it is part of local-fs.target, and any unit which
 2846           (transitively) depends on local-fs.target is started.
 2848         * networkctl list/status/lldp now accept globbing wildcards for network
 2849           interface names to match against all existing interfaces.
 2851         * The $PIDFILE environment variable is set to point the absolute path
 2852           configured with PIDFile= for processes of that service.
 2854         * The fallback DNS server list was augmented with Cloudflare public DNS
 2855           servers. Use `-Ddns-servers=` to set a different fallback.
 2857         * A new special target usb-gadget.target will be started automatically
 2858           when a USB Device Controller is detected (which means that the system
 2859           is a USB peripheral).
 2861         * A new unit setting CPUQuotaPeriodSec= assigns the time period
 2862           relatively to which the CPU time quota specified by CPUQuota= is
 2863           measured.
 2865         * A new unit setting ProtectHostname= may be used to prevent services
 2866           from modifying hostname information (even if they otherwise would
 2867           have privileges to do so).
 2869         * A new unit setting NetworkNamespacePath= may be used to specify a
 2870           namespace for service or socket units through a path referring to a
 2871           Linux network namespace pseudo-file.
 2873         * The PrivateNetwork= setting and JoinsNamespaceOf= dependencies now
 2874           have an effect on .socket units: when used the listening socket is
 2875           created within the configured network namespace instead of the host
 2876           namespace.
 2878         * ExecStart= command lines in unit files may now be prefixed with ':'
 2879           in which case environment variable substitution is
 2880           disabled. (Supported for the other ExecXYZ= settings, too.)
 2882         * .timer units gained two new boolean settings OnClockChange= and
 2883           OnTimezoneChange= which may be used to also trigger a unit when the
 2884           system clock is changed or the local timezone is
 2885           modified. systemd-run has been updated to make these options easily
 2886           accessible from the command line for transient timers.
 2888         * Two new conditions for units have been added: ConditionMemory= may be
 2889           used to conditionalize a unit based on installed system
 2890           RAM. ConditionCPUs= may be used to conditionalize a unit based on
 2891           installed CPU cores.
 2893         * The @default system call filter group understood by SystemCallFilter=
 2894           has been updated to include the new rseq() system call introduced in
 2895           kernel 4.15.
 2897         * A new time-set.target has been added that indicates that the system
 2898           time has been set from a local source (possibly imprecise). The
 2899           existing time-sync.target is stronger and indicates that the time has
 2900           been synchronized with a precise external source. Services where
 2901           approximate time is sufficient should use the new target.
 2903         * "systemctl start" (and related commands) learnt a new
 2904           --show-transaction option. If specified brief information about all
 2905           jobs queued because of the requested operation is shown.
 2907         * systemd-networkd recognizes a new operation state 'enslaved', used
 2908           (instead of 'degraded' or 'carrier') for interfaces which form a
 2909           bridge, bond, or similar, and an new 'degraded-carrier' operational
 2910           state used for the bond or bridge master interface when one of the
 2911           enslaved devices is not operational.
 2913         * .network files learnt the new IgnoreCarrierLoss= option for leaving
 2914           networks configured even if the carrier is lost.
 2916         * The RequiredForOnline= setting in .network files may now specify a
 2917           minimum operational state required for the interface to be considered
 2918           "online" by systemd-networkd-wait-online. Related to this
 2919           systemd-networkd-wait-online gained a new option --operational-state=
 2920           to configure the same, and its --interface= option was updated to
 2921           optionally also take an operational state specific for an interface.
 2923         * systemd-networkd-wait-online gained a new setting --any for waiting
 2924           for only one of the requested interfaces instead of all of them.
 2926         * systemd-networkd now implements L2TP tunnels.
 2928         * Two new .network settings UseAutonomousPrefix= and UseOnLinkPrefix=
 2929           may be used to cause autonomous and onlink prefixes received in IPv6
 2930           Router Advertisements to be ignored.
 2932         * New MulticastFlood=, NeighborSuppression=, and Learning= .network
 2933           file settings may be used to tweak bridge behaviour.
 2935         * The new TripleSampling= option in .network files may be used to
 2936           configure CAN triple sampling.
 2938         * A new .netdev settings PrivateKeyFile= and PresharedKeyFile= may be
 2939           used to point to private or preshared key for a WireGuard interface.
 2941         * /etc/crypttab now supports the same-cpu-crypt and
 2942           submit-from-crypt-cpus options to tweak encryption work scheduling
 2943           details.
 2945         * systemd-tmpfiles will now take a BSD file lock before operating on a
 2946           contents of directory. This may be used to temporarily exclude
 2947           directories from aging by taking the same lock (useful for example
 2948           when extracting a tarball into /tmp or /var/tmp as a privileged user,
 2949           which might create files with really old timestamps, which
 2950           nevertheless should not be deleted). For further details, see:
 2952           https://systemd.io/TEMPORARY_DIRECTORIES
 2954         * systemd-tmpfiles' h line type gained support for the
 2955           FS_PROJINHERIT_FL ('P') file attribute (introduced in kernel 4.5),
 2956           controlling project quota inheritance.
 2958         * sd-boot and bootctl now implement support for an Extended Boot Loader
 2959           (XBOOTLDR) partition, that is intended to be mounted to /boot, in
 2960           addition to the ESP partition mounted to /efi or /boot/efi.
 2961           Configuration file fragments, kernels, initrds and other EFI images
 2962           to boot will be loaded from both the ESP and XBOOTLDR partitions.
 2963           The XBOOTLDR partition was previously described by the Boot Loader
 2964           Specification, but implementation was missing in sd-boot. Support for
 2965           this concept allows using the sd-boot boot loader in more
 2966           conservative scenarios where the boot loader itself is placed in the
 2967           ESP but the kernels to boot (and their metadata) in a separate
 2968           partition.
 2970         * A system may now be booted with systemd.volatile=overlay on the
 2971           kernel command line, which causes the root file system to be set up
 2972           an overlayfs mount combining the root-only root directory with a
 2973           writable tmpfs. In this setup, the underlying root device is not
 2974           modified, and any changes are lost at reboot.
 2976         * Similar, systemd-nspawn can now boot containers with a volatile
 2977           overlayfs root with the new --volatile=overlay switch.
 2979         * systemd-nspawn can now consume OCI runtime bundles using a new
 2980           --oci-bundle= option. This implementation is fully usable, with most
 2981           features in the specification implemented, but since this a lot of
 2982           new code and functionality, this feature should most likely not
 2983           be used in production yet.
 2985         * systemd-nspawn now supports various options described by the OCI
 2986           runtime specification on the command-line and in .nspawn files:
 2987           --inaccessible=/Inaccessible= may be used to mask parts of the file
 2988           system tree, --console=/--pipe may be used to configure how standard
 2989           input, output, and error are set up.
 2991         * busctl learned the `emit` verb to generate D-Bus signals.
 2993         * systemd-analyze cat-config may be used to gather and display
 2994           configuration spread over multiple files, for example system and user
 2995           presets, tmpfiles.d, sysusers.d, udev rules, etc.
 2997         * systemd-analyze calendar now takes an optional new parameter
 2998           --iterations= which may be used to show a maximum number of iterations
 2999           the specified expression will elapse next.
 3001         * The sd-bus C API gained support for naming method parameters in the
 3002           introspection data.
 3004         * systemd-logind gained D-Bus APIs to specify the "reboot parameter"
 3005           the reboot() system call expects.
 3007         * journalctl learnt a new --cursor-file= option that points to a file
 3008           from which a cursor should be loaded in the beginning and to which
 3009           the updated cursor should be stored at the end.
 3011         * ACRN hypervisor and Windows Subsystem for Linux (WSL) are now
 3012           detected by systemd-detect-virt (and may also be used in
 3013           ConditionVirtualization=).
 3015         * The behaviour of systemd-logind may now be modified with environment
 3016           variables $SYSTEMD_REBOOT_TO_FIRMWARE_SETUP,
 3018           $SYSTEMD_REBOOT_TO_BOOT_LOADER_ENTRY. They cause logind to either
 3019           skip the relevant operation completely (when set to false), or to
 3020           create a flag file in /run/systemd (when set to true), instead of
 3021           actually commencing the real operation when requested. The presence
 3022           of /run/systemd/reboot-to-firmware-setup,
 3023           /run/systemd/reboot-to-boot-loader-menu, and
 3024           /run/systemd/reboot-to-boot-loader-entry, may be used by alternative
 3025           boot loader implementations to replace some steps logind performs
 3026           during reboot with their own operations.
 3028         * systemctl can be used to request a reboot into the boot loader menu
 3029           or a specific boot loader entry with the new --boot-load-menu= and
 3030           --boot-loader-entry= options to a reboot command. (This requires a
 3031           boot loader that supports this, for example sd-boot.)
 3033         * kernel-install will no longer unconditionally create the output
 3034           directory (e.g. /efi/<machine-id>/<kernel-version>) for boot loader
 3035           snippets, but will do only if the machine-specific parent directory
 3036           (i.e. /efi/<machine-id>/) already exists. bootctl has been modified
 3037           to create this parent directory during sd-boot installation.
 3039           This makes it easier to use kernel-install with plugins which support
 3040           a different layout of the bootloader partitions (for example grub2).
 3042         * During package installation (with `ninja install`), we would create
 3043           symlinks for getty@tty1.service, systemd-networkd.service,
 3044           systemd-networkd.socket, systemd-resolved.service,
 3045           remote-cryptsetup.target, remote-fs.target,
 3046           systemd-networkd-wait-online.service, and systemd-timesyncd.service
 3047           in /etc, as if `systemctl enable` was called for those units, to make
 3048           the system usable immediately after installation. Now this is not
 3049           done anymore, and instead calling `systemctl preset-all` is
 3050           recommended after the first installation of systemd.
 3052         * A new boolean sandboxing option RestrictSUIDSGID= has been added that
 3053           is built on seccomp. When turned on creation of SUID/SGID files is
 3054           prohibited.
 3056         * The NoNewPrivileges= and the new RestrictSUIDSGID= options are now
 3057           implied if DynamicUser= is turned on for a service. This hardens
 3058           these services, so that they neither can benefit from nor create
 3059           SUID/SGID executables. This is a minor compatibility breakage, given
 3060           that when DynamicUser= was first introduced SUID/SGID behaviour was
 3061           unaffected. However, the security benefit of these two options is
 3062           substantial, and the setting is still relatively new, hence we opted
 3063           to make it mandatory for services with dynamic users.
 3065         Contributions from: Adam Jackson, Alexander Tsoy, Andrey Yashkin,
 3066         Andrzej Pietrasiewicz, Anita Zhang, Balint Reczey, Beniamino Galvani,
 3067         Ben Iofel, Benjamin Berg, Benjamin Dahlhoff, Chris, Chris Morin,
 3068         Christopher Wong, Claudius Ellsel, Clemens Gruber, dana, Daniel Black,
 3069         Davide Cavalca, David Michael, David Rheinsberg, emersion, Evgeny
 3070         Vereshchagin, Filipe Brandenburger, Franck Bui, Frantisek Sumsal,
 3071         Giacinto Cifelli, Hans de Goede, Hugo Kindel, Ignat Korchagin, Insun
 3072         Pyo, Jan Engelhardt, Jonas Dorel, Jonathan Lebon, Jonathon Kowalski,
 3073         Jörg Sommer, Jörg Thalheim, Jussi Pakkanen, Kai-Heng Feng, Lennart
 3074         Poettering, Lubomir Rintel, Luís Ferreira, Martin Pitt, Matthias
 3075         Klumpp, Michael Biebl, Michael Niewöhner, Michael Olbrich, Michal
 3076         Sekletar, Mike Lothian, Paul Menzel, Piotr Drąg, Riccardo Schirone,
 3077         Robin Elvedi, Roman Kulikov, Ronald Tschalär, Ross Burton, Ryan
 3078         Gonzalez, Sebastian Krzyszkowiak, Stephane Chazelas, StKob, Susant
 3079         Sahani, Sylvain Plantefève, Szabolcs Fruhwald, Taro Yamada, Theo
 3080         Ouzhinski, Thomas Haller, Tobias Jungel, Tom Yan, Tony Asleson, Topi
 3081         Miettinen, unixsysadmin, Van Laser, Vesa Jääskeläinen, Yu, Li-Yu,
 3082         Yu Watanabe, Zbigniew Jędrzejewski-Szmek
 3084         — Warsaw, 2019-04-11
 3086 CHANGES WITH 241:
 3088         * The default locale can now be configured at compile time. Otherwise,
 3089           a suitable default will be selected automatically (one of C.UTF-8,
 3090           en_US.UTF-8, and C).
 3092         * The version string shown by systemd and other tools now includes the
 3093           git commit hash when built from git. An override may be specified
 3094           during compilation, which is intended to be used by distributions to
 3095           include the package release information.
 3097         * systemd-cat can now filter standard input and standard error streams
 3098           for different syslog priorities using the new --stderr-priority=
 3099           option.
 3101         * systemd-journald and systemd-journal-remote reject entries which
 3102           contain too many fields (CVE-2018-16865) and set limits on the
 3103           process' command line length (CVE-2018-16864).
 3105         * $DBUS_SESSION_BUS_ADDRESS environment variable is set by pam_systemd
 3106           again.
 3108         * A new network device NamePolicy "keep" is implemented for link files,
 3109           and used by default in 99-default.link (the fallback configuration
 3110           provided by systemd). With this policy, if the network device name
 3111           was already set by userspace, the device will not be renamed again.
 3112           This matches the naming scheme that was implemented before
 3113           systemd-240. If naming-scheme < 240 is specified, the "keep" policy
 3114           is also enabled by default, even if not specified. Effectively, this
 3115           means that if naming-scheme >= 240 is specified, network devices will
 3116           be renamed according to the configuration, even if they have been
 3117           renamed already, if "keep" is not specified as the naming policy in
 3118           the .link file. The 99-default.link file provided by systemd includes
 3119           "keep" for backwards compatibility, but it is recommended for user
 3120           installed .link files to *not* include it.
 3122           The "kernel" policy, which keeps kernel names declared to be
 3123           "persistent", now works again as documented.
 3125         * kernel-install script now optionally takes the paths to one or more
 3126           initrd files, and passes them to all plugins.
 3128         * The mincore() system call has been dropped from the @system-service
 3129           system call filter group, as it is pretty exotic and may potentially
 3130           used for side-channel attacks.
 3132         * -fPIE is dropped from compiler and linker options. Please specify
 3133           -Db_pie=true option to meson to build position-independent
 3134           executables. Note that the meson option is supported since meson-0.49.
 3136         * The fs.protected_regular and fs.protected_fifos sysctls, which were
 3137           added in Linux 4.19 to make some data spoofing attacks harder, are
 3138           now enabled by default. While this will hopefully improve the
 3139           security of most installations, it is technically a backwards
 3140           incompatible change; to disable these sysctls again, place the
 3141           following lines in /etc/sysctl.d/60-protected.conf or a similar file:
 3143               fs.protected_regular = 0
 3144               fs.protected_fifos = 0
 3146           Note that the similar hardlink and symlink protection has been
 3147           enabled since v199, and may be disabled likewise.
 3149         * The files read from the EnvironmentFile= setting in unit files now
 3150           parse backslashes inside quotes literally, matching the behaviour of
 3151           POSIX shells.
 3153         * udevadm trigger, udevadm control, udevadm settle and udevadm monitor
 3154           now automatically become NOPs when run in a chroot() environment.
 3156         * The tmpfiles.d/ "C" line type will now copy directory trees not only
 3157           when the destination is so far missing, but also if it already exists
 3158           as a directory and is empty. This is useful to cater for systems
 3159           where directory trees are put together from multiple separate mount
 3160           points but otherwise empty.
 3162         * A new function sd_bus_close_unref() (and the associated
 3163           sd_bus_close_unrefp()) has been added to libsystemd, that combines
 3164           sd_bus_close() and sd_bus_unref() in one.
 3166         * udevadm control learnt a new option for --ping for testing whether a
 3167           systemd-udevd instance is running and reacting.
 3169         * udevadm trigger learnt a new option for --wait-daemon for waiting
 3170           systemd-udevd daemon to be initialized.
 3172         Contributions from: Aaron Plattner, Alberts Muktupāvels, Alex Mayer,
 3173         Ayman Bagabas, Beniamino Galvani, Burt P, Chris Down, Chris Lamb, Chris
 3174         Morin, Christian Hesse, Claudius Ellsel, dana, Daniel Axtens, Daniele
 3175         Medri, Dave Reisner, David Santamaría Rogado, Diego Canuhe, Dimitri
 3176         John Ledkov, Evgeny Vereshchagin, Fabrice Fontaine, Filipe
 3177         Brandenburger, Franck Bui, Frantisek Sumsal, govwin, Hans de Goede,
 3178         James Hilliard, Jan Engelhardt, Jani Uusitalo, Jan Janssen, Jan
 3179         Synacek, Jonathan McDowell, Jonathan Roemer, Jonathon Kowalski, Joost
 3180         Heitbrink, Jörg Thalheim, Lance, Lennart Poettering, Louis Taylor,
 3181         Lucas Werkmeister, Mantas Mikulėnas, Marc-Antoine Perennou,
 3182         marvelousblack, Michael Biebl, Michael Sloan, Michal Sekletar, Mike
 3183         Auty, Mike Gilbert, Mikhail Kasimov, Neil Brown, Niklas Hambüchen,
 3184         Patrick Williams, Paul Seyfert, Peter Hutterer, Philip Withnall, Roger
 3185         James, Ronnie P. Thomas, Ryan Gonzalez, Sam Morris, Stephan Edel,
 3186         Stephan Gerhold, Susant Sahani, Taro Yamada, Thomas Haller, Topi
 3187         Miettinen, YiFei Zhu, YmrDtnJu, YunQiang Su, Yu Watanabe, Zbigniew
 3188         Jędrzejewski-Szmek, zsergeant77, Дамјан Георгиевски
 3190         — Berlin, 2019-02-14
 3192 CHANGES WITH 240:
 3194         * NoNewPrivileges=yes has been set for all long-running services
 3195           implemented by systemd. Previously, this was problematic due to
 3196           SELinux (as this would also prohibit the transition from PID1's label
 3197           to the service's label). This restriction has since been lifted, but
 3198           an SELinux policy update is required.
 3199           (See e.g. https://github.com/fedora-selinux/selinux-policy/pull/234.)
 3201         * DynamicUser=yes is dropped from systemd-networkd.service,
 3202           systemd-resolved.service and systemd-timesyncd.service, which was
 3203           enabled in v239 for systemd-networkd.service and systemd-resolved.service,
 3204           and since v236 for systemd-timesyncd.service. The users and groups
 3205           systemd-network, systemd-resolve and systemd-timesync are created
 3206           by systemd-sysusers again. Distributors or system administrators
 3207           may need to create these users and groups if they not exist (or need
 3208           to re-enable DynamicUser= for those units) while upgrading systemd.
 3209           Also, the clock file for systemd-timesyncd may need to move from
 3210           /var/lib/private/systemd/timesync/clock to /var/lib/systemd/timesync/clock.
 3212         * When unit files are loaded from disk, previously systemd would
 3213           sometimes (depending on the unit loading order) load units from the
 3214           target path of symlinks in .wants/ or .requires/ directories of other
 3215           units. This meant that unit could be loaded from different paths
 3216           depending on whether the unit was requested explicitly or as a
 3217           dependency of another unit, not honouring the priority of directories
 3218           in search path. It also meant that it was possible to successfully
 3219           load and start units which are not found in the unit search path, as
 3220           long as they were requested as a dependency and linked to from
 3221           .wants/ or .requires/. The target paths of those symlinks are not
 3222           used for loading units anymore and the unit file must be found in
 3223           the search path.
 3225         * A new service type has been added: Type=exec. It's very similar to
 3226           Type=simple but ensures the service manager will wait for both fork()
 3227           and execve() of the main service binary to complete before proceeding
 3228           with follow-up units. This is primarily useful so that the manager
 3229           propagates any errors in the preparation phase of service execution
 3230           back to the job that requested the unit to be started. For example,
 3231           consider a service that has ExecStart= set to a file system binary
 3232           that doesn't exist. With Type=simple starting the unit would be
 3233           considered instantly successful, as only fork() has to complete
 3234           successfully and the manager does not wait for execve(), and hence
 3235           its failure is seen "too late". With the new Type=exec service type
 3236           starting the unit will fail, as the manager will wait for the
 3237           execve() and notice its failure, which is then propagated back to the
 3238           start job.
 3240           NOTE: with the next release 241 of systemd we intend to change the
 3241           systemd-run tool to default to Type=exec for transient services
 3242           started by it. This should be mostly safe, but in specific corner
 3243           cases might result in problems, as the systemd-run tool will then
 3244           block on NSS calls (such as user name look-ups due to User=) done
 3245           between the fork() and execve(), which under specific circumstances
 3246           might cause problems. It is recommended to specify "-p Type=simple"
 3247           explicitly in the few cases where this applies. For regular,
 3248           non-transient services (i.e. those defined with unit files on disk)
 3249           we will continue to default to Type=simple.
 3251         * The Linux kernel's current default RLIMIT_NOFILE resource limit for
 3252           userspace processes is set to 1024 (soft) and 4096
 3253           (hard). Previously, systemd passed this on unmodified to all
 3254           processes it forked off. With this systemd release the hard limit
 3255           systemd passes on is increased to 512K, overriding the kernel's
 3256           defaults and substantially increasing the number of simultaneous file
 3257           descriptors unprivileged userspace processes can allocate. Note that
 3258           the soft limit remains at 1024 for compatibility reasons: the
 3259           traditional UNIX select() call cannot deal with file descriptors >=
 3260           1024 and increasing the soft limit globally might thus result in
 3261           programs unexpectedly allocating a high file descriptor and thus
 3262           failing abnormally when attempting to use it with select() (of
 3263           course, programs shouldn't use select() anymore, and prefer
 3264           poll()/epoll, but the call unfortunately remains undeservedly popular
 3265           at this time). This change reflects the fact that file descriptor
 3266           handling in the Linux kernel has been optimized in more recent
 3267           kernels and allocating large numbers of them should be much cheaper
 3268           both in memory and in performance than it used to be. Programs that
 3269           want to take benefit of the increased limit have to "opt-in" into
 3270           high file descriptors explicitly by raising their soft limit. Of
 3271           course, when they do that they must acknowledge that they cannot use
 3272           select() anymore (and neither can any shared library they use — or
 3273           any shared library used by any shared library they use and so on).
 3274           Which default hard limit is most appropriate is of course hard to
 3275           decide. However, given reports that ~300K file descriptors are used
 3276           in real-life applications we believe 512K is sufficiently high as new
 3277           default for now. Note that there are also reports that using very
 3278           high hard limits (e.g. 1G) is problematic: some software allocates
 3279           large arrays with one element for each potential file descriptor
 3280           (Java, …) — a high hard limit thus triggers excessively large memory
 3281           allocations in these applications. Hopefully, the new default of 512K
 3282           is a good middle ground: higher than what real-life applications
 3283           currently need, and low enough for avoid triggering excessively large
 3284           allocations in problematic software. (And yes, somebody should fix
 3285           Java.)
 3287         * The fs.nr_open and fs.file-max sysctls are now automatically bumped
 3288           to the highest possible values, as separate accounting of file
 3289           descriptors is no longer necessary, as memcg tracks them correctly as
 3290           part of the memory accounting anyway. Thus, from the four limits on
 3291           file descriptors currently enforced (fs.file-max, fs.nr_open,
 3292           RLIMIT_NOFILE hard, RLIMIT_NOFILE soft) we turn off the first two,
 3293           and keep only the latter two. A set of build-time options
 3294           (-Dbump-proc-sys-fs-file-max=false and -Dbump-proc-sys-fs-nr-open=false)
 3295           has been added to revert this change in behaviour, which might be
 3296           an option for systems that turn off memcg in the kernel.
 3298         * When no /etc/locale.conf file exists (and hence no locale settings
 3299           are in place), systemd will now use the "C.UTF-8" locale by default,
 3300           and set LANG= to it. This locale is supported by various
 3301           distributions including Fedora, with clear indications that upstream
 3302           glibc is going to make it available too. This locale enables UTF-8
 3303           mode by default, which appears appropriate for 2018.
 3305         * The "net.ipv4.conf.all.rp_filter" sysctl will now be set to 2 by
 3306           default. This effectively switches the RFC3704 Reverse Path filtering
 3307           from Strict mode to Loose mode. This is more appropriate for hosts
 3308           that have multiple links with routes to the same networks (e.g.
 3309           a client with a Wi-Fi and Ethernet both connected to the internet).
 3311           Consult the kernel documentation for details on this sysctl:
 3312           https://www.kernel.org/doc/Documentation/networking/ip-sysctl.txt
 3314         * CPUAccounting=yes no longer enables the CPU controller when using
 3315           kernel 4.15+ and the unified cgroup hierarchy, as required accounting
 3316           statistics are now provided independently from the CPU controller.
 3318         * Support for disabling a particular cgroup controller within a sub-tree
 3319           has been added through the DisableControllers= directive.
 3321         * cgroup_no_v1=all on the kernel command line now also implies
 3322           using the unified cgroup hierarchy, unless one explicitly passes
 3323           systemd.unified_cgroup_hierarchy=0 on the kernel command line.
 3325         * The new "MemoryMin=" unit file property may now be used to set the
 3326           memory usage protection limit of processes invoked by the unit. This
 3327           controls the cgroup v2 memory.min attribute. Similarly, the new
 3328           "IODeviceLatencyTargetSec=" property has been added, wrapping the new
 3329           cgroup v2 io.latency cgroup property for configuring per-service I/O
 3330           latency.
 3332         * systemd now supports the cgroup v2 devices BPF logic, as counterpart
 3333           to the cgroup v1 "devices" cgroup controller.
 3335         * systemd-escape now is able to combine --unescape with --template. It
 3336           also learnt a new option --instance for extracting and unescaping the
 3337           instance part of a unit name.
 3339         * sd-bus now provides the sd_bus_message_readv() which is similar to
 3340           sd_bus_message_read() but takes a va_list object. The pair
 3341           sd_bus_set_method_call_timeout() and sd_bus_get_method_call_timeout()
 3342           has been added for configuring the default method call timeout to
 3343           use. sd_bus_error_move() may be used to efficiently move the contents
 3344           from one sd_bus_error structure to another, invalidating the
 3345           source. sd_bus_set_close_on_exit() and sd_bus_get_close_on_exit() may
 3346           be used to control whether a bus connection object is automatically
 3347           flushed when an sd-event loop is exited.
 3349         * When processing classic BSD syslog log messages, journald will now
 3350           save the original time-stamp string supplied in the new
 3351           SYSLOG_TIMESTAMP= journal field. This permits consumers to
 3352           reconstruct the original BSD syslog message more correctly.
 3354         * StandardOutput=/StandardError= in service files gained support for
 3355           new "append:…" parameters, for connecting STDOUT/STDERR of a service
 3356           to a file, and appending to it.
 3358         * The signal to use as last step of killing of unit processes is now
 3359           configurable. Previously it was hard-coded to SIGKILL, which may now
 3360           be overridden with the new KillSignal= setting. Note that this is the
 3361           signal used when regular termination (i.e. SIGTERM) does not suffice.
 3362           Similarly, the signal used when aborting a program in case of a
 3363           watchdog timeout may now be configured too (WatchdogSignal=).
 3365         * The XDG_SESSION_DESKTOP environment variable may now be configured in
 3366           the pam_systemd argument line, using the new desktop= switch. This is
 3367           useful to initialize it properly from a display manager without
 3368           having to touch C code.
 3370         * Most configuration options that previously accepted percentage values
 3371           now also accept permille values with the '‰' suffix (instead of '%').
 3373         * systemd-resolved may now optionally use OpenSSL instead of GnuTLS for
 3374           DNS-over-TLS.
 3376         * systemd-resolved's configuration file resolved.conf gained a new
 3377           option ReadEtcHosts= which may be used to turn off processing and
 3378           honoring /etc/hosts entries.
 3380         * The "--wait" switch may now be passed to "systemctl
 3381           is-system-running", in which case the tool will synchronously wait
 3382           until the system finished start-up.
 3384         * hostnamed gained a new bus call to determine the DMI product UUID.
 3386         * On x86-64 systemd will now prefer using the RDRAND processor
 3387           instruction over /dev/urandom whenever it requires randomness that
 3388           neither has to be crypto-grade nor should be reproducible. This
 3389           should substantially reduce the amount of entropy systemd requests
 3390           from the kernel during initialization on such systems, though not
 3391           reduce it to zero. (Why not zero? systemd still needs to allocate
 3392           UUIDs and such uniquely, which require high-quality randomness.)
 3394         * networkd gained support for Foo-Over-UDP, ERSPAN and ISATAP
 3395           tunnels. It also gained a new option ForceDHCPv6PDOtherInformation=
 3396           for forcing the "Other Information" bit in IPv6 RA messages. The
 3397           bonding logic gained four new options AdActorSystemPriority=,
 3398           AdUserPortKey=, AdActorSystem= for configuring various 802.3ad
 3399           aspects, and DynamicTransmitLoadBalancing= for enabling dynamic
 3400           shuffling of flows. The tunnel logic gained a new
 3401           IPv6RapidDeploymentPrefix= option for configuring IPv6 Rapid
 3402           Deployment. The policy rule logic gained four new options IPProtocol=,
 3403           SourcePort= and DestinationPort=, InvertRule=. The bridge logic gained
 3404           support for the MulticastToUnicast= option. networkd also gained
 3405           support for configuring static IPv4 ARP or IPv6 neighbor entries.
 3407         * .preset files (as read by 'systemctl preset') may now be used to
 3408           instantiate services.
 3410         * /etc/crypttab now understands the sector-size= option to configure
 3411           the sector size for an encrypted partition.
 3413         * Key material for encrypted disks may now be placed on a formatted
 3414           medium, and referenced from /etc/crypttab by the UUID of the file
 3415           system, followed by "=" suffixed by the path to the key file.
 3417         * The "collect" udev component has been removed without replacement, as
 3418           it is neither used nor maintained.
 3420         * When the RuntimeDirectory=, StateDirectory=, CacheDirectory=,
 3421           LogsDirectory=, ConfigurationDirectory= settings are used in a
 3422           service the executed processes will now receive a set of environment
 3423           variables containing the full paths of these directories.
 3425           LOGS_DIRECTORY, CONFIGURATION_DIRECTORY are now set if these options
 3426           are used. Note that these options may be used multiple times per
 3427           service in which case the resulting paths will be concatenated and
 3428           separated by colons.
 3430         * Predictable interface naming has been extended to cover InfiniBand
 3431           NICs. They will be exposed with an "ib" prefix.
 3433         * tmpfiles.d/ line types may now be suffixed with a '-' character, in
 3434           which case the respective line failing is ignored.
 3436         * .link files may now be used to configure the equivalent to the
 3437           "ethtool advertise" commands.
 3439         * The sd-device.h and sd-hwdb.h APIs are now exported, as an
 3440           alternative to libudev.h. Previously, the latter was just an internal
 3441           wrapper around the former, but now these two APIs are exposed
 3442           directly.
 3444         * sd-id128.h gained a new function sd_id128_get_boot_app_specific()
 3445           which calculates an app-specific boot ID similar to how
 3446           sd_id128_get_machine_app_specific() generates an app-specific machine
 3447           ID.
 3449         * A new tool systemd-id128 has been added that can be used to determine
 3450           and generate various 128bit IDs.
 3452         * /etc/os-release gained two new standardized fields DOCUMENTATION_URL=
 3453           and LOGO=.
 3455         * systemd-hibernate-resume-generator will now honor the "noresume"
 3456           kernel command line option, in which case it will bypass resuming
 3457           from any hibernated image.
 3459         * The systemd-sleep.conf configuration file gained new options
 3460           AllowSuspend=, AllowHibernation=, AllowSuspendThenHibernate=,
 3461           AllowHybridSleep= for prohibiting specific sleep modes even if the
 3462           kernel exports them.
 3464         * portablectl is now officially supported and has thus moved to
 3465           /usr/bin/.
 3467         * bootctl learnt the two new commands "set-default" and "set-oneshot"
 3468           for setting the default boot loader item to boot to (either
 3469           persistently or only for the next boot). This is currently only
 3470           compatible with sd-boot, but may be implemented on other boot loaders
 3471           too, that follow the boot loader interface. The updated interface is
 3472           now documented here:
 3474           https://systemd.io/BOOT_LOADER_INTERFACE
 3476         * A new kernel command line option systemd.early_core_pattern= is now
 3477           understood which may be used to influence the core_pattern PID 1
 3478           installs during early boot.
 3480         * busctl learnt two new options -j and --json= for outputting method
 3481           call replies, properties and monitoring output in JSON.
 3483         * journalctl's JSON output now supports simple ANSI coloring as well as
 3484           a new "json-seq" mode for generating RFC7464 output.
 3486         * Unit files now support the %g/%G specifiers that resolve to the UNIX
 3487           group/GID of the service manager runs as, similar to the existing
 3488           %u/%U specifiers that resolve to the UNIX user/UID.
 3490         * systemd-logind learnt a new global configuration option
 3491           UserStopDelaySec= that may be set in logind.conf. It specifies how
 3492           long the systemd --user instance shall remain started after a user
 3493           logs out. This is useful to speed up repetitive re-connections of the
 3494           same user, as it means the user's service manager doesn't have to be
 3495           stopped/restarted on each iteration, but can be reused between
 3496           subsequent options. This setting defaults to 10s. systemd-logind also
 3497           exports two new properties on its Manager D-Bus objects indicating
 3498           whether the system's lid is currently closed, and whether the system
 3499           is on AC power.
 3501         * systemd gained support for a generic boot counting logic, which
 3502           generically permits automatic reverting to older boot loader entries
 3503           if newer updated ones don't work. The boot loader side is implemented
 3504           in sd-boot, but is kept open for other boot loaders too. For details
 3505           see:
 3507           https://systemd.io/AUTOMATIC_BOOT_ASSESSMENT
 3509         * The SuccessAction=/FailureAction= unit file settings now learnt two
 3510           new parameters: "exit" and "exit-force", which result in immediate
 3511           exiting of the service manager, and are only useful in systemd --user
 3512           and container environments.
 3514         * Unit files gained support for a pair of options
 3515           FailureActionExitStatus=/SuccessActionExitStatus= for configuring the
 3516           exit status to use as service manager exit status when
 3517           SuccessAction=/FailureAction= is set to exit or exit-force.
 3519         * A pair of LogRateLimitIntervalSec=/LogRateLimitBurst= per-service
 3520           options may now be used to configure the log rate limiting applied by
 3521           journald per-service.
 3523         * systemd-analyze gained a new verb "timespan" for parsing and
 3524           normalizing time span values (i.e. strings like "5min 7s 8us").
 3526         * systemd-analyze also gained a new verb "security" for analyzing the
 3527           security and sand-boxing settings of services in order to determine an
 3528           "exposure level" for them, indicating whether a service would benefit
 3529           from more sand-boxing options turned on for them.
 3531         * "systemd-analyze syscall-filter" will now also show system calls
 3532           supported by the local kernel but not included in any of the defined
 3533           groups.
 3535         * .nspawn files now understand the Ephemeral= setting, matching the
 3536           --ephemeral command line switch.
 3538         * sd-event gained the new APIs sd_event_source_get_floating() and
 3539           sd_event_source_set_floating() for controlling whether a specific
 3540           event source is "floating", i.e. destroyed along with the even loop
 3541           object itself.
 3543         * Unit objects on D-Bus gained a new "Refs" property that lists all
 3544           clients that currently have a reference on the unit (to ensure it is
 3545           not unloaded).
 3547         * The JoinControllers= option in system.conf is no longer supported, as
 3548           it didn't work correctly, is hard to support properly, is legacy (as
 3549           the concept only exists on cgroup v1) and apparently wasn't used.
 3551         * Journal messages that are generated whenever a unit enters the failed
 3552           state are now tagged with a unique MESSAGE_ID. Similarly, messages
 3553           generated whenever a service process exits are now made recognizable,
 3554           too. A tagged message is also emitted whenever a unit enters the
 3555           "dead" state on success.
 3557         * systemd-run gained a new switch --working-directory= for configuring
 3558           the working directory of the service to start. A shortcut -d is
 3559           equivalent, setting the working directory of the service to the
 3560           current working directory of the invoking program. The new --shell
 3561           (or just -S) option has been added for invoking the $SHELL of the
 3562           caller as a service, and implies --pty --same-dir --wait --collect
 3563           --service-type=exec. Or in other words, "systemd-run -S" is now the
 3564           quickest way to quickly get an interactive in a fully clean and
 3565           well-defined system service context.
 3567         * machinectl gained a new verb "import-fs" for importing an OS tree
 3568           from a directory. Moreover, when a directory or tarball is imported
 3569           and single top-level directory found with the OS itself below the OS
 3570           tree is automatically mangled and moved one level up.
 3572         * systemd-importd will no longer set up an implicit btrfs loop-back
 3573           file system on /var/lib/machines. If one is already set up, it will
 3574           continue to be used.
 3576         * A new generator "systemd-run-generator" has been added. It will
 3577           synthesize a unit from one or more program command lines included in
 3578           the kernel command line. This is very useful in container managers
 3579           for example:
 3581           # systemd-nspawn -i someimage.raw -b systemd.run='"some command line"'
 3583           This will run "systemd-nspawn" on an image, invoke the specified
 3584           command line and immediately shut down the container again, returning
 3585           the command line's exit code.
 3587         * The block device locking logic is now documented:
 3589           https://systemd.io/BLOCK_DEVICE_LOCKING
 3591         * loginctl and machinectl now optionally output the various tables in
 3592           JSON using the --output= switch. It is our intention to add similar
 3593           support to systemctl and all other commands.
 3595         * udevadm's query and trigger verb now optionally take a .device unit
 3596           name as argument.
 3598         * systemd-udevd's network naming logic now understands a new
 3599           net.naming-scheme= kernel command line switch, which may be used to
 3600           pick a specific version of the naming scheme. This helps stabilizing
 3601           interface names even as systemd/udev are updated and the naming logic
 3602           is improved.
 3604         * sd-id128.h learnt two new auxiliary helpers: sd_id128_is_allf() and
 3605           SD_ID128_ALLF to test if a 128bit ID is set to all 0xFF bytes, and to
 3606           initialize one to all 0xFF.
 3608         * After loading the SELinux policy systemd will now recursively relabel
 3609           all files and directories listed in
 3610           /run/systemd/relabel-extra.d/*.relabel (which should be simple
 3611           newline separated lists of paths) in addition to the ones it already
 3612           implicitly relabels in /run, /dev and /sys. After the relabelling is
 3613           completed the *.relabel files (and /run/systemd/relabel-extra.d/) are
 3614           removed. This is useful to permit initrds (i.e. code running before
 3615           the SELinux policy is in effect) to generate files in the host
 3616           filesystem safely and ensure that the correct label is applied during
 3617           the transition to the host OS.
 3619         * KERNEL API BREAKAGE: Linux kernel 4.18 changed behaviour regarding
 3620           mknod() handling in user namespaces. Previously mknod() would always
 3621           fail with EPERM in user namespaces. Since 4.18 mknod() will succeed
 3622           but device nodes generated that way cannot be opened, and attempts to
 3623           open them result in EPERM. This breaks the "graceful fallback" logic
 3624           in systemd's PrivateDevices= sand-boxing option. This option is
 3625           implemented defensively, so that when systemd detects it runs in a
 3626           restricted environment (such as a user namespace, or an environment
 3627           where mknod() is blocked through seccomp or absence of CAP_SYS_MKNOD)
 3628           where device nodes cannot be created the effect of PrivateDevices= is
 3629           bypassed (following the logic that 2nd-level sand-boxing is not
 3630           essential if the system systemd runs in is itself already sand-boxed
 3631           as a whole). This logic breaks with 4.18 in container managers where
 3632           user namespacing is used: suddenly PrivateDevices= succeeds setting
 3633           up a private /dev/ file system containing devices nodes — but when
 3634           these are opened they don't work.
 3636           At this point it is recommended that container managers utilizing
 3637           user namespaces that intend to run systemd in the payload explicitly
 3638           block mknod() with seccomp or similar, so that the graceful fallback
 3639           logic works again.
 3641           We are very sorry for the breakage and the requirement to change
 3642           container configurations for newer kernels. It's purely caused by an
 3643           incompatible kernel change. The relevant kernel developers have been
 3644           notified about this userspace breakage quickly, but they chose to
 3645           ignore it.
 3647         * PermissionsStartOnly= setting is deprecated (but is still supported
 3648           for backwards compatibility). The same functionality is provided by
 3649           the more flexible "+", "!", and "!!" prefixes to ExecStart= and other
 3650           commands.
 3652         * $DBUS_SESSION_BUS_ADDRESS environment variable is not set by
 3653           pam_systemd anymore.
 3655         * The naming scheme for network devices was changed to always rename
 3656           devices, even if they were already renamed by userspace. The "kernel"
 3657           policy was changed to only apply as a fallback, if no other naming
 3658           policy took effect.
 3660         * The requirements to build systemd is bumped to meson-0.46 and
 3661           python-3.5.
 3663         Contributions from: afg, Alan Jenkins, Aleksei Timofeyev, Alexander
 3664         Filippov, Alexander Kurtz, Alexey Bogdanenko, Andreas Henriksson,
 3665         Andrew Jorgensen, Anita Zhang, apnix-uk, Arkan49, Arseny Maslennikov,
 3666         asavah, Asbjørn Apeland, aszlig, Bastien Nocera, Ben Boeckel, Benedikt
 3667         Morbach, Benjamin Berg, Bruce Zhang, Carlo Caione, Cedric Viou, Chen
 3668         Qi, Chris Chiu, Chris Down, Chris Morin, Christian Rebischke, Claudius
 3669         Ellsel, Colin Guthrie, dana, Daniel, Daniele Medri, Daniel Kahn
 3670         Gillmor, Daniel Rusek, Daniel van Vugt, Dariusz Gadomski, Dave Reisner,
 3671         David Anderson, Davide Cavalca, David Leeds, David Malcolm, David
 3672         Strauss, David Tardon, Dimitri John Ledkov, Dmitry Torokhov, dj-kaktus,
 3673         Dongsu Park, Elias Probst, Emil Soleyman, Erik Kooistra, Ervin Peters,
 3674         Evgeni Golov, Evgeny Vereshchagin, Fabrice Fontaine, Faheel Ahmad,
 3675         Faizal Luthfi, Felix Yan, Filipe Brandenburger, Franck Bui, Frank
 3676         Schaefer, Frantisek Sumsal, Gautier Husson, Gianluca Boiano, Giuseppe
 3677         Scrivano, glitsj16, Hans de Goede, Harald Hoyer, Harry Mallon, Harshit
 3678         Jain, Helmut Grohne, Henry Tung, Hui Yiqun, imayoda, Insun Pyo, Iwan
 3679         Timmer, Jan Janssen, Jan Pokorný, Jan Synacek, Jason A. Donenfeld,
 3680         javitoom, Jérémy Nouhaud, Jeremy Su, Jiuyang Liu, João Paulo Rechi
 3681         Vita, Joe Hershberger, Joe Rayhawk, Joerg Behrmann, Joerg Steffens,
 3682         Jonas Dorel, Jon Ringle, Josh Soref, Julian Andres Klode, Jun Bo Bi,
 3683         Jürg Billeter, Keith Busch, Khem Raj, Kirill Marinushkin, Larry
 3684         Bernstone, Lennart Poettering, Lion Yang, Li Song, Lorenz
 3685         Hübschle-Schneider, Lubomir Rintel, Lucas Werkmeister, Ludwin Janvier,
 3686         Lukáš Nykrýn, Luke Shumaker, mal, Marc-Antoine Perennou, Marcin
 3687         Skarbek, Marco Trevisan (Treviño), Marian Cepok, Mario Hros, Marko
 3688         Myllynen, Markus Grimm, Martin Pitt, Martin Sobotka, Martin Wilck,
 3689         Mathieu Trudel-Lapierre, Matthew Leeds, Michael Biebl, Michael Olbrich,
 3690         Michael 'pbone' Pobega, Michael Scherer, Michal Koutný, Michal
 3691         Sekletar, Michal Soltys, Mike Gilbert, Mike Palmer, Muhammet Kara, Neal
 3692         Gompa, Neil Brown, Network Silence, Niklas Tibbling, Nikolas Nyby,
 3693         Nogisaka Sadata, Oliver Smith, Patrik Flykt, Pavel Hrdina, Paweł
 3694         Szewczyk, Peter Hutterer, Piotr Drąg, Ray Strode, Reinhold Mueller,
 3695         Renaud Métrich, Roman Gushchin, Ronny Chevalier, Rubén Suárez Alvarez,
 3696         Ruixin Bao, RussianNeuroMancer, Ryutaroh Matsumoto, Saleem Rashid, Sam
 3697         Morris, Samuel Morris, Sandy Carter, scootergrisen, Sébastien Bacher,
 3698         Sergey Ptashnick, Shawn Landden, Shengyao Xue, Shih-Yuan Lee
 3699         (FourDollars), Silvio Knizek, Sjoerd Simons, Stasiek Michalski, Stephen
 3700         Gallagher, Steven Allen, Steve Ramage, Susant Sahani, Sven Joachim,
 3701         Sylvain Plantefève, Tanu Kaskinen, Tejun Heo, Thiago Macieira, Thomas
 3702         Blume, Thomas Haller, Thomas H. P. Andersen, Tim Ruffing, TJ, Tobias
 3703         Jungel, Todd Walton, Tommi Rantala, Tomsod M, Tony Novak, Tore
 3704         Anderson, Trevonn, Victor Laskurain, Victor Tapia, Violet Halo, Vojtech
 3705         Trefny, welaq, William A. Kennington III, William Douglas, Wyatt Ward,
 3706         Xiang Fan, Xi Ruoyao, Xuanwo, Yann E. Morin, YmrDtnJu, Yu Watanabe,
 3707         Zbigniew Jędrzejewski-Szmek, Zhang Xianwei, Zsolt Dollenstein
 3709         — Warsaw, 2018-12-21
 3711 CHANGES WITH 239:
 3713         * NETWORK INTERFACE DEVICE NAMING CHANGES: systemd-udevd's "net_id"
 3714           builtin will name network interfaces differently than in previous
 3715           versions for virtual network interfaces created with SR-IOV and NPAR
 3716           and for devices where the PCI network controller device does not have
 3717           a slot number associated.
 3719           SR-IOV virtual devices are now named based on the name of the parent
 3720           interface, with a suffix of "v<N>", where <N> is the virtual device
 3721           number. Previously those virtual devices were named as if completely
 3722           independent.
 3724           The ninth and later NPAR virtual devices will be named following the
 3725           scheme used for the first eight NPAR partitions. Previously those
 3726           devices were not renamed and the kernel default (eth<n>) was used.
 3728           "net_id" will also generate names for PCI devices where the PCI
 3729           network controller device does not have an associated slot number
 3730           itself, but one of its parents does. Previously those devices were
 3731           not renamed and the kernel default (eth<n>) was used.
 3733         * AF_INET and AF_INET6 are dropped from RestrictAddressFamilies= in
 3734           systemd-logind.service. Since v235, IPAddressDeny=any has been set to
 3735           the unit. So, it is expected that the default behavior of
 3736           systemd-logind is not changed. However, if distribution packagers or
 3737           administrators disabled or modified IPAddressDeny= setting by a
 3738           drop-in config file, then it may be necessary to update the file to
 3739           re-enable AF_INET and AF_INET6 to support network user name services,
 3740           e.g. NIS.
 3742         * When the RestrictNamespaces= unit property is specified multiple
 3743           times, then the specified types are merged now. Previously, only the
 3744           last assignment was used. So, if distribution packagers or
 3745           administrators modified the setting by a drop-in config file, then it
 3746           may be necessary to update the file.
 3748         * When OnFailure= is used in combination with Restart= on a service
 3749           unit, then the specified units will no longer be triggered on
 3750           failures that result in restarting. Previously, the specified units
 3751           would be activated each time the unit failed, even when the unit was
 3752           going to be restarted automatically. This behaviour contradicted the
 3753           documentation. With this release the code is adjusted to match the
 3754           documentation.
 3756         * systemd-tmpfiles will now print a notice whenever it encounters
 3757           tmpfiles.d/ lines referencing the /var/run/ directory. It will
 3758           recommend reworking them to use the /run/ directory instead (for
 3759           which /var/run/ is simply a symlinked compatibility alias). This way
 3760           systemd-tmpfiles can properly detect line conflicts and merge lines
 3761           referencing the same file by two paths, without having to access
 3762           them.
 3764         * systemctl disable/unmask/preset/preset-all cannot be used with
 3765           --runtime. Previously this was allowed, but resulted in unintuitive
 3766           behaviour that wasn't useful. systemctl disable/unmask will now undo
 3767           both runtime and persistent enablement/masking, i.e. it will remove
 3768           any relevant symlinks both in /run and /etc.
 3770         * Note that all long-running system services shipped with systemd will
 3771           now default to a system call allow list (rather than a deny list, as
 3772           before). In particular, systemd-udevd will now enforce one too. For
 3773           most cases this should be safe, however downstream distributions
 3774           which disabled sandboxing of systemd-udevd (specifically the
 3775           MountFlags= setting), might want to disable this security feature
 3776           too, as the default allow-listing will prohibit all mount, swap,
 3777           reboot and clock changing operations from udev rules.
 3779         * sd-boot acquired new loader configuration settings to optionally turn
 3780           off Windows and MacOS boot partition discovery as well as
 3781           reboot-into-firmware menu items. It is also able to pick a better
 3782           screen resolution for HiDPI systems, and now provides loader
 3783           configuration settings to change the resolution explicitly.
 3785         * systemd-resolved now supports DNS-over-TLS. It's still
 3786           turned off by default, use DNSOverTLS=opportunistic to turn it on in
 3787           resolved.conf. We intend to make this the default as soon as couple
 3788           of additional techniques for optimizing the initial latency caused by
 3789           establishing a TLS/TCP connection are implemented.
 3791         * systemd-resolved.service and systemd-networkd.service now set
 3792           DynamicUser=yes. The users systemd-resolve and systemd-network are
 3793           not created by systemd-sysusers anymore.
 3795           NOTE: This has a chance of breaking nss-ldap and similar NSS modules
 3796           that embed a network facing module into any process using getpwuid()
 3797           or related call: the dynamic allocation of the user ID for
 3798           systemd-resolved.service means the service manager has to check NSS
 3799           if the user name is already taken when forking off the service. Since
 3800           the user in the common case won't be defined in /etc/passwd the
 3801           lookup is likely to trigger nss-ldap which in turn might use NSS to
 3802           ask systemd-resolved for hostname lookups. This will hence result in
 3803           a deadlock: a user name lookup in order to start
 3804           systemd-resolved.service will result in a hostname lookup for which
 3805           systemd-resolved.service needs to be started already. There are
 3806           multiple ways to work around this problem: pre-allocate the
 3807           "systemd-resolve" user on such systems, so that nss-ldap won't be
 3808           triggered; or use a different NSS package that doesn't do networking
 3809           in-process but provides a local asynchronous name cache; or configure
 3810           the NSS package to avoid lookups for UIDs in the range `pkg-config
 3811           systemd --variable=dynamicuidmin` … `pkg-config systemd
 3812           --variable=dynamicuidmax`, so that it does not consider itself
 3813           authoritative for the same UID range systemd allocates dynamic users
 3814           from.
 3816         * The systemd-resolve tool has been renamed to resolvectl (it also
 3817           remains available under the old name, for compatibility), and its
 3818           interface is now verb-based, similar in style to the other <xyz>ctl
 3819           tools, such as systemctl or loginctl.
 3821         * The resolvectl/systemd-resolve tool also provides 'resolvconf'
 3822           compatibility. It may be symlinked under the 'resolvconf' name, in
 3823           which case it will take arguments and input compatible with the
 3824           Debian and FreeBSD resolvconf tool.
 3826         * Support for suspend-then-hibernate has been added, i.e. a sleep mode
 3827           where the system initially suspends, and after a timeout resumes and
 3828           hibernates again.
 3830         * networkd's ClientIdentifier= now accepts a new option "duid-only". If
 3831           set the client will only send a DUID as client identifier.
 3833         * The nss-systemd glibc NSS module will now enumerate dynamic users and
 3834           groups in effect. Previously, it could resolve UIDs/GIDs to user
 3835           names/groups and vice versa, but did not support enumeration.
 3837         * journald's Compress= configuration setting now optionally accepts a
 3838           byte threshold value. All journal objects larger than this threshold
 3839           will be compressed, smaller ones will not. Previously this threshold
 3840           was not configurable and set to 512.
 3842         * A new system.conf setting NoNewPrivileges= is now available which may
 3843           be used to turn off acquisition of new privileges system-wide
 3844           (i.e. set Linux' PR_SET_NO_NEW_PRIVS for PID 1 itself, and thus also
 3845           for all its children). Note that turning this option on means setuid
 3846           binaries and file system capabilities lose their special powers.
 3847           While turning on this option is a big step towards a more secure
 3848           system, doing so is likely to break numerous pre-existing UNIX tools,
 3849           in particular su and sudo.
 3851         * A new service systemd-time-sync-wait.service has been added. If
 3852           enabled it will delay the time-sync.target unit at boot until time
 3853           synchronization has been received from the network. This
 3854           functionality is useful on systems lacking a local RTC or where it is
 3855           acceptable that the boot process shall be delayed by external network
 3856           services.
 3858         * When hibernating, systemd will now inform the kernel of the image
 3859           write offset, on kernels new enough to support this. This means swap
 3860           files should work for hibernation now.
 3862         * When loading unit files, systemd will now look for drop-in unit files
 3863           extensions in additional places. Previously, for a unit file name
 3864           "foo-bar-baz.service" it would look for dropin files in
 3865           "foo-bar-baz.service.d/*.conf". Now, it will also look in
 3866           "foo-bar-.service.d/*.conf" and "foo-.service.d/", i.e. at the
 3867           service name truncated after all inner dashes. This scheme allows
 3868           writing drop-ins easily that apply to a whole set of unit files at
 3869           once. It's particularly useful for mount and slice units (as their
 3870           naming is prefix based), but is also useful for service and other
 3871           units, for packages that install multiple unit files at once,
 3872           following a strict naming regime of beginning the unit file name with
 3873           the package's name. Two new specifiers are now supported in unit
 3874           files to match this: %j and %J are replaced by the part of the unit
 3875           name following the last dash.
 3877         * Unit files and other configuration files that support specifier
 3878           expansion now understand another three new specifiers: %T and %V will
 3879           resolve to /tmp and /var/tmp respectively, or whatever temporary
 3880           directory has been set for the calling user. %E will expand to either
 3881           /etc (for system units) or $XDG_CONFIG_HOME (for user units).
 3883         * The ExecStart= lines of unit files are no longer required to
 3884           reference absolute paths. If non-absolute paths are specified the
 3885           specified binary name is searched within the service manager's
 3886           built-in $PATH, which may be queried with 'systemd-path
 3887           search-binaries-default'. It's generally recommended to continue to
 3888           use absolute paths for all binaries specified in unit files.
 3890         * Units gained a new load state "bad-setting", which is used when a
 3891           unit file was loaded, but contained fatal errors which prevent it
 3892           from being started (for example, a service unit has been defined
 3893           lacking both ExecStart= and ExecStop= lines).
 3895         * coredumpctl's "gdb" verb has been renamed to "debug", in order to
 3896           support alternative debuggers, for example lldb. The old name
 3897           continues to be available however, for compatibility reasons. Use the
 3898           new --debugger= switch or the $SYSTEMD_DEBUGGER environment variable
 3899           to pick an alternative debugger instead of the default gdb.
 3901         * systemctl and the other tools will now output escape sequences that
 3902           generate proper clickable hyperlinks in various terminal emulators
 3903           where useful (for example, in the "systemctl status" output you can
 3904           now click on the unit file name to quickly open it in the
 3905           editor/viewer of your choice). Note that not all terminal emulators
 3906           support this functionality yet, but many do. Unfortunately, the
 3907           "less" pager doesn't support this yet, hence this functionality is
 3908           currently automatically turned off when a pager is started (which
 3909           happens quite often due to auto-paging). We hope to remove this
 3910           limitation as soon as "less" learns these escape sequences. This new
 3911           behaviour may also be turned off explicitly with the $SYSTEMD_URLIFY
 3912           environment variable. For details on these escape sequences see:
 3913           https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda
 3915         * networkd's .network files now support a new IPv6MTUBytes= option for
 3916           setting the MTU used by IPv6 explicitly as well as a new MTUBytes=
 3917           option in the [Route] section to configure the MTU to use for
 3918           specific routes. It also gained support for configuration of the DHCP
 3919           "UserClass" option through the new UserClass= setting. It gained
 3920           three new options in the new [CAN] section for configuring CAN
 3921           networks. The MULTICAST and ALLMULTI interface flags may now be
 3922           controlled explicitly with the new Multicast= and AllMulticast=
 3923           settings.
 3925         * networkd will now automatically make use of the kernel's route
 3926           expiration feature, if it is available.
 3928         * udevd's .link files now support setting the number of receive and
 3929           transmit channels, using the RxChannels=, TxChannels=,
 3930           OtherChannels=, CombinedChannels= settings.
 3932         * Support for UDPSegmentationOffload= has been removed, given its
 3933           limited support in hardware, and waning software support.
 3935         * networkd's .netdev files now support creating "netdevsim" interfaces.
 3937         * PID 1 learnt a new bus call GetUnitByControlGroup() which may be used
 3938           to query the unit belonging to a specific kernel control group.
 3940         * systemd-analyze gained a new verb "cat-config", which may be used to
 3941           dump the contents of any configuration file, with all its matching
 3942           drop-in files added in, and honouring the usual search and masking
 3943           logic applied to systemd configuration files. For example use
 3944           "systemd-analyze cat-config systemd/system.conf" to get the complete
 3945           system configuration file of systemd how it would be loaded by PID 1
 3946           itself. Similar to this, various tools such as systemd-tmpfiles or
 3947           systemd-sysusers, gained a new option "--cat-config", which does the
 3948           corresponding operation for their own configuration settings. For
 3949           example, "systemd-tmpfiles --cat-config" will now output the full
 3950           list of tmpfiles.d/ lines in place.
 3952         * timedatectl gained three new verbs: "show" shows bus properties of
 3953           systemd-timedated, "timesync-status" shows the current NTP
 3954           synchronization state of systemd-timesyncd, and "show-timesync"
 3955           shows bus properties of systemd-timesyncd.
 3957         * systemd-timesyncd gained a bus interface on which it exposes details
 3958           about its state.
 3960         * A new environment variable $SYSTEMD_TIMEDATED_NTP_SERVICES is now
 3961           understood by systemd-timedated. It takes a colon-separated list of
 3962           unit names of NTP client services. The list is used by
 3963           "timedatectl set-ntp".
 3965         * systemd-nspawn gained a new --rlimit= switch for setting initial
 3966           resource limits for the container payload. There's a new switch
 3967           --hostname= to explicitly override the container's hostname. A new
 3968           --no-new-privileges= switch may be used to control the
 3969           PR_SET_NO_NEW_PRIVS flag for the container payload. A new
 3970           --oom-score-adjust= switch controls the OOM scoring adjustment value
 3971           for the payload. The new --cpu-affinity= switch controls the CPU
 3972           affinity of the container payload. The new --resolv-conf= switch
 3973           allows more detailed control of /etc/resolv.conf handling of the
 3974           container. Similarly, the new --timezone= switch allows more detailed
 3975           control of /etc/localtime handling of the container.
 3977         * systemd-detect-virt gained a new --list switch, which will print a
 3978           list of all currently known VM and container environments.
 3980         * Support for "Portable Services" has been added, see
 3981           doc/PORTABLE_SERVICES.md for details. Currently, the support is still
 3982           experimental, but this is expected to change soon. Reflecting this
 3983           experimental state, the "portablectl" binary is not installed into
 3984           /usr/bin yet. The binary has to be called with the full path
 3985           /usr/lib/systemd/portablectl instead.
 3987         * journalctl's and systemctl's -o switch now knows a new log output
 3988           mode "with-unit". The output it generates is very similar to the
 3989           regular "short" mode, but displays the unit name instead of the
 3990           syslog tag for each log line. Also, the date is shown with timezone
 3991           information. This mode is probably more useful than the classic
 3992           "short" output mode for most purposes, except where pixel-perfect
 3993           compatibility with classic /var/log/messages formatting is required.
 3995         * A new --dump-bus-properties switch has been added to the systemd
 3996           binary, which may be used to dump all supported D-Bus properties.
 3997           (Options which are still supported, but are deprecated, are *not*
 3998           shown.)
 4000         * sd-bus gained a set of new calls:
 4001           sd_bus_slot_set_floating()/sd_bus_slot_get_floating() may be used to
 4002           enable/disable the "floating" state of a bus slot object,
 4003           i.e. whether the slot object pins the bus it is allocated for into
 4004           memory or if the bus slot object gets disconnected when the bus goes
 4005           away. sd_bus_open_with_description(),
 4006           sd_bus_open_user_with_description(),
 4007           sd_bus_open_system_with_description() may be used to allocate bus
 4008           objects and set their description string already during allocation.
 4010         * sd-event gained support for watching inotify events from the event
 4011           loop, in an efficient way, sharing inotify handles between multiple
 4012           users. For this a new function sd_event_add_inotify() has been added.
 4014         * sd-event and sd-bus gained support for calling special user-supplied
 4015           destructor functions for userdata pointers associated with
 4016           sd_event_source, sd_bus_slot, and sd_bus_track objects. For this new
 4017           functions sd_bus_slot_set_destroy_callback,
 4018           sd_bus_slot_get_destroy_callback, sd_bus_track_set_destroy_callback,
 4019           sd_bus_track_get_destroy_callback,
 4020           sd_event_source_set_destroy_callback,
 4021           sd_event_source_get_destroy_callback have been added.
 4023         * The "net.ipv4.tcp_ecn" sysctl will now be turned on by default.
 4025         * PID 1 will now automatically reschedule .timer units whenever the
 4026           local timezone changes. (They previously got rescheduled
 4027           automatically when the system clock changed.)
 4029         * New documentation has been added to document cgroups delegation,
 4030           portable services and the various code quality tools we have set up:
 4032           https://github.com/systemd/systemd/blob/master/docs/CGROUP_DELEGATION.md
 4033           https://github.com/systemd/systemd/blob/master/docs/PORTABLE_SERVICES.md
 4034           https://github.com/systemd/systemd/blob/master/docs/CODE_QUALITY.md
 4036         * The Boot Loader Specification has been added to the source tree.
 4038           https://github.com/systemd/systemd/blob/master/docs/BOOT_LOADER_SPECIFICATION.md
 4040           While moving it into our source tree we have updated it and further
 4041           changes are now accepted through the usual github PR workflow.
 4043         * pam_systemd will now look for PAM userdata fields systemd.memory_max,
 4044           systemd.tasks_max, systemd.cpu_weight, systemd.io_weight set by
 4045           earlier PAM modules. The data in these fields is used to initialize
 4046           the session scope's resource properties. Thus external PAM modules
 4047           may now configure per-session limits, for example sourced from
 4048           external user databases.
 4050         * socket units with Accept=yes will now maintain a "refused" counter in
 4051           addition to the existing "accepted" counter, counting connections
 4052           refused due to the enforced limits.
 4054         * The "systemd-path search-binaries-default" command may now be use to
 4055           query the default, built-in $PATH PID 1 will pass to the services it
 4056           manages.
 4058         * A new unit file setting PrivateMounts= has been added. It's a boolean
 4059           option. If enabled the unit's processes are invoked in their own file
 4060           system namespace. Note that this behaviour is also implied if any
 4061           other file system namespacing options (such as PrivateTmp=,
 4062           PrivateDevices=, ProtectSystem=, …) are used. This option is hence
 4063           primarily useful for services that do not use any of the other file
 4064           system namespacing options. One such service is systemd-udevd.service
 4065           where this is now used by default.
 4067         * ConditionSecurity= gained a new value "uefi-secureboot" that is true
 4068           when the system is booted in UEFI "secure mode".
 4070         * A new unit "system-update-pre.target" is added, which defines an
 4071           optional synchronization point for offline system updates, as
 4072           implemented by the pre-existing "system-update.target" unit. It
 4073           allows ordering services before the service that executes the actual
 4074           update process in a generic way.
 4076         * Systemd now emits warnings whenever .include syntax is used.
 4078         Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale,
 4079         Alexander Kurtz, Alex Gartrell, Anssi Hannula, Arnaud Rebillout, Brian
 4080         J. Murrell, Bruno Vernay, Chris Lamb, Chris Lesiak, Christian Brauner,
 4081         Christian Hesse, Christian Rebischke, Colin Guthrie, Daniel Dao, Daniel
 4082         Lin, Danylo Korostil, Davide Cavalca, David Tardon, Dimitri John
 4083         Ledkov, Dmitriy Geels, Douglas Christman, Elia Geretto, emelenas, Emil
 4084         Velikov, Evgeny Vereshchagin, Felipe Sateler, Feng Sun, Filipe
 4085         Brandenburger, Franck Bui, futpib, Giuseppe Scrivano, Guillem Jover,
 4086         guixxx, Hannes Reinecke, Hans de Goede, Harald Hoyer, Henrique Dante de
 4087         Almeida, Hiram van Paassen, Ian Miell, Igor Gnatenko, Ivan Shapovalov,
 4088         Iwan Timmer, James Cowgill, Jan Janssen, Jan Synacek, Jared Kazimir,
 4089         Jérémy Rosen, João Paulo Rechi Vita, Joost Heitbrink, Jui-Chi Ricky
 4090         Liang, Jürg Billeter, Kai-Heng Feng, Karol Augustin, Kay Sievers,
 4091         Krzysztof Nowicki, Lauri Tirkkonen, Lennart Poettering, Leonard König,
 4092         Long Li, Luca Boccassi, Lucas Werkmeister, Marcel Hoppe, Marc
 4093         Kleine-Budde, Mario Limonciello, Martin Jansa, Martin Wilck, Mathieu
 4094         Malaterre, Matteo F. Vescovi, Matthew McGinn, Matthias-Christian Ott,
 4095         Michael Biebl, Michael Olbrich, Michael Prokop, Michal Koutný, Michal
 4096         Sekletar, Mike Gilbert, Mikhail Kasimov, Milan Broz, Milan Pässler,
 4097         Mladen Pejaković, Muhammet Kara, Nicolas Boichat, Omer Katz, Paride
 4098         Legovini, Paul Menzel, Paul Milliken, Pavel Hrdina, Peter A. Bigot,
 4099         Peter D'Hoye, Peter Hutterer, Peter Jones, Philip Sequeira, Philip
 4100         Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de Araujo,
 4101         Ronny Chevalier, Rosen Penev, Rubén Suárez Alvarez, Ryan Gonzalez,
 4102         Salvo Tomaselli, Sebastian Reichel, Sergey Ptashnick, Sergio Lindo
 4103         Mansilla, Stefan Schweter, Stephen Hemminger, Stuart Hayes, Susant
 4104         Sahani, Sylvain Plantefève, Thomas H. P. Andersen, Tobias Jungel,
 4105         Tomasz Torcz, Vito Caputo, Will Dietz, Will Thompson, Wim van Mourik,
 4106         Yu Watanabe, Zbigniew Jędrzejewski-Szmek
 4108         — Berlin, 2018-06-22
 4110 CHANGES WITH 238:
 4112         * The MemoryAccounting= unit property now defaults to on. After
 4113           discussions with the upstream control group maintainers we learnt
 4114           that the negative impact of cgroup memory accounting on current
 4115           kernels is finally relatively minimal, so that it should be safe to
 4116           enable this by default without affecting system performance. Besides
 4117           memory accounting only task accounting is turned on by default, all
 4118           other forms of resource accounting (CPU, IO, IP) remain off for now,
 4119           because it's not clear yet that their impact is small enough to move
 4120           from opt-in to opt-out. We recommend downstreams to leave memory
 4121           accounting on by default if kernel 4.14 or higher is primarily
 4122           used. On very resource constrained systems or when support for old
 4123           kernels is a necessity, -Dmemory-accounting-default=false can be used
 4124           to revert this change.
 4126         * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update,
 4127           %udev_rules_update) and the journal catalog (%journal_catalog_update)
 4128           from the upgrade scriptlets of individual packages now do nothing.
 4129           Transfiletriggers have been added which will perform those updates
 4130           once at the end of the transaction.
 4132           Similar transfiletriggers have been added to execute any sysctl.d
 4133           and binfmt.d rules. Thus, it should be unnecessary to provide any
 4134           scriptlets to execute this configuration from package installation
 4135           scripts.
 4137         * systemd-sysusers gained a mode where the configuration to execute is
 4138           specified on the command line, but this configuration is not executed
 4139           directly, but instead it is merged with the configuration on disk,
 4140           and the result is executed. This is useful for package installation
 4141           scripts which want to create the user before installing any files on
 4142           disk (in case some of those files are owned by that user), while
 4143           still allowing local admin overrides.
 4145           This functionality is exposed to rpm scriptlets through a new
 4146           %sysusers_create_package macro. Old %sysusers_create and
 4147           %sysusers_create_inline macros are deprecated.
 4149           A transfiletrigger for sysusers.d configuration is now installed,
 4150           which means that it should be unnecessary to call systemd-sysusers from
 4151           package installation scripts, unless the package installs any files
 4152           owned by those newly-created users, in which case
 4153           %sysusers_create_package should be used.
 4155         * Analogous change has been done for systemd-tmpfiles: it gained a mode
 4156           where the command-line configuration is merged with the configuration
 4157           on disk. This is exposed as the new %tmpfiles_create_package macro,
 4158           and %tmpfiles_create is deprecated. A transfiletrigger is installed
 4159           for tmpfiles.d, hence it should be unnecessary to call systemd-tmpfiles
 4160           from package installation scripts.
 4162         * sysusers.d configuration for a user may now also specify the group
 4163           number, in addition to the user number ("u username 123:456"), or
 4164           without the user number ("u username -:456").
 4166         * Configution items for systemd-sysusers can now be specified as
 4167           positional arguments when the new --inline switch is used.
 4169         * The login shell of users created through sysusers.d may now be
 4170           specified (previously, it was always /bin/sh for root and
 4171           /sbin/nologin for other users).
 4173         * systemd-analyze gained a new --global switch to look at global user
 4174           configuration. It also gained a unit-paths verb to list the unit load
 4175           paths that are compiled into systemd (which can be used with
 4176           --systemd, --user, or --global).
 4178         * udevadm trigger gained a new --settle/-w option to wait for any
 4179           triggered events to finish (but just those, and not any other events
 4180           which are triggered meanwhile).
 4182         * The action that systemd-logind takes when the lid is closed and the
 4183           machine is connected to external power can now be configured using
 4184           HandleLidSwitchExternalPower= in logind.conf. Previously, this action
 4185           was determined by HandleLidSwitch=, and, for backwards compatibility,
 4186           is still is, if HandleLidSwitchExternalPower= is not explicitly set.
 4188         * journalctl will periodically call sd_journal_process() to make it
 4189           resilient against inotify queue overruns when journal files are
 4190           rotated very quickly.
 4192         * Two new functions in libsystemd — sd_bus_get_n_queued_read and
 4193           sd_bus_get_n_queued_write — may be used to check the number of
 4194           pending bus messages.
 4196         * systemd gained a new
 4197           org.freedesktop.systemd1.Manager.AttachProcessesToUnit dbus call
 4198           which can be used to migrate foreign processes to scope and service
 4199           units. The primary user for this new API is systemd itself: the
 4200           systemd --user instance uses this call of the systemd --system
 4201           instance to migrate processes if it itself gets the request to
 4202           migrate processes and the kernel refuses this due to access
 4203           restrictions.  Thanks to this "systemd-run --scope --user …" works
 4204           again in pure cgroup v2 environments when invoked from the user
 4205           session scope.
 4207         * A new TemporaryFileSystem= setting can be used to mask out part of
 4208           the real file system tree with tmpfs mounts. This may be combined
 4209           with BindPaths= and BindReadOnlyPaths= to hide files or directories
 4210           not relevant to the unit, while still allowing some paths lower in
 4211           the tree to be accessed.
 4213           ProtectHome=tmpfs may now be used to hide user home and runtime
 4214           directories from units, in a way that is mostly equivalent to
 4215           "TemporaryFileSystem=/home /run/user /root".
 4217         * Non-service units are now started with KeyringMode=shared by default.
 4218           This means that mount and swapon and other mount tools have access
 4219           to keys in the main keyring.
 4221         * /sys/fs/bpf is now mounted automatically.
 4223         * QNX virtualization is now detected by systemd-detect-virt and may
 4224           be used in ConditionVirtualization=.
 4226         * IPAccounting= may now be enabled also for slice units.
 4228         * A new -Dsplit-bin= build configuration switch may be used to specify
 4229           whether bin and sbin directories are merged, or if they should be
 4230           included separately in $PATH and various listings of executable
 4231           directories. The build configuration scripts will try to autodetect
 4232           the proper values of -Dsplit-usr= and -Dsplit-bin= based on build
 4233           system, but distributions are encouraged to configure this
 4234           explicitly.
 4236         * A new -Dok-color= build configuration switch may be used to change
 4237           the colour of "OK" status messages.
 4239         * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with
 4240           PrivateNetwork=yes was buggy in previous versions of systemd. This
 4241           means that after the upgrade and daemon-reexec, any such units must
 4242           be restarted.
 4244         * INCOMPATIBILITY: as announced in the NEWS for 237, systemd-tmpfiles
 4245           will not exclude read-only files owned by root from cleanup.
 4247         Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet,
 4248         Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337,
 4249         Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo
 4250         de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel
 4251         Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny
 4252         Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib,
 4253         Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer,
 4254         Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld,
 4255         Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas
 4256         Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt,
 4257         MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren,
 4258         Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert
 4259         Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
 4260         Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain
 4261         Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe,
 4262         Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić)
 4264         — Warsaw, 2018-03-05
 4266 CHANGES WITH 237:
 4268         * Some keyboards come with a zoom see-saw or rocker which until now got
 4269           mapped to the Linux "zoomin/out" keys in hwdb. However, these
 4270           keycodes are not recognized by any major desktop. They now produce
 4271           Up/Down key events so that they can be used for scrolling.
 4273         * INCOMPATIBILITY: systemd-tmpfiles' "f" lines changed behaviour
 4274           slightly: previously, if an argument was specified for lines of this
 4275           type (i.e. the right-most column was set) this string was appended to
 4276           existing files each time systemd-tmpfiles was run. This behaviour was
 4277           different from what the documentation said, and not particularly
 4278           useful, as repeated systemd-tmpfiles invocations would not be
 4279           idempotent and grow such files without bounds. With this release
 4280           behaviour has been altered to match what the documentation says:
 4281           lines of this type only have an effect if the indicated files don't
 4282           exist yet, and only then the argument string is written to the file.
 4284         * FUTURE INCOMPATIBILITY: In systemd v238 we intend to slightly change
 4285           systemd-tmpfiles behaviour: previously, read-only files owned by root
 4286           were always excluded from the file "aging" algorithm (i.e. the
 4287           automatic clean-up of directories like /tmp based on
 4288           atime/mtime/ctime). We intend to drop this restriction, and age files
 4289           by default even when owned by root and read-only. This behaviour was
 4290           inherited from older tools, but there have been requests to remove
 4291           it, and it's not obvious why this restriction was made in the first
 4292           place. Please speak up now, if you are aware of software that requires
 4293           this behaviour, otherwise we'll remove the restriction in v238.
 4295         * A new environment variable $SYSTEMD_OFFLINE is now understood by
 4296           systemctl. It takes a boolean argument. If on, systemctl assumes it
 4297           operates on an "offline" OS tree, and will not attempt to talk to the
 4298           service manager. Previously, this mode was implicitly enabled if a
 4299           chroot() environment was detected, and this new environment variable
 4300           now provides explicit control.
 4302         * .path and .socket units may now be created transiently, too.
 4303           Previously only service, mount, automount and timer units were
 4304           supported as transient units. The systemd-run tool has been updated
 4305           to expose this new functionality, you may hence use it now to bind
 4306           arbitrary commands to path or socket activation on-the-fly from the
 4307           command line. Moreover, almost all properties are now exposed for the
 4308           unit types that already supported transient operation.
 4310         * The systemd-mount command gained support for a new --owner= parameter
 4311           which takes a user name, which is then resolved and included in uid=
 4312           and gid= mount options string of the file system to mount.
 4314         * A new unit condition ConditionControlGroupController= has been added
 4315           that checks whether a specific cgroup controller is available.
 4317         * Unit files, udev's .link files, and systemd-networkd's .netdev and
 4318           .network files all gained support for a new condition
 4319           ConditionKernelVersion= for checking against specific kernel
 4320           versions.
 4322         * In systemd-networkd, the [IPVLAN] section in .netdev files gained
 4323           support for configuring device flags in the Flags= setting. In the
 4324           same files, the [Tunnel] section gained support for configuring
 4325           AllowLocalRemote=.  The [Route] section in .network files gained
 4326           support for configuring InitialCongestionWindow=,
 4327           InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now
 4328           understands RapidCommit=.
 4330         * systemd-networkd's DHCPv6 support gained support for Prefix
 4331           Delegation.
 4333         * sd-bus gained support for a new "watch-bind" feature. When this
 4334           feature is enabled, an sd_bus connection may be set up to connect to
 4335           an AF_UNIX socket in the file system as soon as it is created. This
 4336           functionality is useful for writing early-boot services that
 4337           automatically connect to the system bus as soon as it is started,
 4338           without ugly time-based polling. systemd-networkd and
 4339           systemd-resolved have been updated to make use of this
 4340           functionality. busctl exposes this functionality in a new
 4341           --watch-bind= command line switch.
 4343         * sd-bus will now optionally synthesize a local "Connected" signal as
 4344           soon as a D-Bus connection is set up fully. This message mirrors the
 4345           already existing "Disconnected" signal which is synthesized when the
 4346           connection is terminated. This signal is generally useful but
 4347           particularly handy in combination with the "watch-bind" feature
 4348           described above. Synthesizing of this message has to be requested
 4349           explicitly through the new API call sd_bus_set_connected_signal(). In
 4350           addition a new call sd_bus_is_ready() has been added that checks
 4351           whether a connection is fully set up (i.e. between the "Connected" and
 4352           "Disconnected" signals).
 4354         * sd-bus gained two new calls sd_bus_request_name_async() and
 4355           sd_bus_release_name_async() for asynchronously registering bus
 4356           names. Similar, there is now sd_bus_add_match_async() for installing
 4357           a signal match asynchronously. All of systemd's own services have
 4358           been updated to make use of these calls. Doing these operations
 4359           asynchronously has two benefits: it reduces the risk of deadlocks in
 4360           case of cyclic dependencies between bus services, and it speeds up
 4361           service initialization since synchronization points for bus
 4362           round-trips are removed.
 4364         * sd-bus gained two new calls sd_bus_match_signal() and
 4365           sd_bus_match_signal_async(), which are similar to sd_bus_add_match()
 4366           and sd_bus_add_match_async() but instead of taking a D-Bus match
 4367           string take match fields as normal function parameters.
 4369         * sd-bus gained two new calls sd_bus_set_sender() and
 4370           sd_bus_message_set_sender() for setting the sender name of outgoing
 4371           messages (either for all outgoing messages or for just one specific
 4372           one). These calls are only useful in direct connections as on
 4373           brokered connections the broker fills in the sender anyway,
 4374           overwriting whatever the client filled in.
 4376         * sd-event gained a new pseudo-handle that may be specified on all API
 4377           calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When
 4378           used this refers to the default event loop object of the calling
 4379           thread. Note however that this does not implicitly allocate one —
 4380           which has to be done prior by using sd_event_default(). Similarly
 4381           sd-bus gained three new pseudo-handles SD_BUS_DEFAULT,
 4382           SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer
 4383           to the default bus of the specified type of the calling thread. Here
 4384           too this does not implicitly allocate bus connection objects, this
 4385           has to be done prior with sd_bus_default() and friends.
 4387         * sd-event gained a new call pair
 4388           sd_event_source_{get|set}_io_fd_own(). This may be used to request
 4389           automatic closure of the file descriptor an IO event source watches
 4390           when the event source is destroyed.
 4392         * systemd-networkd gained support for natively configuring WireGuard
 4393           connections.
 4395         * In previous versions systemd synthesized user records both for the
 4396           "nobody" (UID 65534) and "root" (UID 0) users in nss-systemd and
 4397           internally. In order to simplify distribution-wide renames of the
 4398           "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a
 4399           new transitional flag file has been added: if
 4400           /etc/systemd/dont-synthesize-nobody exists synthesizing of the 65534
 4401           user and group record within the systemd codebase is disabled.
 4403         * systemd-notify gained a new --uid= option for selecting the source
 4404           user/UID to use for notification messages sent to the service
 4405           manager.
 4407         * journalctl gained a new --grep= option to list only entries in which
 4408           the message matches a certain pattern. By default matching is case
 4409           insensitive if the pattern is lowercase, and case sensitive
 4410           otherwise. Option --case-sensitive=yes|no can be used to override
 4411           this an specify case sensitivity or case insensitivity.
 4413         * There's now a "systemd-analyze service-watchdogs" command for printing
 4414           the current state of the service runtime watchdog, and optionally
 4415           enabling or disabling the per-service watchdogs system-wide if given a
 4416           boolean argument (i.e. the concept you configure in WatchdogSec=), for
 4417           debugging purposes. There's also a kernel command line option
 4418           systemd.service_watchdogs= for controlling the same.
 4420         * Two new "log-level" and "log-target" options for systemd-analyze were
 4421           added that merge the now deprecated get-log-level, set-log-level and
 4422           get-log-target, set-log-target pairs. The deprecated options are still
 4423           understood for backwards compatibility. The two new options print the
 4424           current value when no arguments are given, and set them when a
 4425           level/target is given as an argument.
 4427         * sysusers.d's "u" lines now optionally accept both a UID and a GID
 4428           specification, separated by a ":" character, in order to create users
 4429           where UID and GID do not match.
 4431         Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov,
 4432         Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman
 4433         Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton
 4434         Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov,
 4435         Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui,
 4436         Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian
 4437         Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander
 4438         Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen,
 4439         Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg
 4440         Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering,
 4441         Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt,
 4442         Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy,
 4443         Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał
 4444         Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf
 4445         Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer,
 4446         Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer,
 4447         Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani,
 4448         Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz
 4449         Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary
 4450         Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян
 4451         Палаузов
 4453         — Brno, 2018-01-28
 4455 CHANGES WITH 236:
 4457         * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
 4458           in v235 has been extended to also set the dummy.ko module option
 4459           numdummies=0, preventing the kernel from automatically creating
 4460           dummy0. All dummy interfaces must now be explicitly created.
 4462         * Unknown '%' specifiers in configuration files are now rejected. This
 4463           applies to units and tmpfiles.d configuration. Any percent characters
 4464           that are followed by a letter or digit that are not supposed to be
 4465           interpreted as the beginning of a specifier should be escaped by
 4466           doubling ("%%").  (So "size=5%" is still accepted, as well as
 4467           "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
 4468           valid specifiers today.)
 4470         * systemd-resolved now maintains a new dynamic
 4471           /run/systemd/resolve/stub-resolv.conf compatibility file. It is
 4472           recommended to make /etc/resolv.conf a symlink to it. This file
 4473           points at the systemd-resolved stub DNS resolver and
 4474           includes dynamically acquired search domains, achieving more correct
 4475           DNS resolution by software that bypasses local DNS APIs such as NSS.
 4477         * The "uaccess" udev tag has been dropped from /dev/kvm and
 4478           /dev/dri/renderD*.  These devices now have the 0666 permissions by
 4479           default (but this may be changed at build-time). /dev/dri/renderD*
 4480           will now be owned by the "render" group along with /dev/kfd.
 4482         * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
 4483           systemd-journal-gatewayd.service and
 4484           systemd-journal-upload.service. This means "nss-systemd" must be
 4485           enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
 4486           services are resolved properly.
 4488         * In /etc/fstab two new mount options are now understood:
 4489           x-systemd.makefs and x-systemd.growfs. The former has the effect that
 4490           the configured file system is formatted before it is mounted, the
 4491           latter that the file system is resized to the full block device size
 4492           after it is mounted (i.e. if the file system is smaller than the
 4493           partition it resides on, it's grown). This is similar to the fsck
 4494           logic in /etc/fstab, and pulls in systemd-makefs@.service and
 4495           systemd-growfs@.service as necessary, similar to
 4496           systemd-fsck@.service. Resizing is currently only supported on ext4
 4497           and btrfs.
 4499         * In systemd-networkd, the IPv6 RA logic now optionally may announce
 4500           DNS server and domain information.
 4502         * Support for the LUKS2 on-disk format for encrypted partitions has
 4503           been added. This requires libcryptsetup2 during compilation and
 4504           runtime.
 4506         * The systemd --user instance will now signal "readiness" when its
 4507           basic.target unit has been reached, instead of when the run queue ran
 4508           empty for the first time.
 4510         * Tmpfiles.d with user configuration are now also supported.
 4511           systemd-tmpfiles gained a new --user switch, and snippets placed in
 4512           ~/.config/user-tmpfiles.d/ and corresponding directories will be
 4513           executed by systemd-tmpfiles --user running in the new
 4514           systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
 4515           running in the user session.
 4517         * Unit files and tmpfiles.d snippets learnt three new % specifiers:
 4518           %S resolves to the top-level state directory (/var/lib for the system
 4519           instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
 4520           top-level cache directory (/var/cache for the system instance,
 4521           $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
 4522           logs directory (/var/log for the system instance,
 4523           $XDG_CONFIG_HOME/log/ for the user instance). This matches the
 4524           existing %t specifier, that resolves to the top-level runtime
 4525           directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
 4526           user instance).
 4528         * journalctl learnt a new parameter --output-fields= for limiting the
 4529           set of journal fields to output in verbose and JSON output modes.
 4531         * systemd-timesyncd's configuration file gained a new option
 4532           RootDistanceMaxSec= for setting the maximum root distance of servers
 4533           it'll use, as well as the new options PollIntervalMinSec= and
 4534           PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
 4536         * bootctl gained a new command "list" for listing all available boot
 4537           menu items on systems that follow the boot loader specification.
 4539         * systemctl gained a new --dry-run switch that shows what would be done
 4540           instead of doing it, and is currently supported by the shutdown and
 4541           sleep verbs.
 4543         * ConditionSecurity= can now detect the TOMOYO security module.
 4545         * Unit file [Install] sections are now also respected in unit drop-in
 4546           files. This is intended to be used by drop-ins under /usr/lib/.
 4548         * systemd-firstboot may now also set the initial keyboard mapping.
 4550         * Udev "changed" events for devices which are exposed as systemd
 4551           .device units are now propagated to units specified in
 4552           ReloadPropagatedFrom= as reload requests.
 4554         * If a udev device has a SYSTEMD_WANTS= property containing a systemd
 4555           unit template name (i.e. a name in the form of 'foobar@.service',
 4556           without the instance component between the '@' and - the '.'), then
 4557           the escaped sysfs path of the device is automatically used as the
 4558           instance.
 4560         * SystemCallFilter= in unit files has been extended so that an "errno"
 4561           can be specified individually for each system call. Example:
 4562           SystemCallFilter=~uname:EILSEQ.
 4564         * The cgroup delegation logic has been substantially updated. Delegate=
 4565           now optionally takes a list of controllers (instead of a boolean, as
 4566           before), which lists the controllers to delegate at least.
 4568         * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
 4570         * A new LogLevelMax= setting configures the maximum log level any
 4571           process of the service may log at (i.e. anything with a lesser
 4572           priority than what is specified is automatically dropped). A new
 4573           LogExtraFields= setting allows configuration of additional journal
 4574           fields to attach to all log records generated by any of the unit's
 4575           processes.
 4577         * New StandardInputData= and StandardInputText= settings along with the
 4578           new option StandardInput=data may be used to configure textual or
 4579           binary data that shall be passed to the executed service process via
 4580           standard input, encoded in-line in the unit file.
 4582         * StandardInput=, StandardOutput= and StandardError= may now be used to
 4583           connect stdin/stdout/stderr of executed processes directly with a
 4584           file or AF_UNIX socket in the file system, using the new "file:" option.
 4586         * A new unit file option CollectMode= has been added, that allows
 4587           tweaking the garbage collection logic for units. It may be used to
 4588           tell systemd to garbage collect units that have failed automatically
 4589           (normally it only GCs units that exited successfully). systemd-run
 4590           and systemd-mount expose this new functionality with a new -G option.
 4592         * "machinectl bind" may now be used to bind mount non-directories
 4593           (i.e. regularfiles, devices, fifos, sockets).
 4595         * systemd-analyze gained a new verb "calendar" for validating and
 4596           testing calendar time specifications to use for OnCalendar= in timer
 4597           units. Besides validating the expression it will calculate the next
 4598           time the specified expression would elapse.
 4600         * In addition to the pre-existing FailureAction= unit file setting
 4601           there's now SuccessAction=, for configuring a shutdown action to
 4602           execute when a unit completes successfully. This is useful in
 4603           particular inside containers that shall terminate after some workload
 4604           has been completed. Also, both options are now supported for all unit
 4605           types, not just services.
 4607         * networkds's IP rule support gained two new options
 4608           IncomingInterface= and OutgoingInterface= for configuring the incoming
 4609           and outgoing interfaces of configured rules. systemd-networkd also
 4610           gained support for "vxcan" network devices.
 4612         * networkd gained a new setting RequiredForOnline=, taking a
 4613           boolean. If set, systemd-wait-online will take it into consideration
 4614           when determining that the system is up, otherwise it will ignore the
 4615           interface for this purpose.
 4617         * The sd_notify() protocol gained support for a new operation: with
 4618           FDSTOREREMOVE=1 file descriptors may be removed from the per-service
 4619           store again, ahead of POLLHUP or POLLERR when they are removed
 4620           anyway.
 4622         * A new document doc/UIDS-GIDS.md has been added to the source tree,
 4623           that documents the UID/GID range and assignment assumptions and
 4624           requirements of systemd.
 4626         * The watchdog device PID 1 will ping may now be configured through the
 4627           WatchdogDevice= configuration file setting, or by setting the
 4628           systemd.watchdog_service= kernel commandline option.
 4630         * systemd-resolved's gained support for registering DNS-SD services on
 4631           the local network using MulticastDNS. Services may either be
 4632           registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
 4633           the same dir below /run, /usr/lib), or through its D-Bus API.
 4635         * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
 4636           extend the effective start, runtime, and stop time. The service must
 4637           continue to send EXTEND_TIMEOUT_USEC within the period specified to
 4638           prevent the service manager from making the service as timedout.
 4640         * systemd-resolved's DNSSEC support gained support for RFC 8080
 4641           (Ed25519 keys and signatures).
 4643         * The systemd-resolve command line tool gained a new set of options
 4644           --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
 4645           --set-nta= and --revert to configure per-interface DNS configuration
 4646           dynamically during runtime. It's useful for pushing DNS information
 4647           into systemd-resolved from DNS hook scripts that various interface
 4648           managing software supports (such as pppd).
 4650         * systemd-nspawn gained a new --network-namespace-path= command line
 4651           option, which may be used to make a container join an existing
 4652           network namespace, by specifying a path to a "netns" file.
 4654         Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
 4655         Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
 4656         Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
 4657         Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
 4658         John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
 4659         Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
 4660         Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
 4661         Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
 4662         Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
 4663         Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
 4664         Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
 4665         Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
 4666         Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
 4667         Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
 4668         Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
 4669         Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
 4670         Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
 4671         Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
 4672         Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
 4673         Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
 4674         Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
 4675         Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
 4676         Jędrzejewski-Szmek, Zeal Jagannatha
 4678         — Berlin, 2017-12-14
 4680 CHANGES WITH 235:
 4682         * INCOMPATIBILITY: systemd-logind.service and other long-running
 4683           services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
 4684           communication with the outside. This generally improves security of
 4685           the system, and is in almost all cases a safe and good choice, as
 4686           these services do not and should not provide any network-facing
 4687           functionality. However, systemd-logind uses the glibc NSS API to
 4688           query the user database. This creates problems on systems where NSS
 4689           is set up to directly consult network services for user database
 4690           lookups. In particular, this creates incompatibilities with the
 4691           "nss-nis" module, which attempts to directly contact the NIS/YP
 4692           network servers it is configured for, and will now consistently
 4693           fail. In such cases, it is possible to turn off IP sandboxing for
 4694           systemd-logind.service (set IPAddressDeny= in its [Service] section
 4695           to the empty string, via a .d/ unit file drop-in). Downstream
 4696           distributions might want to update their nss-nis packaging to include
 4697           such a drop-in snippet, accordingly, to hide this incompatibility
 4698           from the user. Another option is to make use of glibc's nscd service
 4699           to proxy such network requests through a privilege-separated, minimal
 4700           local caching daemon, or to switch to more modern technologies such
 4701           sssd, whose NSS hook-ups generally do not involve direct network
 4702           access. In general, we think it's definitely time to question the
 4703           implementation choices of nss-nis, i.e. whether it's a good idea
 4704           today to embed a network-facing loadable module into all local
 4705           processes that need to query the user database, including the most
 4706           trivial and benign ones, such as "ls". For more details about
 4707           IPAddressDeny= see below.
 4709         * A new modprobe.d drop-in is now shipped by default that sets the
 4710           bonding module option max_bonds=0. This overrides the kernel default,
 4711           to avoid conflicts and ambiguity as to whether or not bond0 should be
 4712           managed by systemd-networkd or not. This resolves multiple issues
 4713           with bond0 properties not being applied, when bond0 is configured
 4714           with systemd-networkd. Distributors may choose to not package this,
 4715           however in that case users will be prevented from correctly managing
 4716           bond0 interface using systemd-networkd.
 4718         * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
 4719           which print the logging level and target of the system manager. They
 4720           complement the existing "set-log-level" and "set-log-target" verbs
 4721           used to change those values.
 4723         * journald.conf gained a new boolean setting ReadKMsg= which defaults
 4724           to on. If turned off kernel log messages will not be read by
 4725           systemd-journald or included in the logs. It also gained a new
 4726           setting LineMax= for configuring the maximum line length in
 4727           STDOUT/STDERR log streams. The new default for this value is 48K, up
 4728           from the previous hardcoded 2048.
 4730         * A new unit setting RuntimeDirectoryPreserve= has been added, which
 4731           allows more detailed control of what to do with a runtime directory
 4732           configured with RuntimeDirectory= (i.e. a directory below /run or
 4733           $XDG_RUNTIME_DIR) after a unit is stopped.
 4735         * The RuntimeDirectory= setting for units gained support for creating
 4736           deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
 4737           one top-level directory.
 4739         * Units gained new options StateDirectory=, CacheDirectory=,
 4740           LogsDirectory= and ConfigurationDirectory= which are closely related
 4741           to RuntimeDirectory= but manage per-service directories below
 4742           /var/lib, /var/cache, /var/log and /etc. By making use of them it is
 4743           possible to write unit files which when activated automatically gain
 4744           properly owned service specific directories in these locations, thus
 4745           making unit files self-contained and increasing compatibility with
 4746           stateless systems and factory reset where /etc or /var are
 4747           unpopulated at boot. Matching these new settings there's also
 4748           StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
 4749           ConfigurationDirectoryMode= for configuring the access mode of these
 4750           directories. These settings are particularly useful in combination
 4751           with DynamicUser=yes as they provide secure, properly-owned,
 4752           writable, and stateful locations for storage, excluded from the
 4753           sandbox that such services live in otherwise.
 4755         * Automake support has been removed from this release. systemd is now
 4756           Meson-only.
 4758         * systemd-journald will now aggressively cache client metadata during
 4759           runtime, speeding up log write performance under pressure. This comes
 4760           at a small price though: as much of the metadata is read
 4761           asynchronously from /proc/ (and isn't implicitly attached to log
 4762           datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
 4763           metadata stored alongside a log entry might be slightly
 4764           out-of-date. Previously it could only be slightly newer than the log
 4765           message. The time window is small however, and given that the kernel
 4766           is unlikely to be improved anytime soon in this regard, this appears
 4767           acceptable to us.
 4769         * nss-myhostname/systemd-resolved will now by default synthesize an
 4770           A/AAAA resource record for the "_gateway" hostname, pointing to the
 4771           current default IP gateway. Previously it did that for the "gateway"
 4772           name, hampering adoption, as some distributions wanted to leave that
 4773           hostname open for local use. The old behaviour may still be
 4774           requested at build time.
 4776         * systemd-networkd's [Address] section in .network files gained a new
 4777           Scope= setting for configuring the IP address scope. The [Network]
 4778           section gained a new boolean setting ConfigureWithoutCarrier= that
 4779           tells systemd-networkd to ignore link sensing when configuring the
 4780           device. The [DHCP] section gained a new Anonymize= boolean option for
 4781           turning on a number of options suggested in RFC 7844. A new
 4782           [RoutingPolicyRule] section has been added for configuring the IP
 4783           routing policy. The [Route] section has gained support for a new
 4784           Type= setting which permits configuring
 4785           blackhole/unreachable/prohibit routes.
 4787         * The [VRF] section in .netdev files gained a new Table= setting for
 4788           configuring the routing table to use. The [Tunnel] section gained a
 4789           new Independent= boolean field for configuring tunnels independent of
 4790           an underlying network interface. The [Bridge] section gained a new
 4791           GroupForwardMask= option for configuration of propagation of link
 4792           local frames between bridge ports.
 4794         * The WakeOnLan= setting in .link files gained support for a number of
 4795           new modes. A new TCP6SegmentationOffload= setting has been added for
 4796           configuring TCP/IPv6 hardware segmentation offload.
 4798         * The IPv6 RA sender implementation may now optionally send out RDNSS
 4799           and RDNSSL records to supply DNS configuration to peers.
 4801         * systemd-nspawn gained support for a new --system-call-filter= command
 4802           line option for adding and removing entries in the default system
 4803           call filter it applies. Moreover systemd-nspawn has been changed to
 4804           implement a system call allow list instead of a deny list.
 4806         * systemd-run gained support for a new --pipe command line option. If
 4807           used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
 4808           are directly passed on to the activated transient service
 4809           executable. This allows invoking arbitrary processes as systemd
 4810           services (for example to take benefit of dependency management,
 4811           accounting management, resource management or log management that is
 4812           done automatically for services) — while still allowing them to be
 4813           integrated in a classic UNIX shell pipeline.
 4815         * When a service sends RELOAD=1 via sd_notify() and reload propagation
 4816           using ReloadPropagationTo= is configured, a reload is now propagated
 4817           to configured units. (Previously this was only done on explicitly
 4818           requested reloads, using "systemctl reload" or an equivalent
 4819           command.)
 4821         * For each service unit a restart counter is now kept: it is increased
 4822           each time the service is restarted due to Restart=, and may be
 4823           queried using "systemctl show -p NRestarts …".
 4825         * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
 4826           @signal and @timer have been added, for usage with SystemCallFilter=
 4827           in unit files and the new --system-call-filter= command line option
 4828           of systemd-nspawn (see above).
 4830         * ExecStart= lines in unit files gained two new modifiers: when a
 4831           command line is prefixed with "!" the command will be executed as
 4832           configured, except for the credentials applied by
 4833           setuid()/setgid()/setgroups(). It is very similar to the pre-existing
 4834           "+", but does still apply namespacing options unlike "+". There's
 4835           also "!!" now, which is mostly identical, but becomes a NOP on
 4836           systems that support ambient capabilities. This is useful to write
 4837           unit files that work with ambient capabilities where possible but
 4838           automatically fall back to traditional privilege dropping mechanisms
 4839           on systems where this is not supported.
 4841         * ListenNetlink= settings in socket units now support RDMA netlink
 4842           sockets.
 4844         * A new unit file setting LockPersonality= has been added which permits
 4845           locking down the chosen execution domain ("personality") of a service
 4846           during runtime.
 4848         * A new special target "getty-pre.target" has been added, which is
 4849           ordered before all text logins, and may be used to order services
 4850           before textual logins acquire access to the console.
 4852         * systemd will now attempt to load the virtio-rng.ko kernel module very
 4853           early on if a VM environment supporting this is detected. This should
 4854           improve entropy during early boot in virtualized environments.
 4856         * A _netdev option is now supported in /etc/crypttab that operates in a
 4857           similar way as the same option in /etc/fstab: it permits configuring
 4858           encrypted devices that need to be ordered after the network is up.
 4859           Following this logic, two new special targets
 4860           remote-cryptsetup-pre.target and remote-cryptsetup.target have been
 4861           added that are to cryptsetup.target what remote-fs.target and
 4862           remote-fs-pre.target are to local-fs.target.
 4864         * Service units gained a new UnsetEnvironment= setting which permits
 4865           unsetting specific environment variables for services that are
 4866           normally passed to it (for example in order to mask out locale
 4867           settings for specific services that can't deal with it).
 4869         * Units acquired a new boolean option IPAccounting=. When turned on, IP
 4870           traffic accounting (packet count as well as byte count) is done for
 4871           the service, and shown as part of "systemctl status" or "systemd-run
 4872           --wait".
 4874         * Service units acquired two new options IPAddressAllow= and
 4875           IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
 4876           for configuring a simple IP access control list for all sockets of
 4877           the unit. These options are available also on .slice and .socket
 4878           units, permitting flexible access list configuration for individual
 4879           services as well as groups of services (as defined by a slice unit),
 4880           including system-wide. Note that IP ACLs configured this way are
 4881           enforced on every single IPv4 and IPv6 socket created by any process
 4882           of the service unit, and apply to ingress as well as egress traffic.
 4884         * If CPUAccounting= or IPAccounting= is turned on for a unit a new
 4885           structured log message is generated each time the unit is stopped,
 4886           containing information about the consumed resources of this
 4887           invocation.
 4889         * A new setting KeyringMode= has been added to unit files, which may be
 4890           used to control how the kernel keyring is set up for executed
 4891           processes.
 4893         * "systemctl poweroff", "systemctl reboot", "systemctl halt",
 4894           "systemctl kexec" and "systemctl exit" are now always asynchronous in
 4895           behaviour (that is: these commands return immediately after the
 4896           operation was enqueued instead of waiting for the operation to
 4897           complete). Previously, "systemctl poweroff" and "systemctl reboot"
 4898           were asynchronous on systems using systemd-logind (i.e. almost
 4899           always, and like they were on sysvinit), and the other three commands
 4900           were unconditionally synchronous. With this release this is cleaned
 4901           up, and callers will see the same asynchronous behaviour on all
 4902           systems for all five operations.
 4904         * systemd-logind gained new Halt() and CanHalt() bus calls for halting
 4905           the system.
 4907         * .timer units now accept calendar specifications in other timezones
 4908           than UTC or the local timezone.
 4910         * The tmpfiles snippet var.conf has been changed to create
 4911           /var/log/btmp with access mode 0660 instead of 0600. It was owned by
 4912           the "utmp" group already, and it appears to be generally understood
 4913           that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
 4914           databases. Previously this was implemented correctly for all these
 4915           databases excepts btmp, which has been opened up like this now
 4916           too. Note that while the other databases are world-readable
 4917           (i.e. 0644), btmp is not and remains more restrictive.
 4919         * The systemd-resolve tool gained a new --reset-server-features
 4920           switch. When invoked like this systemd-resolved will forget
 4921           everything it learnt about the features supported by the configured
 4922           upstream DNS servers, and restarts the feature probing logic on the
 4923           next resolver look-up for them at the highest feature level
 4924           again.
 4926         * The status dump systemd-resolved sends to the logs upon receiving
 4927           SIGUSR1 now also includes information about all DNS servers it is
 4928           configured to use, and the features levels it probed for them.
 4930         Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
 4931         Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
 4932         Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
 4933         Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
 4934         Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
 4935         Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
 4936         ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
 4937         Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
 4938         Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
 4939         John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
 4940         Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
 4941         Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
 4942         Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
 4943         Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
 4944         Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
 4945         Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
 4946         Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
 4947         Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
 4948         Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
 4949         Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
 4951         — Berlin, 2017-10-06
 4953 CHANGES WITH 234:
 4955         * Meson is now supported as build system in addition to Automake. It is
 4956           our plan to remove Automake in one of our next releases, so that
 4957           Meson becomes our exclusive build system. Hence, please start using
 4958           the Meson build system in your downstream packaging. There's plenty
 4959           of documentation around how to use Meson, the extremely brief
 4960           summary:
 4962               ./autogen.sh && ./configure && make && sudo make install
 4964           becomes:
 4966               meson build && ninja -C build && sudo ninja -C build install
 4968         * Unit files gained support for a new JobRunningTimeoutUSec= setting,
 4969           which permits configuring a timeout on the time a job is
 4970           running. This is particularly useful for setting timeouts on jobs for
 4971           .device units.
 4973         * Unit files gained two new options ConditionUser= and ConditionGroup=
 4974           for conditionalizing units based on the identity of the user/group
 4975           running a systemd user instance.
 4977         * systemd-networkd now understands a new FlowLabel= setting in the
 4978           [VXLAN] section of .network files, as well as a Priority= in
 4979           [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
 4980           and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
 4981           gained support for configuration of GENEVE links, and IPv6 address
 4982           labels. The [Network] section gained the new IPv6ProxyNDP= setting.
 4984         * .link files now understand a new Port= setting.
 4986         * systemd-networkd's DHCP support gained support for DHCP option 119
 4987           (domain search list).
 4989         * systemd-networkd gained support for serving IPv6 address ranges using
 4990           the Router Advertisement protocol. The new .network configuration
 4991           section [IPv6Prefix] may be used to configure the ranges to
 4992           serve. This is implemented based on a new, minimal, native server
 4993           implementation of RA.
 4995         * journalctl's --output= switch gained support for a new parameter
 4996           "short-iso-precise" for a mode where timestamps are shown as precise
 4997           ISO date values.
 4999         * systemd-udevd's "net_id" builtin may now generate stable network
 5000           interface names from IBM PowerVM VIO devices as well as ACPI platform
 5001           devices.
 5003         * MulticastDNS support in systemd-resolved may now be explicitly
 5004           enabled/disabled using the new MulticastDNS= configuration file
 5005           option.
 5007         * systemd-resolved may now optionally use libidn2 instead of the libidn
 5008           for processing internationalized domain names. Support for libidn2
 5009           should be considered experimental and should not be enabled by
 5010           default yet.
 5012         * "machinectl pull-tar" and related call may now do verification of
 5013           downloaded images using SUSE-style .sha256 checksum files in addition
 5014           to the already existing support for validating using Ubuntu-style
 5015           SHA256SUMS files.
 5017         * sd-bus gained support for a new sd_bus_message_appendv() call which
 5018           is va_list equivalent of sd_bus_message_append().
 5020         * sd-boot gained support for validating images using SHIM/MOK.
 5022         * The SMACK code learnt support for "onlycap".
 5024         * systemd-mount --umount is now much smarter in figuring out how to
 5025           properly unmount a device given its mount or device path.
 5027         * The code to call libnss_dns as a fallback from libnss_resolve when
 5028           the communication with systemd-resolved fails was removed. This
 5029           fallback was redundant and interfered with the [!UNAVAIL=return]
 5030           suffix. See nss-resolve(8) for the recommended configuration.
 5032         * systemd-logind may now be restarted without losing state. It stores
 5033           the file descriptors for devices it manages in the system manager
 5034           using the FDSTORE= mechanism. Please note that further changes in
 5035           other components may be required to make use of this (for example
 5036           Xorg has code to listen for stops of systemd-logind and terminate
 5037           itself when logind is stopped or restarted, in order to avoid using
 5038           stale file descriptors for graphical devices, which is now
 5039           counterproductive and must be reverted in order for restarts of
 5040           systemd-logind to be safe. See
 5041           https://cgit.freedesktop.org/xorg/xserver/commit/?id=dc48bd653c7e101.)
 5043         * All kernel-install plugins are called with the environment variable
 5044           KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
 5045           /etc/machine-id. If the machine ID could not be determined,
 5046           $KERNEL_INSTALL_MACHINE_ID will be empty. Plugins should not put
 5047           anything in the entry directory (passed as the second argument) if
 5048           $KERNEL_INSTALL_MACHINE_ID is empty. For backwards compatibility, a
 5049           temporary directory is passed as the entry directory and removed
 5050           after all the plugins exit.
 5052         Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
 5053         Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
 5054         Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
 5055         Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
 5056         Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
 5057         Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
 5058         Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
 5059         Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
 5060         Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
 5061         Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
 5062         hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
 5063         Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
 5064         Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
 5065         Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
 5066         Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
 5067         Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
 5068         Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
 5069         Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
 5070         Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
 5071         Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
 5072         Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
 5073         Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
 5074         Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
 5075         Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
 5076         Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
 5077         H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
 5078         Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
 5079         userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
 5080         Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
 5081         Георгиевски
 5083         — Berlin, 2017-07-12
 5085 CHANGES WITH 233:
 5087         * The "hybrid" control group mode has been modified to improve
 5088           compatibility with "legacy" cgroups-v1 setups. Specifically, the
 5089           "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
 5090           "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
 5091           cgroups-v1 hierarchy), the only externally visible change being that
 5092           the cgroups-v2 hierarchy is also mounted, to
 5093           /sys/fs/cgroup/unified. This should provide a large degree of
 5094           compatibility with "legacy" cgroups-v1, while taking benefit of the
 5095           better management capabilities of cgroups-v2.
 5097         * The default control group setup mode may be selected both a boot-time
 5098           via a set of kernel command line parameters (specifically:
 5099           systemd.unified_cgroup_hierarchy= and
 5100           systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
 5101           default selected on the configure command line
 5102           (--with-default-hierarchy=). The upstream default is "hybrid"
 5103           (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
 5104           this will change in a future systemd version to be "unified" (pure
 5105           cgroups-v2 mode). The third option for the compile time option is
 5106           "legacy", to enter pure cgroups-v1 mode. We recommend downstream
 5107           distributions to default to "hybrid" mode for release distributions,
 5108           starting with v233. We recommend "unified" for development
 5109           distributions (specifically: distributions such as Fedora's rawhide)
 5110           as that's where things are headed in the long run. Use "legacy" for
 5111           greatest stability and compatibility only.
 5113         * Note one current limitation of "unified" and "hybrid" control group
 5114           setup modes: the kernel currently does not permit the systemd --user
 5115           instance (i.e. unprivileged code) to migrate processes between two
 5116           disconnected cgroup subtrees, even if both are managed and owned by
 5117           the user. This effectively means "systemd-run --user --scope" doesn't
 5118           work when invoked from outside of any "systemd --user" service or
 5119           scope. Specifically, it is not supported from session scopes. We are
 5120           working on fixing this in a future systemd version. (See #3388 for
 5121           further details about this.)
 5123         * DBus policy files are now installed into /usr rather than /etc. Make
 5124           sure your system has dbus >= 1.9.18 running before upgrading to this
 5125           version, or override the install path with --with-dbuspolicydir= .
 5127         * All python scripts shipped with systemd (specifically: the various
 5128           tests written in Python) now require Python 3.
 5130         * systemd unit tests can now run standalone (without the source or
 5131           build directories), and can be installed into /usr/lib/systemd/tests/
 5132           with 'make install-tests'.
 5134         * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
 5135           CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
 5136           kernel.
 5138         * Support for the %c, %r, %R specifiers in unit files has been
 5139           removed. Specifiers are not supposed to be dependent on configuration
 5140           in the unit file itself (so that they resolve the same regardless
 5141           where used in the unit files), but these specifiers were influenced
 5142           by the Slice= option.
 5144         * The shell invoked by debug-shell.service now defaults to /bin/sh in
 5145           all cases. If distributions want to use a different shell for this
 5146           purpose (for example Fedora's /sbin/sushell) they need to specify
 5147           this explicitly at configure time using --with-debug-shell=.
 5149         * The confirmation spawn prompt has been reworked to offer the
 5150           following choices:
 5152            (c)ontinue, proceed without asking anymore
 5153            (D)ump, show the state of the unit
 5154            (f)ail, don't execute the command and pretend it failed
 5155            (h)elp
 5156            (i)nfo, show a short summary of the unit
 5157            (j)obs, show jobs that are in progress
 5158            (s)kip, don't execute the command and pretend it succeeded
 5159            (y)es, execute the command
 5161           The 'n' choice for the confirmation spawn prompt has been removed,
 5162           because its meaning was confusing.
 5164           The prompt may now also be redirected to an alternative console by
 5165           specifying the console as parameter to systemd.confirm_spawn=.
 5167         * Services of Type=notify require a READY=1 notification to be sent
 5168           during startup. If no such message is sent, the service now fails,
 5169           even if the main process exited with a successful exit code.
 5171         * Services that fail to start up correctly now always have their
 5172           ExecStopPost= commands executed. Previously, they'd enter "failed"
 5173           state directly, without executing these commands.
 5175         * The option MulticastDNS= of network configuration files has acquired
 5176           an actual implementation. With MulticastDNS=yes a host can resolve
 5177           names of remote hosts and reply to mDNS A and AAAA requests.
 5179         * When units are about to be started an additional check is now done to
 5180           ensure that all dependencies of type BindsTo= (when used in
 5181           combination with After=) have been started.
 5183         * systemd-analyze gained a new verb "syscall-filter" which shows which
 5184           system call groups are defined for the SystemCallFilter= unit file
 5185           setting, and which system calls they contain.
 5187         * A new system call filter group "@filesystem" has been added,
 5188           consisting of various file system related system calls. Group
 5189           "@reboot" has been added, covering reboot, kexec and shutdown related
 5190           calls. Finally, group "@swap" has been added covering swap
 5191           configuration related calls.
 5193         * A new unit file option RestrictNamespaces= has been added that may be
 5194           used to restrict access to the various process namespace types the
 5195           Linux kernel provides. Specifically, it may be used to take away the
 5196           right for a service unit to create additional file system, network,
 5197           user, and other namespaces. This sandboxing option is particularly
 5198           relevant due to the high amount of recently discovered namespacing
 5199           related vulnerabilities in the kernel.
 5201         * systemd-udev's .link files gained support for a new AutoNegotiation=
 5202           setting for configuring Ethernet auto-negotiation.
 5204         * systemd-networkd's .network files gained support for a new
 5205           ListenPort= setting in the [DHCP] section to explicitly configure the
 5206           UDP client port the DHCP client shall listen on.
 5208         * .network files gained a new Unmanaged= boolean setting for explicitly
 5209           excluding one or more interfaces from management by systemd-networkd.
 5211         * The systemd-networkd ProxyARP= option has been renamed to
 5212           IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
 5213           renamed to ReduceARPProxy=. The old names continue to be available
 5214           for compatibility.
 5216         * systemd-networkd gained support for configuring IPv6 Proxy NDP
 5217           addresses via the new IPv6ProxyNDPAddress= .network file setting.
 5219         * systemd-networkd's bonding device support gained support for two new
 5220           configuration options ActiveSlave= and PrimarySlave=.
 5222         * The various options in the [Match] section of .network files gained
 5223           support for negative matching.
 5225         * New systemd-specific mount options are now understood in /etc/fstab:
 5227           x-systemd.mount-timeout= may be used to configure the maximum
 5228           permitted runtime of the mount command.
 5230           x-systemd.device-bound may be set to bind a mount point to its
 5231           backing device unit, in order to automatically remove a mount point
 5232           if its backing device is unplugged. This option may also be
 5233           configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
 5234           on the block device, which is now automatically set for all CDROM
 5235           drives, so that mounted CDs are automatically unmounted when they are
 5236           removed from the drive.
 5238           x-systemd.after= and x-systemd.before= may be used to explicitly
 5239           order a mount after or before another unit or mount point.
 5241         * Enqueued start jobs for device units are now automatically garbage
 5242           collected if there are no jobs waiting for them anymore.
 5244         * systemctl list-jobs gained two new switches: with --after, for every
 5245           queued job the jobs it's waiting for are shown; with --before the
 5246           jobs which it's blocking are shown.
 5248         * systemd-nspawn gained support for ephemeral boots from disk images
 5249           (or in other words: --ephemeral and --image= may now be
 5250           combined). Moreover, ephemeral boots are now supported for normal
 5251           directories, even if the backing file system is not btrfs. Of course,
 5252           if the file system does not support file system snapshots or
 5253           reflinks, the initial copy operation will be relatively expensive, but
 5254           this should still be suitable for many use cases.
 5256         * Calendar time specifications in .timer units now support
 5257           specifications relative to the end of a month by using "~" instead of
 5258           "-" as separator between month and day. For example, "*-02~03" means
 5259           "the third last day in February". In addition a new syntax for
 5260           repeated events has been added using the "/" character. For example,
 5261           "9..17/2:00" means "every two hours from 9am to 5pm".
 5263         * systemd-socket-proxyd gained a new parameter --connections-max= for
 5264           configuring the maximum number of concurrent connections.
 5266         * sd-id128 gained a new API for generating unique IDs for the host in a
 5267           way that does not leak the machine ID. Specifically,
 5268           sd_id128_get_machine_app_specific() derives an ID based on the
 5269           machine ID a in well-defined, non-reversible, stable way. This is
 5270           useful whenever an identifier for the host is needed but where the
 5271           identifier shall not be useful to identify the system beyond the
 5272           scope of the application itself. (Internally this uses HMAC-SHA256 as
 5273           keyed hash function using the machine ID as input.)
 5275         * NotifyAccess= gained a new supported value "exec". When set
 5276           notifications are accepted from all processes systemd itself invoked,
 5277           including all control processes.
 5279         * .nspawn files gained support for defining overlay mounts using the
 5280           Overlay= and OverlayReadOnly= options. Previously this functionality
 5281           was only available on the systemd-nspawn command line.
 5283         * systemd-nspawn's --bind= and --overlay= options gained support for
 5284           bind/overlay mounts whose source lies within the container tree by
 5285           prefixing the source path with "+".
 5287         * systemd-nspawn's --bind= and --overlay= options gained support for
 5288           automatically allocating a temporary source directory in /var/tmp
 5289           that is removed when the container dies. Specifically, if the source
 5290           directory is specified as empty string this mechanism is selected. An
 5291           example usage is --overlay=+/var::/var, which creates an overlay
 5292           mount based on the original /var contained in the image, overlaid
 5293           with a temporary directory in the host's /var/tmp. This way changes
 5294           to /var are automatically flushed when the container shuts down.
 5296         * systemd-nspawn --image= option does now permit raw file system block
 5297           devices (in addition to images containing partition tables, as
 5298           before).
 5300         * The disk image dissection logic in systemd-nspawn gained support for
 5301           automatically setting up LUKS encrypted as well as Verity protected
 5302           partitions. When a container is booted from an encrypted image the
 5303           passphrase is queried at start-up time. When a container with Verity
 5304           data is started, the root hash is search in a ".roothash" file
 5305           accompanying the disk image (alternatively, pass the root hash via
 5306           the new --root-hash= command line option).
 5308         * A new tool /usr/lib/systemd/systemd-dissect has been added that may
 5309           be used to dissect disk images the same way as systemd-nspawn does
 5310           it, following the Bootable Partition Specification. It may even be
 5311           used to mount disk images with complex partition setups (including
 5312           LUKS and Verity partitions) to a local host directory, in order to
 5313           inspect them. This tool is not considered public API (yet), and is
 5314           thus not installed into /usr/bin. Please do not rely on its
 5315           existence, since it might go away or be changed in later systemd
 5316           versions.
 5318         * A new generator "systemd-verity-generator" has been added, similar in
 5319           style to "systemd-cryptsetup-generator", permitting automatic setup of
 5320           Verity root partitions when systemd boots up. In order to make use of
 5321           this your partition setup should follow the Discoverable Partitions
 5322           Specification, and the GPT partition ID of the root file system
 5323           partition should be identical to the upper 128bit of the Verity root
 5324           hash. The GPT partition ID of the Verity partition protecting it
 5325           should be the lower 128bit of the Verity root hash. If the partition
 5326           image follows this model it is sufficient to specify a single
 5327           "roothash=" kernel command line argument to both configure which root
 5328           image and verity partition to use as well as the root hash for
 5329           it. Note that systemd-nspawn's Verity support follows the same
 5330           semantics, meaning that disk images with proper Verity data in place
 5331           may be booted in containers with systemd-nspawn as well as on
 5332           physical systems via the verity generator. Also note that the "mkosi"
 5333           tool available at https://github.com/systemd/mkosi has been updated
 5334           to generate Verity protected disk images following this scheme. In
 5335           fact, it has been updated to generate disk images that optionally
 5336           implement a complete UEFI SecureBoot trust chain, involving a signed
 5337           kernel and initrd image that incorporates such a root hash as well as
 5338           a Verity-enabled root partition.
 5340         * The hardware database (hwdb) udev supports has been updated to carry
 5341           accelerometer quirks.
 5343         * All system services are now run with a fresh kernel keyring set up
 5344           for them. The invocation ID is stored by default in it, thus
 5345           providing a safe, non-overridable way to determine the invocation
 5346           ID of each service.
 5348         * Service unit files gained new BindPaths= and BindReadOnlyPaths=
 5349           options for bind mounting arbitrary paths in a service-specific
 5350           way. When these options are used, arbitrary host or service files and
 5351           directories may be mounted to arbitrary locations in the service's
 5352           view.
 5354         * Documentation has been added that lists all of systemd's low-level
 5355           environment variables:
 5357           https://github.com/systemd/systemd/blob/master/docs/ENVIRONMENT.md
 5359         * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
 5360           whether a specific socket file descriptor matches a specified socket
 5361           address.
 5363         * systemd-firstboot has been updated to check for the
 5364           systemd.firstboot= kernel command line option. It accepts a boolean
 5365           and when set to false the first boot questions are skipped.
 5367         * systemd-fstab-generator has been updated to check for the
 5368           systemd.volatile= kernel command line option, which either takes an
 5369           optional boolean parameter or the special value "state". If used the
 5370           system may be booted in a "volatile" boot mode. Specifically,
 5371           "systemd.volatile" is used, the root directory will be mounted as
 5372           tmpfs, and only /usr is mounted from the actual root file system. If
 5373           "systemd.volatile=state" is used, the root directory will be mounted
 5374           as usual, but /var is mounted as tmpfs. This concept provides similar
 5375           functionality as systemd-nspawn's --volatile= option, but provides it
 5376           on physical boots. Use this option for implementing stateless
 5377           systems, or testing systems with all state and/or configuration reset
 5378           to the defaults. (Note though that many distributions are not
 5379           prepared to boot up without a populated /etc or /var, though.)
 5381         * systemd-gpt-auto-generator gained support for LUKS encrypted root
 5382           partitions. Previously it only supported LUKS encrypted partitions
 5383           for all other uses, except for the root partition itself.
 5385         * Socket units gained support for listening on AF_VSOCK sockets for
 5386           communication in virtualized QEMU environments.
 5388         * The "configure" script gained a new option --with-fallback-hostname=
 5389           for specifying the fallback hostname to use if none is configured in
 5390           /etc/hostname. For example, by specifying
 5391           --with-fallback-hostname=fedora it is possible to default to a
 5392           hostname of "fedora" on pristine installations.
 5394         * systemd-cgls gained support for a new --unit= switch for listing only
 5395           the control groups of a specific unit. Similar --user-unit= has been
 5396           added for listing only the control groups of a specific user unit.
 5398         * systemd-mount gained a new --umount switch for unmounting a mount or
 5399           automount point (and all mount/automount points below it).
 5401         * systemd will now refuse full configuration reloads (via systemctl
 5402           daemon-reload and related calls) unless at least 16MiB of free space
 5403           are available in /run. This is a safety precaution in order to ensure
 5404           that generators can safely operate after the reload completed.
 5406         * A new unit file option RootImage= has been added, which has a similar
 5407           effect as RootDirectory= but mounts the service's root directory from
 5408           a disk image instead of plain directory. This logic reuses the same
 5409           image dissection and mount logic that systemd-nspawn already uses,
 5410           and hence supports any disk images systemd-nspawn supports, including
 5411           those following the Discoverable Partition Specification, as well as
 5412           Verity enabled images. This option enables systemd to run system
 5413           services directly off disk images acting as resource bundles,
 5414           possibly even including full integrity data.
 5416         * A new MountAPIVFS= unit file option has been added, taking a boolean
 5417           argument. If enabled /proc, /sys and /dev (collectively called the
 5418           "API VFS") will be mounted for the service. This is only relevant if
 5419           RootDirectory= or RootImage= is used for the service, as these mounts
 5420           are of course in place in the host mount namespace anyway.
 5422         * systemd-nspawn gained support for a new --pivot-root= switch. If
 5423           specified the root directory within the container image is pivoted to
 5424           the specified mount point, while the original root disk is moved to a
 5425           different place. This option enables booting of ostree images
 5426           directly with systemd-nspawn.
 5428         * The systemd build scripts will no longer complain if the NTP server
 5429           addresses are not changed from the defaults. Google now supports
 5430           these NTP servers officially. We still recommend downstreams to
 5431           properly register an NTP pool with the NTP pool project though.
 5433         * coredumpctl gained a new "--reverse" option for printing the list
 5434           of coredumps in reverse order.
 5436         * coredumpctl will now show additional information about truncated and
 5437           inaccessible coredumps, as well as coredumps that are still being
 5438           processed. It also gained a new --quiet switch for suppressing
 5439           additional informational message in its output.
 5441         * coredumpctl gained support for only showing coredumps newer and/or
 5442           older than specific timestamps, using the new --since= and --until=
 5443           options, reminiscent of journalctl's options by the same name.
 5445         * The systemd-coredump logic has been improved so that it may be reused
 5446           to collect backtraces in non-compiled languages, for example in
 5447           scripting languages such as Python.
 5449         * machinectl will now show the UID shift of local containers, if user
 5450           namespacing is enabled for them.
 5452         * systemd will now optionally run "environment generator" binaries at
 5453           configuration load time. They may be used to add environment
 5454           variables to the environment block passed to services invoked. One
 5455           user environment generator is shipped by default that sets up
 5456           environment variables based on files dropped into /etc/environment.d
 5457           and ~/.config/environment.d/.
 5459         * systemd-resolved now includes the new, recently published 2017 DNSSEC
 5460           root key (KSK).
 5462         * hostnamed has been updated to report a new chassis type of
 5463           "convertible" to cover "foldable" laptops that can both act as a
 5464           tablet and as a laptop, such as various Lenovo Yoga devices.
 5466         Contributions from: Adrián López, Alexander Galanin, Alexander
 5467         Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
 5468         Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
 5469         Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
 5470         Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
 5471         David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
 5472         Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
 5473         Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
 5474         Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
 5475         Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
 5476         Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
 5477         Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
 5478         Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
 5479         Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
 5480         Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
 5481         Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
 5482         Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
 5483         Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
 5484         Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
 5485         Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
 5486         Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
 5487         Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
 5488         Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
 5489         Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
 5490         Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
 5491         Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
 5492         Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
 5493         YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
 5494         Тихонов
 5496         — Berlin, 2017-03-01
 5498 CHANGES WITH 232:
 5500         * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
 5501           RestrictAddressFamilies= enabled. These sandboxing options should
 5502           generally be compatible with the various external udev call-out
 5503           binaries we are aware of, however there may be exceptions, in
 5504           particular when exotic languages for these call-outs are used. In
 5505           this case, consider turning off these settings locally.
 5507         * The new RemoveIPC= option can be used to remove IPC objects owned by
 5508           the user or group of a service when that service exits.
 5510         * The new ProtectKernelModules= option can be used to disable explicit
 5511           load and unload operations of kernel modules by a service. In
 5512           addition access to /usr/lib/modules is removed if this option is set.
 5514         * ProtectSystem= option gained a new value "strict", which causes the
 5515           whole file system tree with the exception of /dev, /proc, and /sys,
 5516           to be remounted read-only for a service.
 5518         * The new ProtectKernelTunables= option can be used to disable
 5519           modification of configuration files in /sys and /proc by a service.
 5520           Various directories and files are remounted read-only, so access is
 5521           restricted even if the file permissions would allow it.
 5523         * The new ProtectControlGroups= option can be used to disable write
 5524           access by a service to /sys/fs/cgroup.
 5526         * Various systemd services have been hardened with
 5527           ProtectKernelTunables=yes, ProtectControlGroups=yes,
 5528           RestrictAddressFamilies=.
 5530         * Support for dynamically creating users for the lifetime of a service
 5531           has been added. If DynamicUser=yes is specified, user and group IDs
 5532           will be allocated from the range 61184…65519 for the lifetime of the
 5533           service. They can be resolved using the new nss-systemd.so NSS
 5534           module. The module must be enabled in /etc/nsswitch.conf. Services
 5535           started in this way have PrivateTmp= and RemoveIPC= enabled, so that
 5536           any resources allocated by the service will be cleaned up when the
 5537           service exits. They also have ProtectHome=read-only and
 5538           ProtectSystem=strict enabled, so they are not able to make any
 5539           permanent modifications to the system.
 5541         * The nss-systemd module also always resolves root and nobody, making
 5542           it possible to have no /etc/passwd or /etc/group files in minimal
 5543           container or chroot environments.
 5545         * Services may be started with their own user namespace using the new
 5546           boolean PrivateUsers= option. Only root, nobody, and the uid/gid
 5547           under which the service is running are mapped. All other users are
 5548           mapped to nobody.
 5550         * Support for the cgroup namespace has been added to systemd-nspawn. If
 5551           supported by kernel, the container system started by systemd-nspawn
 5552           will have its own view of the cgroup hierarchy. This new behaviour
 5553           can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
 5555         * The new MemorySwapMax= option can be used to limit the maximum swap
 5556           usage under the unified cgroup hierarchy.
 5558         * Support for the CPU controller in the unified cgroup hierarchy has
 5559           been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
 5560           options. This controller requires out-of-tree patches for the kernel
 5561           and the support is provisional.
 5563         * Mount and automount units may now be created transiently
 5564           (i.e. dynamically at runtime via the bus API, instead of requiring
 5565           unit files in the file system).
 5567         * systemd-mount is a new tool which may mount file systems – much like
 5568           mount(8), optionally pulling in additional dependencies through
 5569           transient .mount and .automount units. For example, this tool
 5570           automatically runs fsck on a backing block device before mounting,
 5571           and allows the automount logic to be used dynamically from the
 5572           command line for establishing mount points. This tool is particularly
 5573           useful when dealing with removable media, as it will ensure fsck is
 5574           run – if necessary – before the first access and that the file system
 5575           is quickly unmounted after each access by utilizing the automount
 5576           logic. This maximizes the chance that the file system on the
 5577           removable media stays in a clean state, and if it isn't in a clean
 5578           state is fixed automatically.
 5580         * LazyUnmount=yes option for mount units has been added to expose the
 5581           umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
 5582           option.
 5584         * /efi will be used as the mount point of the EFI boot partition, if
 5585           the directory is present, and the mount point was not configured
 5586           through other means (e.g. fstab). If /efi directory does not exist,
 5587           /boot will be used as before. This makes it easier to automatically
 5588           mount the EFI partition on systems where /boot is used for something
 5589           else.
 5591         * When operating on GPT disk images for containers, systemd-nspawn will
 5592           now mount the ESP to /boot or /efi according to the same rules as PID
 5593           1 running on a host. This allows tools like "bootctl" to operate
 5594           correctly within such containers, in order to make container images
 5595           bootable on physical systems.
 5597         * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
 5599         * Two new user session targets have been added to support running
 5600           graphical sessions under the systemd --user instance:
 5601           graphical-session.target and graphical-session-pre.target. See
 5602           systemd.special(7) for a description of how those targets should be
 5603           used.
 5605         * The vconsole initialization code has been significantly reworked to
 5606           use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
 5607           support unicode keymaps. Font and keymap configuration will now be
 5608           copied to all allocated virtual consoles.
 5610         * FreeBSD's bhyve virtualization is now detected.
 5612         * Information recorded in the journal for core dumps now includes the
 5613           contents of /proc/mountinfo and the command line of the process at
 5614           the top of the process hierarchy (which is usually the init process
 5615           of the container).
 5617         * systemd-journal-gatewayd learned the --directory= option to serve
 5618           files from the specified location.
 5620         * journalctl --root=… can be used to peruse the journal in the
 5621           /var/log/ directories inside of a container tree. This is similar to
 5622           the existing --machine= option, but does not require the container to
 5623           be active.
 5625         * The hardware database has been extended to support
 5626           ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
 5627           trackball devices.
 5629           MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
 5630           specify the click rate for mice which include a horizontal wheel with
 5631           a click rate that is different than the one for the vertical wheel.
 5633         * systemd-run gained a new --wait option that makes service execution
 5634           synchronous. (Specifically, the command will not return until the
 5635           specified service binary exited.)
 5637         * systemctl gained a new --wait option that causes the start command to
 5638           wait until the units being started have terminated again.
 5640         * A new journal output mode "short-full" has been added which displays
 5641           timestamps with abbreviated English day names and adds a timezone
 5642           suffix. Those timestamps include more information than the default
 5643           "short" output mode, and can be passed directly to journalctl's
 5644           --since= and --until= options.
 5646         * /etc/resolv.conf will be bind-mounted into containers started by
 5647           systemd-nspawn, if possible, so any changes to resolv.conf contents
 5648           are automatically propagated to the container.
 5650         * The number of instances for socket-activated services originating
 5651           from a single IP address can be limited with
 5652           MaxConnectionsPerSource=, extending the existing setting of
 5653           MaxConnections=.
 5655         * systemd-networkd gained support for vcan ("Virtual CAN") interface
 5656           configuration.
 5658         * .netdev and .network configuration can now be extended through
 5659           drop-ins.
 5661         * UDP Segmentation Offload, TCP Segmentation Offload, Generic
 5662           Segmentation Offload, Generic Receive Offload, Large Receive Offload
 5663           can be enabled and disabled using the new UDPSegmentationOffload=,
 5664           TCPSegmentationOffload=, GenericSegmentationOffload=,
 5665           GenericReceiveOffload=, LargeReceiveOffload= options in the
 5666           [Link] section of .link files.
 5668         * The Spanning Tree Protocol, Priority, Aging Time, and the Default
 5669           Port VLAN ID can be configured for bridge devices using the new STP=,
 5670           Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
 5671           section of .netdev files.
 5673         * The route table to which routes received over DHCP or RA should be
 5674           added can be configured with the new RouteTable= option in the [DHCP]
 5675           and [IPv6AcceptRA] sections of .network files.
 5677         * The Address Resolution Protocol can be disabled on links managed by
 5678           systemd-networkd using the ARP=no setting in the [Link] section of
 5679           .network files.
 5681         * New environment variables $SERVICE_RESULT, $EXIT_CODE and
 5682           $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
 5683           encode information about the result and exit codes of the current
 5684           service runtime cycle.
 5686         * systemd-sysctl will now configure kernel parameters in the order
 5687           they occur in the configuration files. This matches what sysctl
 5688           has been traditionally doing.
 5690         * kernel-install "plugins" that are executed to perform various
 5691           tasks after a new kernel is added and before an old one is removed
 5692           can now return a special value to terminate the procedure and
 5693           prevent any later plugins from running.
 5695         * Journald's SplitMode=login setting has been deprecated. It has been
 5696           removed from documentation, and its use is discouraged. In a future
 5697           release it will be completely removed, and made equivalent to current
 5698           default of SplitMode=uid.
 5700         * Storage=both option setting in /etc/systemd/coredump.conf has been
 5701           removed. With fast LZ4 compression storing the core dump twice is not
 5702           useful.
 5704         * The --share-system systemd-nspawn option has been replaced with an
 5705           (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
 5706           this functionality is discouraged. In addition the variables
 5708           $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
 5709           individual namespaces.
 5711         * "machinectl list" now shows the IP address of running containers in
 5712           the output, as well as OS release information.
 5714         * "loginctl list" now shows the TTY of each session in the output.
 5716         * sd-bus gained new API calls sd_bus_track_set_recursive(),
 5717           sd_bus_track_get_recursive(), sd_bus_track_count_name(),
 5718           sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
 5719           tracking objects in a "recursive" mode, where a single client can be
 5720           counted multiple times, if it takes multiple references.
 5722         * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
 5723           sd_bus_get_exit_on_disconnect(). They may be used to make a
 5724           process using sd-bus automatically exit if the bus connection is
 5725           severed.
 5727         * Bus clients of the service manager may now "pin" loaded units into
 5728           memory, by taking an explicit reference on them. This is useful to
 5729           ensure the client can retrieve runtime data about the service even
 5730           after the service completed execution. Taking such a reference is
 5731           available only for privileged clients and should be helpful to watch
 5732           running services in a race-free manner, and in particular collect
 5733           information about exit statuses and results.
 5735         * The nss-resolve module has been changed to strictly return UNAVAIL
 5736           when communication via D-Bus with resolved failed, and NOTFOUND when
 5737           a lookup completed but was negative. This means it is now possible to
 5738           neatly configure fallbacks using nsswitch.conf result checking
 5739           expressions. Taking benefit of this, the new recommended
 5740           configuration line for the "hosts" entry in /etc/nsswitch.conf is:
 5742               hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
 5744         * A new setting CtrlAltDelBurstAction= has been added to
 5745           /etc/systemd/system.conf which may be used to configure the precise
 5746           behaviour if the user on the console presses Ctrl-Alt-Del more often
 5747           than 7 times in 2s. Previously this would unconditionally result in
 5748           an expedited, immediate reboot. With this new setting the precise
 5749           operation may be configured in more detail, and also turned off
 5750           entirely.
 5752         * In .netdev files two new settings RemoteChecksumTx= and
 5753           RemoteChecksumRx= are now understood that permit configuring the
 5754           remote checksumming logic for VXLAN networks.
 5756         * The service manager learnt a new "invocation ID" concept for invoked
 5757           services. Each runtime cycle of a service will get a new invocation
 5758           ID (a 128bit random UUID) assigned that identifies the current
 5759           run of the service uniquely and globally. A new invocation ID
 5760           is generated each time a service starts up. The journal will store
 5761           the invocation ID of a service along with any logged messages, thus
 5762           making the invocation ID useful for matching the online runtime of a
 5763           service with the offline log data it generated in a safe way without
 5764           relying on synchronized timestamps. In many ways this new service
 5765           invocation ID concept is similar to the kernel's boot ID concept that
 5766           uniquely and globally identifies the runtime of each boot. The
 5767           invocation ID of a service is passed to the service itself via an
 5768           environment variable ($INVOCATION_ID). A new bus call
 5769           GetUnitByInvocationID() has been added that is similar to GetUnit()
 5770           but instead of retrieving the bus path for a unit by its name
 5771           retrieves it by its invocation ID. The returned path is valid only as
 5772           long as the passed invocation ID is current.
 5774         * systemd-resolved gained a new "DNSStubListener" setting in
 5775           resolved.conf. It either takes a boolean value or the special values
 5776           "udp" and "tcp", and configures whether to enable the stub DNS
 5777           listener on
 5779         * IP addresses configured via networkd may now carry additional
 5780           configuration settings supported by the kernel. New options include:
 5781           HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
 5782           PrefixRoute=, AutoJoin=.
 5784         * The PAM configuration fragment file for "user@.service" shipped with
 5785           systemd (i.e. the --user instance of systemd) has been stripped to
 5786           the minimum necessary to make the system boot. Previously, it
 5787           contained Fedora-specific stanzas that did not apply to other
 5788           distributions. It is expected that downstream distributions add
 5789           additional configuration lines, matching their needs to this file,
 5790           using it only as rough template of what systemd itself needs. Note
 5791           that this reduced fragment does not even include an invocation of
 5792           pam_limits which most distributions probably want to add, even though
 5793           systemd itself does not need it. (There's also the new build time
 5794           option --with-pamconfdir=no to disable installation of the PAM
 5795           fragment entirely.)
 5797         * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
 5798           capability is now also dropped from its set (in addition to
 5799           CAP_SYS_MKNOD as before).
 5801         * In service unit files it is now possible to connect a specific named
 5802           file descriptor with stdin/stdout/stdout of an executed service. The
 5803           name may be specified in matching .socket units using the
 5804           FileDescriptorName= setting.
 5806         * A number of journal settings may now be configured on the kernel
 5807           command line. Specifically, the following options are now understood:
 5808           systemd.journald.max_level_console=,
 5809           systemd.journald.max_level_store=,
 5810           systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
 5811           systemd.journald.max_level_wall=.
 5813         * "systemctl is-enabled --full" will now show by which symlinks a unit
 5814           file is enabled in the unit dependency tree.
 5816         * Support for VeraCrypt encrypted partitions has been added to the
 5817           "cryptsetup" logic and /etc/crypttab.
 5819         * systemd-detect-virt gained support for a new --private-users switch
 5820           that checks whether the invoking processes are running inside a user
 5821           namespace. Similar, a new special value "private-users" for the
 5822           existing ConditionVirtualization= setting has been added, permitting
 5823           skipping of specific units in user namespace environments.
 5825         Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
 5826         Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
 5827         Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
 5828         Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
 5829         Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
 5830         Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
 5831         Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
 5832         Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
 5833         Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
 5834         Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
 5835         Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
 5836         Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
 5837         Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
 5838         Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
 5839         Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
 5840         Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
 5841         Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
 5842         Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
 5843         Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
 5844         Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
 5845         Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
 5846         Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
 5847         Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
 5848         Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
 5849         E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
 5850         Jędrzejewski-Szmek, Zeal Jagannatha
 5852         — Santa Fe, 2016-11-03
 5854 CHANGES WITH 231:
 5856         * In service units the various ExecXYZ= settings have been extended
 5857           with an additional special character as first argument of the
 5858           assigned value: if the character '+' is used the specified command
 5859           line it will be run with full privileges, regardless of User=,
 5860           Group=, CapabilityBoundingSet= and similar options. The effect is
 5861           similar to the existing PermissionsStartOnly= option, but allows
 5862           configuration of this concept for each executed command line
 5863           independently.
 5865         * Services may now alter the service watchdog timeout at runtime by
 5866           sending a WATCHDOG_USEC= message via sd_notify().
 5868         * MemoryLimit= and related unit settings now optionally take percentage
 5869           specifications. The percentage is taken relative to the amount of
 5870           physical memory in the system (or in case of containers, the assigned
 5871           amount of memory). This allows scaling service resources neatly with
 5872           the amount of RAM available on the system. Similarly, systemd-logind's
 5873           RuntimeDirectorySize= option now also optionally takes percentage
 5874           values.
 5876         * In similar fashion TasksMax= takes percentage values now, too. The
 5877           value is taken relative to the configured maximum number of processes
 5878           on the system. The per-service task maximum has been changed to 15%
 5879           using this functionality. (Effectively this is an increase of 512 →
 5880           4915 for service units, given the kernel's default pid_max setting.)
 5882         * Calendar time specifications in .timer units now understand a ".."
 5883           syntax for time ranges. Example: "4..7:10" may now be used for
 5884           defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
 5885           7:10am every day.
 5887         * The InaccessableDirectories=, ReadOnlyDirectories= and
 5888           ReadWriteDirectories= unit file settings have been renamed to
 5889           InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
 5890           applied to all kinds of file nodes, and not just directories, with
 5891           the exception of symlinks. Specifically these settings may now be
 5892           used on block and character device nodes, UNIX sockets and FIFOS as
 5893           well as regular files. The old names of these settings remain
 5894           available for compatibility.
 5896         * systemd will now log about all service processes it kills forcibly
 5897           (using SIGKILL) because they remained after the clean shutdown phase
 5898           of the service completed. This should help identifying services that
 5899           shut down uncleanly. Moreover if KillUserProcesses= is enabled in
 5900           systemd-logind's configuration a similar log message is generated for
 5901           processes killed at the end of each session due to this setting.
 5903         * systemd will now set the $JOURNAL_STREAM environment variable for all
 5904           services whose stdout/stderr are connected to the Journal (which
 5905           effectively means by default: all services). The variable contains
 5906           the device and inode number of the file descriptor used for
 5907           stdout/stderr. This may be used by invoked programs to detect whether
 5908           their stdout/stderr is connected to the Journal, in which case they
 5909           can switch over to direct Journal communication, thus being able to
 5910           pass extended, structured metadata along with their log messages. As
 5911           one example, this is now used by glib's logging primitives.
 5913         * When using systemd's default tmp.mount unit for /tmp, the mount point
 5914           will now be established with the "nosuid" and "nodev" options. This
 5915           avoids privilege escalation attacks that put traps and exploits into
 5916           /tmp.  However, this might cause problems if you e. g. put container
 5917           images or overlays into /tmp; if you need this, override tmp.mount's
 5918           "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
 5919           desired options.
 5921         * systemd now supports the "memory" cgroup controller also on
 5922           cgroup v2.
 5924         * The systemd-cgtop tool now optionally takes a control group path as
 5925           command line argument. If specified, the control group list shown is
 5926           limited to subgroups of that group.
 5928         * The SystemCallFilter= unit file setting gained support for
 5929           pre-defined, named system call filter sets. For example
 5930           SystemCallFilter=@clock is now an effective way to make all clock
 5931           changing-related system calls unavailable to a service. A number of
 5932           similar pre-defined groups are defined. Writing system call filters
 5933           for system services is simplified substantially with this new
 5934           concept. Accordingly, all of systemd's own, long-running services now
 5935           enable system call filtering based on this, by default.
 5937         * A new service setting MemoryDenyWriteExecute= has been added, taking
 5938           a boolean value. If turned on, a service may no longer create memory
 5939           mappings that are writable and executable at the same time. This
 5940           enhances security for services where this is enabled as it becomes
 5941           harder to dynamically write and then execute memory in exploited
 5942           service processes. This option has been enabled for all of systemd's
 5943           own long-running services.
 5945         * A new RestrictRealtime= service setting has been added, taking a
 5946           boolean argument. If set the service's processes may no longer
 5947           acquire realtime scheduling. This improves security as realtime
 5948           scheduling may otherwise be used to easily freeze the system.
 5950         * systemd-nspawn gained a new switch --notify-ready= taking a boolean
 5951           value. This may be used for requesting that the system manager inside
 5952           of the container reports start-up completion to nspawn which then
 5953           propagates this notification further to the service manager
 5954           supervising nspawn itself. A related option NotifyReady= in .nspawn
 5955           files has been added too. This functionality allows ordering of the
 5956           start-up of multiple containers using the usual systemd ordering
 5957           primitives.
 5959         * machinectl gained a new command "stop" that is an alias for
 5960           "terminate".
 5962         * systemd-resolved gained support for contacting DNS servers on
 5963           link-local IPv6 addresses.
 5965         * If systemd-resolved receives the SIGUSR2 signal it will now flush all
 5966           its caches. A method call for requesting the same operation has been
 5967           added to the bus API too, and is made available via "systemd-resolve
 5968           --flush-caches".
 5970         * systemd-resolve gained a new --status switch. If passed a brief
 5971           summary of the used DNS configuration with per-interface information
 5972           is shown.
 5974         * resolved.conf gained a new Cache= boolean option, defaulting to
 5975           on. If turned off local DNS caching is disabled. This comes with a
 5976           performance penalty in particular when DNSSEC is enabled. Note that
 5977           resolved disables its internal caching implicitly anyway, when the
 5978           configured DNS server is on a host-local IP address such as ::1 or
 5979 , thus automatically avoiding double local caching.
 5981         * systemd-resolved now listens on the local IP address
 5982           for DNS requests. This improves compatibility with local programs
 5983           that do not use the libc NSS or systemd-resolved's bus APIs for name
 5984           resolution. This minimal DNS service is only available to local
 5985           programs and does not implement the full DNS protocol, but enough to
 5986           cover local DNS clients. A new, static resolv.conf file, listing just
 5987           this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
 5988           now recommended to make /etc/resolv.conf a symlink to this file in
 5989           order to route all DNS lookups to systemd-resolved, regardless if
 5990           done via NSS, the bus API or raw DNS packets. Note that this local
 5991           DNS service is not as fully featured as the libc NSS or
 5992           systemd-resolved's bus APIs. For example, as unicast DNS cannot be
 5993           used to deliver link-local address information (as this implies
 5994           sending a local interface index along), LLMNR/mDNS support via this
 5995           interface is severely restricted. It is thus strongly recommended for
 5996           all applications to use the libc NSS API or native systemd-resolved
 5997           bus API instead.
 5999         * systemd-networkd's bridge support learned a new setting
 6000           VLANFiltering= for controlling VLAN filtering. Moreover a new section
 6001           in .network files has been added for configuring VLAN bridging in
 6002           more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
 6004         * systemd-networkd's IPv6 Router Advertisement code now makes use of
 6005           the DNSSL and RDNSS options. This means IPv6 DNS configuration may
 6006           now be acquired without relying on DHCPv6. Two new options
 6007           UseDomains= and UseDNS= have been added to configure this behaviour.
 6009         * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
 6010           renamed IPv6AcceptRA=, without altering its behaviour. The old
 6011           setting name remains available for compatibility reasons.
 6013         * The systemd-networkd VTI/VTI6 tunneling support gained new options
 6014           Key=, InputKey= and OutputKey=.
 6016         * systemd-networkd gained support for VRF ("Virtual Routing Function")
 6017           interface configuration.
 6019         * "systemctl edit" may now be used to create new unit files by
 6020           specifying the --force switch.
 6022         * sd-event gained a new function sd_event_get_iteration() for
 6023           requesting the current iteration counter of the event loop. It starts
 6024           at zero and is increased by one with each event loop iteration.
 6026         * A new rpm macro %systemd_ordering is provided by the macros.systemd
 6027           file. It can be used in lieu of %systemd_requires in packages which
 6028           don't use any systemd functionality and are intended to be installed
 6029           in minimal containers without systemd present. This macro provides
 6030           ordering dependencies to ensure that if the package is installed in
 6031           the same rpm transaction as systemd, systemd will be installed before
 6032           the scriptlets for the package are executed, allowing unit presets
 6033           to be handled.
 6035           New macros %_systemdgeneratordir and %_systemdusergeneratordir have
 6036           been added to simplify packaging of generators.
 6038         * The os-release file gained VERSION_CODENAME field for the
 6039           distribution nickname (e.g. VERSION_CODENAME=woody).
 6041         * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1
 6042           can be set to disable parsing of metadata and the creation
 6043           of persistent symlinks for that device.
 6045         * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
 6046           to make them available to logged-in users has been reverted.
 6048         * Much of the common code of the various systemd components is now
 6049           built into an internal shared library libsystemd-shared-231.so
 6050           (incorporating the systemd version number in the name, to be updated
 6051           with future releases) that the components link to. This should
 6052           decrease systemd footprint both in memory during runtime and on
 6053           disk. Note that the shared library is not for public use, and is
 6054           neither API nor ABI stable, but is likely to change with every new
 6055           released update. Packagers need to make sure that binaries
 6056           linking to libsystemd-shared.so are updated in step with the
 6057           library.
 6059         * Configuration for "mkosi" is now part of the systemd
 6060           repository. mkosi is a tool to easily build legacy-free OS images,
 6061           and is available on github: https://github.com/systemd/mkosi. If
 6062           "mkosi" is invoked in the build tree a new raw OS image is generated
 6063           incorporating the systemd sources currently being worked on and a
 6064           clean, fresh distribution installation. The generated OS image may be
 6065           booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
 6066           UEFI PC. This functionality is particularly useful to easily test
 6067           local changes made to systemd in a pristine, defined environment. See
 6068           doc/HACKING for details.
 6070         * configure learned the --with-support-url= option to specify the
 6071           distribution's bugtracker.
 6073         Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
 6074         Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
 6075         Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
 6076         Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
 6077         Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
 6078         Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
 6079         Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
 6080         Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
 6081         Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
 6082         Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
 6083         Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
 6084         Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
 6085         Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
 6086         Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
 6087         Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
 6088         Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
 6089         Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
 6090         Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
 6091         WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
 6093         — Berlin, 2016-07-25
 6095 CHANGES WITH 230:
 6097         * DNSSEC is now turned on by default in systemd-resolved (in
 6098           "allow-downgrade" mode), but may be turned off during compile time by
 6099           passing "--with-default-dnssec=no" to "configure" (and of course,
 6100           during runtime with DNSSEC= in resolved.conf). We recommend
 6101           downstreams to leave this on at least during development cycles and
 6102           report any issues with the DNSSEC logic upstream. We are very
 6103           interested in collecting feedback about the DNSSEC validator and its
 6104           limitations in the wild. Note however, that DNSSEC support is
 6105           probably nothing downstreams should turn on in stable distros just
 6106           yet, as it might create incompatibilities with a few DNS servers and
 6107           networks. We tried hard to make sure we downgrade to non-DNSSEC mode
 6108           automatically whenever we detect such incompatible setups, but there
 6109           might be systems we do not cover yet. Hence: please help us testing
 6110           the DNSSEC code, leave this on where you can, report back, but then
 6111           again don't consider turning this on in your stable, LTS or
 6112           production release just yet. (Note that you have to enable
 6113           nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
 6114           and its DNSSEC mode for hostname resolution from local
 6115           applications.)
 6117         * systemd-resolve conveniently resolves DANE records with the --tlsa
 6118           option and OPENPGPKEY records with the --openpgp option. It also
 6119           supports dumping raw DNS record data via the new --raw= switch.
 6121         * systemd-logind will now by default terminate user processes that are
 6122           part of the user session scope unit (session-XX.scope) when the user
 6123           logs out. This behavior is controlled by the KillUserProcesses=
 6124           setting in logind.conf, and the previous default of "no" is now
 6125           changed to "yes". This means that user sessions will be properly
 6126           cleaned up after, but additional steps are necessary to allow
 6127           intentionally long-running processes to survive logout.
 6129           While the user is logged in at least once, user@.service is running,
 6130           and any service that should survive the end of any individual login
 6131           session can be started at a user service or scope using systemd-run.
 6132           systemd-run(1) man page has been extended with an example which shows
 6133           how to run screen in a scope unit underneath user@.service. The same
 6134           command works for tmux.
 6136           After the user logs out of all sessions, user@.service will be
 6137           terminated too, by default, unless the user has "lingering" enabled.
 6138           To effectively allow users to run long-term tasks even if they are
 6139           logged out, lingering must be enabled for them. See loginctl(1) for
 6140           details. The default polkit policy was modified to allow users to
 6141           set lingering for themselves without authentication.
 6143           Previous defaults can be restored at compile time by the
 6144           --without-kill-user-processes option to "configure".
 6146         * systemd-logind gained new configuration settings SessionsMax= and
 6147           InhibitorsMax=, both with a default of 8192. It will not register new
 6148           user sessions or inhibitors above this limit.
 6150         * systemd-logind will now reload configuration on SIGHUP.
 6152         * The unified cgroup hierarchy added in Linux 4.5 is now supported.
 6153           Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
 6154           enable. Also, support for the "io" cgroup controller in the unified
 6155           hierarchy has been added, so that the "memory", "pids" and "io" are
 6156           now the controllers that are supported on the unified hierarchy.
 6158           WARNING: it is not possible to use previous systemd versions with
 6159           systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
 6160           is necessary to also update systemd in the initramfs if using the
 6161           unified hierarchy. An updated SELinux policy is also required.
 6163         * LLDP support has been extended, and both passive (receive-only) and
 6164           active (sender) modes are supported. Passive mode ("routers-only") is
 6165           enabled by default in systemd-networkd. Active LLDP mode is enabled
 6166           by default for containers on the internal network. The "networkctl
 6167           lldp" command may be used to list information gathered. "networkctl
 6168           status" will also show basic LLDP information on connected peers now.
 6170         * The IAID and DUID unique identifier sent in DHCP requests may now be
 6171           configured for the system and each .network file managed by
 6172           systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
 6174         * systemd-networkd gained support for configuring proxy ARP support for
 6175           each interface, via the ProxyArp= setting in .network files. It also
 6176           gained support for configuring the multicast querier feature of
 6177           bridge devices, via the new MulticastQuerier= setting in .netdev
 6178           files. Similarly, snooping on the IGMP traffic can be controlled
 6179           via the new setting MulticastSnooping=.
 6181           A new setting PreferredLifetime= has been added for addresses
 6182           configured in .network file to configure the lifetime intended for an
 6183           address.
 6185           The systemd-networkd DHCP server gained the option EmitRouter=, which
 6186           defaults to yes, to configure whether the DHCP Option 3 (Router)
 6187           should be emitted.
 6189         * The testing tool /usr/lib/systemd/systemd-activate is renamed to
 6190           systemd-socket-activate and installed into /usr/bin. It is now fully
 6191           supported.
 6193         * systemd-journald now uses separate threads to flush changes to disk
 6194           when closing journal files, thus reducing impact of slow disk I/O on
 6195           logging performance.
 6197         * The sd-journal API gained two new calls
 6198           sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
 6199           can be used to open journal files using file descriptors instead of
 6200           file or directory paths. sd_journal_open_container() has been
 6201           deprecated, sd_journal_open_directory_fd() should be used instead
 6202           with the flag SD_JOURNAL_OS_ROOT.
 6204         * journalctl learned a new output mode "-o short-unix" that outputs log
 6205           lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
 6206           UTC). It also gained support for a new --no-hostname setting to
 6207           suppress the hostname column in the family of "short" output modes.
 6209         * systemd-ask-password now optionally skips printing of the password to
 6210           stdout with --no-output which can be useful in scripts.
 6212         * Framebuffer devices (/dev/fb*) and 3D printers and scanners
 6213           (devices tagged with ID_MAKER_TOOL) are now tagged with
 6214           "uaccess" and are available to logged in users.
 6216         * The DeviceAllow= unit setting now supports specifiers (with "%").
 6218         * "systemctl show" gained a new --value switch, which allows print a
 6219           only the contents of a specific unit property, without also printing
 6220           the property's name. Similar support was added to "show*" verbs
 6221           of loginctl and machinectl that output "key=value" lists.
 6223         * A new unit type "generated" was added for files dynamically generated
 6224           by generator tools. Similarly, a new unit type "transient" is used
 6225           for unit files created using the runtime API. "systemctl enable" will
 6226           refuse to operate on such files.
 6228         * A new command "systemctl revert" has been added that may be used to
 6229           revert to the vendor version of a unit file, in case local changes
 6230           have been made by adding drop-ins or overriding the unit file.
 6232         * "machinectl clean" gained a new verb to automatically remove all or
 6233           just hidden container images.
 6235         * systemd-tmpfiles gained support for a new line type "e" for emptying
 6236           directories, if they exist, without creating them if they don't.
 6238         * systemd-nspawn gained support for automatically patching the UID/GIDs
 6239           of the owners and the ACLs of all files and directories in a
 6240           container tree to match the UID/GID user namespacing range selected
 6241           for the container invocation. This mode is enabled via the new
 6242           --private-users-chown switch. It also gained support for
 6243           automatically choosing a free, previously unused UID/GID range when
 6244           starting a container, via the new --private-users=pick setting (which
 6245           implies --private-users-chown). Together, these options for the first
 6246           time make user namespacing for nspawn containers fully automatic and
 6247           thus deployable. The systemd-nspawn@.service template unit file has
 6248           been changed to use this functionality by default.
 6250         * systemd-nspawn gained a new --network-zone= switch, that allows
 6251           creating ad-hoc virtual Ethernet links between multiple containers,
 6252           that only exist as long as at least one container referencing them is
 6253           running. This allows easy connecting of multiple containers with a
 6254           common link that implements an Ethernet broadcast domain. Each of
 6255           these network "zones" may be named relatively freely by the user, and
 6256           may be referenced by any number of containers, but each container may
 6257           only reference one of these "zones". On the lower level, this is
 6258           implemented by an automatically managed bridge network interface for
 6259           each zone, that is created when the first container referencing its
 6260           zone is created and removed when the last one referencing its zone
 6261           terminates.
 6263         * The default start timeout may now be configured on the kernel command
 6264           line via systemd.default_timeout_start_sec=. It was already
 6265           configurable via the DefaultTimeoutStartSec= option in
 6266           /etc/systemd/system.conf.
 6268         * Socket units gained a new TriggerLimitIntervalSec= and
 6269           TriggerLimitBurst= setting to configure a limit on the activation
 6270           rate of the socket unit.
 6272         * The LimitNICE= setting now optionally takes normal UNIX nice values
 6273           in addition to the raw integer limit value. If the specified
 6274           parameter is prefixed with "+" or "-" and is in the range -20…19 the
 6275           value is understood as UNIX nice value. If not prefixed like this it
 6276           is understood as raw RLIMIT_NICE limit.
 6278         * Note that the effect of the PrivateDevices= unit file setting changed
 6279           slightly with this release: the per-device /dev file system will be
 6280           mounted read-only from this version on, and will have "noexec"
 6281           set. This (minor) change of behavior might cause some (exceptional)
 6282           legacy software to break, when PrivateDevices=yes is set for its
 6283           service. Please leave PrivateDevices= off if you run into problems
 6284           with this.
 6286         * systemd-bootchart has been split out to a separate repository:
 6287           https://github.com/systemd/systemd-bootchart
 6289         * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
 6290           merged into the kernel in its current form.
 6292         * The compatibility libraries libsystemd-daemon.so,
 6293           libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
 6294           which have been deprecated since systemd-209 have been removed along
 6295           with the corresponding pkg-config files. All symbols provided by
 6296           those libraries are provided by libsystemd.so.
 6298         * The Capabilities= unit file setting has been removed (it is ignored
 6299           for backwards compatibility). AmbientCapabilities= and
 6300           CapabilityBoundingSet= should be used instead.
 6302         * A new special target has been added, initrd-root-device.target,
 6303           which creates a synchronization point for dependencies of the root
 6304           device in early userspace. Initramfs builders must ensure that this
 6305           target is now included in early userspace.
 6307         Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
 6308         Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
 6309         Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
 6310         Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
 6311         Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
 6312         R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
 6313         Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
 6314         Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
 6315         Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
 6316         Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
 6317         John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
 6318         Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
 6319         Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
 6320         Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
 6321         Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
 6322         mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
 6323         Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
 6324         Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
 6325         Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
 6326         Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
 6327         Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
 6328         Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
 6329         Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
 6330         Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
 6331         Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
 6332         Jędrzejewski-Szmek
 6334         — Fairfax, 2016-05-21
 6336 CHANGES WITH 229:
 6338         * The systemd-resolved DNS resolver service has gained a substantial
 6339           set of new features, most prominently it may now act as a DNSSEC
 6340           validating stub resolver. DNSSEC mode is currently turned off by
 6341           default, but is expected to be turned on by default in one of the
 6342           next releases. For now, we invite everybody to test the DNSSEC logic
 6343           by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
 6344           service also gained a full set of D-Bus interfaces, including calls
 6345           to configure DNS and DNSSEC settings per link (for use by external
 6346           network management software). systemd-resolved and systemd-networkd
 6347           now distinguish between "search" and "routing" domains. The former
 6348           are used to qualify single-label names, the latter are used purely
 6349           for routing lookups within certain domains to specific links.
 6350           resolved now also synthesizes RRs for all entries from /etc/hosts.
 6352         * The systemd-resolve tool (which is a client utility for
 6353           systemd-resolved) has been improved considerably and is now fully
 6354           supported and documented. Hence it has moved from /usr/lib/systemd to
 6355           /usr/bin.
 6357         * /dev/disk/by-path/ symlink support has been (re-)added for virtio
 6358           devices.
 6360         * The coredump collection logic has been reworked: when a coredump is
 6361           collected it is now written to disk, compressed and processed
 6362           (including stacktrace extraction) from a new instantiated service
 6363           systemd-coredump@.service, instead of directly from the
 6364           /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
 6365           processing large coredumps can take up a substantial amount of
 6366           resources and time, and this previously happened entirely outside of
 6367           systemd's service supervision. With the new logic the core_pattern
 6368           hook only does minimal metadata collection before passing off control
 6369           to the new instantiated service, which is configured with a time
 6370           limit, a nice level and other settings to minimize negative impact on
 6371           the rest of the system. Also note that the new logic will honour the
 6372           RLIMIT_CORE setting of the crashed process, which now allows users
 6373           and processes to turn off coredumping for their processes by setting
 6374           this limit.
 6376         * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
 6377           and all forked processes by default. Previously, PID 1 would leave
 6378           the setting at "0" for all processes, as set by the kernel. Note that
 6379           the resource limit traditionally has no effect on the generated
 6380           coredumps on the system if the /proc/sys/kernel/core_pattern hook
 6381           logic is used. Since the limit is now honoured (see above) its
 6382           default has been changed so that the coredumping logic is enabled by
 6383           default for all processes, while allowing specific opt-out.
 6385         * When the stacktrace is extracted from processes of system users, this
 6386           is now done as "systemd-coredump" user, in order to sandbox this
 6387           potentially security sensitive parsing operation. (Note that when
 6388           processing coredumps of normal users this is done under the user ID
 6389           of process that crashed, as before.) Packagers should take notice
 6390           that it is now necessary to create the "systemd-coredump" system user
 6391           and group at package installation time.
 6393         * The systemd-activate socket activation testing tool gained support
 6394           for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
 6395           and --seqpacket switches. It also has been extended to support both
 6396           new-style and inetd-style file descriptor passing. Use the new
 6397           --inetd switch to request inetd-style file descriptor passing.
 6399         * Most systemd tools now honor a new $SYSTEMD_COLORS environment
 6400           variable, which takes a boolean value. If set to false, ANSI color
 6401           output is disabled in the tools even when run on a terminal that
 6402           supports it.
 6404         * The VXLAN support in networkd now supports two new settings
 6405           DestinationPort= and PortRange=.
 6407         * A new systemd.machine_id= kernel command line switch has been added,
 6408           that may be used to set the machine ID in /etc/machine-id if it is
 6409           not initialized yet. This command line option has no effect if the
 6410           file is already initialized.
 6412         * systemd-nspawn gained a new --as-pid2 switch that invokes any
 6413           specified command line as PID 2 rather than PID 1 in the
 6414           container. In this mode PID 1 is a minimal stub init process that
 6415           implements the special POSIX and Linux semantics of PID 1 regarding
 6416           signal and child process management. Note that this stub init process
 6417           is implemented in nspawn itself and requires no support from the
 6418           container image. This new logic is useful to support running
 6419           arbitrary commands in the container, as normal processes are
 6420           generally not prepared to run as PID 1.
 6422         * systemd-nspawn gained a new --chdir= switch for setting the current
 6423           working directory for the process started in the container.
 6425         * "journalctl /dev/sda" will now output all kernel log messages for
 6426           specified device from the current boot, in addition to all devices
 6427           that are parents of it. This should make log output about devices
 6428           pretty useful, as long as kernel drivers attach enough metadata to
 6429           the log messages. (The usual SATA drivers do.)
 6431         * The sd-journal API gained two new calls
 6432           sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
 6433           that report whether log data from /run or /var has been found.
 6435         * journalctl gained a new switch "--fields" that prints all journal
 6436           record field names currently in use in the journal.  This is backed
 6437           by two new sd-journal API calls sd_journal_enumerate_fields() and
 6438           sd_journal_restart_fields().
 6440         * Most configurable timeouts in systemd now expect an argument of
 6441           "infinity" to turn them off, instead of "0" as before. The semantics
 6442           from now on is that a timeout of "0" means "now", and "infinity"
 6443           means "never". To maintain backwards compatibility, "0" continues to
 6444           turn off previously existing timeout settings.
 6446         * "systemctl reload-or-try-restart" has been renamed to "systemctl
 6447           try-reload-or-restart" to clarify what it actually does: the "try"
 6448           logic applies to both reloading and restarting, not just restarting.
 6449           The old name continues to be accepted for compatibility.
 6451         * On boot-up, when PID 1 detects that the system clock is behind the
 6452           release date of the systemd version in use, the clock is now set
 6453           to the latter. Previously, this was already done in timesyncd, in order
 6454           to avoid running with clocks set to the various clock epochs such as
 6455           1902, 1938 or 1970. With this change the logic is now done in PID 1
 6456           in addition to timesyncd during early boot-up, so that it is enforced
 6457           before the first process is spawned by systemd. Note that the logic
 6458           in timesyncd remains, as it is more comprehensive and ensures
 6459           clock monotonicity by maintaining a persistent timestamp file in
 6460           /var. Since /var is generally not available in earliest boot or the
 6461           initrd, this part of the logic remains in timesyncd, and is not done
 6462           by PID 1.
 6464         * Support for tweaking details in net_cls.class_id through the
 6465           NetClass= configuration directive has been removed, as the kernel
 6466           people have decided to deprecate that controller in cgroup v2.
 6467           Userspace tools such as nftables are moving over to setting rules
 6468           that are specific to the full cgroup path of a task, which obsoletes
 6469           these controllers anyway. The NetClass= directive is kept around for
 6470           legacy compatibility reasons. For a more in-depth description of the
 6471           kernel change, please refer to the respective upstream commit:
 6473             https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
 6475         * A new service setting RuntimeMaxSec= has been added that may be used
 6476           to specify a maximum runtime for a service. If the timeout is hit, the
 6477           service is terminated and put into a failure state.
 6479         * A new service setting AmbientCapabilities= has been added. It allows
 6480           configuration of additional Linux process capabilities that are
 6481           passed to the activated processes. This is only available on very
 6482           recent kernels.
 6484         * The process resource limit settings in service units may now be used
 6485           to configure hard and soft limits individually.
 6487         * The various libsystemd APIs such as sd-bus or sd-event now publicly
 6488           expose support for gcc's __attribute__((cleanup())) C extension.
 6489           Specifically, for many object destructor functions alternative
 6490           versions have been added that have names suffixed with "p" and take a
 6491           pointer to a pointer to the object to destroy, instead of just a
 6492           pointer to the object itself. This is useful because these destructor
 6493           functions may be used directly as parameters to the cleanup
 6494           construct. Internally, systemd has been a heavy user of this GCC
 6495           extension for a long time, and with this change similar support is
 6496           now available to consumers of the library outside of systemd. Note
 6497           that by using this extension in your sources compatibility with old
 6498           and strictly ANSI compatible C compilers is lost. However, all gcc or
 6499           LLVM versions of recent years support this extension.
 6501         * Timer units gained support for a new setting RandomizedDelaySec= that
 6502           allows configuring some additional randomized delay to the configured
 6503           time. This is useful to spread out timer events to avoid load peaks in
 6504           clusters or larger setups.
 6506         * Calendar time specifications now support sub-second accuracy.
 6508         * Socket units now support listening on SCTP and UDP-lite protocol
 6509           sockets.
 6511         * The sd-event API now comes with a full set of man pages.
 6513         * Older versions of systemd contained experimental support for
 6514           compressing journal files and coredumps with the LZ4 compressor that
 6515           was not compatible with the lz4 binary (due to API limitations of the
 6516           lz4 library). This support has been removed; only support for files
 6517           compatible with the lz4 binary remains. This LZ4 logic is now
 6518           officially supported and no longer considered experimental.
 6520         * The dkr image import logic has been removed again from importd. dkr's
 6521           micro-services focus doesn't fit into the machine image focus of
 6522           importd, and quickly got out of date with the upstream dkr API.
 6524         * Creation of the /run/lock/lockdev/ directory was dropped from
 6525           tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
 6526           been available for many years. If you still need this, you need to
 6527           create your own tmpfiles.d config file with:
 6529                   d /run/lock/lockdev 0775 root lock -
 6531         * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction=
 6532           and RebootArgument= have been moved from the [Service] section of
 6533           unit files to [Unit], and they are now supported on all unit types,
 6534           not just service units. Of course, systemd will continue to
 6535           understand these settings also at the old location, in order to
 6536           maintain compatibility.
 6538         Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
 6539         Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
 6540         Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
 6541         Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
 6542         Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
 6543         David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
 6544         Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
 6545         Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
 6546         Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
 6547         Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
 6548         Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
 6549         lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
 6550         Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
 6551         Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
 6552         Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
 6553         Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
 6554         Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
 6555         Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
 6556         Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
 6558         — Berlin, 2016-02-11
 6560 CHANGES WITH 228:
 6562         * A number of properties previously only settable in unit
 6563           files are now also available as properties to set when
 6564           creating transient units programmatically via the bus, as it
 6565           is exposed with systemd-run's --property=
 6566           setting. Specifically, these are: SyslogIdentifier=,
 6567           SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
 6568           EnvironmentFile=, ReadWriteDirectories=,
 6569           ReadOnlyDirectories=, InaccessibleDirectories=,
 6570           ProtectSystem=, ProtectHome=, RuntimeDirectory=.
 6572         * When creating transient services via the bus API it is now
 6573           possible to pass in a set of file descriptors to use as
 6574           STDIN/STDOUT/STDERR for the invoked process.
 6576         * Slice units may now be created transiently via the bus APIs,
 6577           similar to the way service and scope units may already be
 6578           created transiently.
 6580         * Wherever systemd expects a calendar timestamp specification
 6581           (like in journalctl's --since= and --until= switches) UTC
 6582           timestamps are now supported. Timestamps suffixed with "UTC"
 6583           are now considered to be in Universal Time Coordinated
 6584           instead of the local timezone. Also, timestamps may now
 6585           optionally be specified with sub-second accuracy. Both of
 6586           these additions also apply to recurring calendar event
 6587           specification, such as OnCalendar= in timer units.
 6589         * journalctl gained a new "--sync" switch that asks the
 6590           journal daemon to write all so far unwritten log messages to
 6591           disk and sync the files, before returning.
 6593         * systemd-tmpfiles learned two new line types "q" and "Q" that
 6594           operate like "v", but also set up a basic btrfs quota
 6595           hierarchy when used on a btrfs file system with quota
 6596           enabled.
 6598         * tmpfiles' "v", "q" and "Q" will now create a plain directory
 6599           instead of a subvolume (even on a btrfs file system) if the
 6600           root directory is a plain directory, and not a
 6601           subvolume. This should simplify things with certain chroot()
 6602           environments which are not aware of the concept of btrfs
 6603           subvolumes.
 6605         * systemd-detect-virt gained a new --chroot switch to detect
 6606           whether execution takes place in a chroot() environment.
 6608         * CPUAffinity= now takes CPU index ranges in addition to
 6609           individual indexes.
 6611         * The various memory-related resource limit settings (such as
 6612           LimitAS=) now understand the usual K, M, G, … suffixes to
 6613           the base of 1024 (IEC). Similar, the time-related resource
 6614           limit settings understand the usual min, h, day, … suffixes
 6615           now.
 6617         * There's a new system.conf setting DefaultTasksMax= to
 6618           control the default TasksMax= setting for services and
 6619           scopes running on the system. (TasksMax= is the primary
 6620           setting that exposes the "pids" cgroup controller on systemd
 6621           and was introduced in the previous systemd release.) The
 6622           setting now defaults to 512, which means services that are
 6623           not explicitly configured otherwise will only be able to
 6624           create 512 processes or threads at maximum, from this
 6625           version on. Note that this means that thread- or
 6626           process-heavy services might need to be reconfigured to set
 6627           TasksMax= to a higher value. It is sufficient to set
 6628           TasksMax= in these specific unit files to a higher value, or
 6629           even "infinity". Similar, there's now a logind.conf setting
 6630           UserTasksMax= that defaults to 4096 and limits the total
 6631           number of processes or tasks each user may own
 6632           concurrently. nspawn containers also have the TasksMax=
 6633           value set by default now, to 8192. Note that all of this
 6634           only has an effect if the "pids" cgroup controller is
 6635           enabled in the kernel. The general benefit of these changes
 6636           should be a more robust and safer system, that provides a
 6637           certain amount of per-service fork() bomb protection.
 6639         * systemd-nspawn gained the new --network-veth-extra= switch
 6640           to define additional and arbitrarily-named virtual Ethernet
 6641           links between the host and the container.
 6643         * A new service execution setting PassEnvironment= has been
 6644           added that allows importing select environment variables
 6645           from PID1's environment block into the environment block of
 6646           the service.
 6648         * Timer units gained support for a new RemainAfterElapse=
 6649           setting which takes a boolean argument. It defaults to on,
 6650           exposing behaviour unchanged to previous releases. If set to
 6651           off, timer units are unloaded after they elapsed if they
 6652           cannot elapse again. This is particularly useful for
 6653           transient timer units, which shall not stay around longer
 6654           than until they first elapse.
 6656         * systemd will now bump the net.unix.max_dgram_qlen to 512 by
 6657           default now (the kernel default is 16). This is beneficial
 6658           for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
 6659           allows substantially larger numbers of queued
 6660           datagrams. This should increase the capability of systemd to
 6661           parallelize boot-up, as logging and sd_notify() are unlikely
 6662           to stall execution anymore. If you need to change the value
 6663           from the new defaults, use the usual sysctl.d/ snippets.
 6665         * The compression framing format used by the journal or
 6666           coredump processing has changed to be in line with what the
 6667           official LZ4 tools generate. LZ4 compression support in
 6668           systemd was considered unsupported previously, as the format
 6669           was not compatible with the normal tools. With this release
 6670           this has changed now, and it is hence safe for downstream
 6671           distributions to turn it on. While not compressing as well
 6672           as the XZ, LZ4 is substantially faster, which makes
 6673           it a good default choice for the compression logic in the
 6674           journal and in coredump handling.
 6676         * Any reference to /etc/mtab has been dropped from
 6677           systemd. The file has been obsolete since a while, but
 6678           systemd refused to work on systems where it was incorrectly
 6679           set up (it should be a symlink or non-existent). Please make
 6680           sure to update to util-linux 2.27.1 or newer in conjunction
 6681           with this systemd release, which also drops any reference to
 6682           /etc/mtab. If you maintain a distribution make sure that no
 6683           software you package still references it, as this is a
 6684           likely source of bugs. There's also a glibc bug pending,
 6685           asking for removal of any reference to this obsolete file:
 6687           https://sourceware.org/bugzilla/show_bug.cgi?id=19108
 6689           Note that only util-linux versions built with
 6690           --enable-libmount-force-mountinfo are supported.
 6692         * Support for the ".snapshot" unit type has been removed. This
 6693           feature turned out to be little useful and little used, and
 6694           has now been removed from the core and from systemctl.
 6696         * The dependency types RequiresOverridable= and
 6697           RequisiteOverridable= have been removed from systemd. They
 6698           have been used only very sparingly to our knowledge and
 6699           other options that provide a similar effect (such as
 6700           systemctl --mode=ignore-dependencies) are much more useful
 6701           and commonly used. Moreover, they were only half-way
 6702           implemented as the option to control behaviour regarding
 6703           these dependencies was never added to systemctl. By removing
 6704           these dependency types the execution engine becomes a bit
 6705           simpler. Unit files that use these dependencies should be
 6706           changed to use the non-Overridable dependency types
 6707           instead. In fact, when parsing unit files with these
 6708           options, that's what systemd will automatically convert them
 6709           too, but it will also warn, asking users to fix the unit
 6710           files accordingly. Removal of these dependency types should
 6711           only affect a negligible number of unit files in the wild.
 6713         * Behaviour of networkd's IPForward= option changed
 6714           (again). It will no longer maintain a per-interface setting,
 6715           but propagate one way from interfaces where this is enabled
 6716           to the global kernel setting. The global setting will be
 6717           enabled when requested by a network that is set up, but
 6718           never be disabled again. This change was made to make sure
 6719           IPv4 and IPv6 behaviour regarding packet forwarding is
 6720           similar (as the Linux IPv6 stack does not support
 6721           per-interface control of this setting) and to minimize
 6722           surprises.
 6724         * In unit files the behaviour of %u, %U, %h, %s has
 6725           changed. These specifiers will now unconditionally resolve
 6726           to the various user database fields of the user that the
 6727           systemd instance is running as, instead of the user
 6728           configured in the specific unit via User=. Note that this
 6729           effectively doesn't change much, as resolving of these
 6730           specifiers was already turned off in the --system instance
 6731           of systemd, as we cannot do NSS lookups from PID 1. In the
 6732           --user instance of systemd these specifiers where correctly
 6733           resolved, but hardly made any sense, since the user instance
 6734           lacks privileges to do user switches anyway, and User= is
 6735           hence useless. Moreover, even in the --user instance of
 6736           systemd behaviour was awkward as it would only take settings
 6737           from User= assignment placed before the specifier into
 6738           account. In order to unify and simplify the logic around
 6739           this the specifiers will now always resolve to the
 6740           credentials of the user invoking the manager (which in case
 6741           of PID 1 is the root user).
 6743         Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
 6744         Yang, Daniel Machon, Daniel Mack, David Herrmann, David
 6745         Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
 6746         Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
 6747         Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
 6748         Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
 6749         Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
 6750         Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
 6751         Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
 6752         Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
 6753         Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
 6754         Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
 6755         Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
 6756         Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
 6757         Jędrzejewski-Szmek
 6759         — Berlin, 2015-11-18
 6761 CHANGES WITH 227:
 6763         * systemd now depends on util-linux v2.27. More specifically,
 6764           the newly added mount monitor feature in libmount now
 6765           replaces systemd's former own implementation.
 6767         * libmount mandates /etc/mtab not to be regular file, and
 6768           systemd now enforces this condition at early boot.
 6769           /etc/mtab has been deprecated and warned about for a very
 6770           long time, so systems running systemd should already have
 6771           stopped having this file around as anything else than a
 6772           symlink to /proc/self/mounts.
 6774         * Support for the "pids" cgroup controller has been added.  It
 6775           allows accounting the number of tasks in a cgroup and
 6776           enforcing limits on it. This adds two new setting
 6777           TasksAccounting= and TasksMax= to each unit, as well as a
 6778           global option DefaultTasksAccounting=.
 6780         * Support for the "net_cls" cgroup controller has been added.
 6781           It allows assigning a net class ID to each task in the
 6782           cgroup, which can then be used in firewall rules and traffic
 6783           shaping configurations. Note that the kernel netfilter net
 6784           class code does not currently work reliably for ingress
 6785           packets on unestablished sockets.
 6787           This adds a new config directive called NetClass= to CGroup
 6788           enabled units. Allowed values are positive numbers for fixed
 6789           assignments and "auto" for picking a free value
 6790           automatically.
 6792         * 'systemctl is-system-running' now returns 'offline' if the
 6793           system is not booted with systemd. This command can now be
 6794           used as a substitute for 'systemd-notify --booted'.
 6796         * Watchdog timeouts have been increased to 3 minutes for all
 6797           in-tree service files. Apparently, disk IO issues are more
 6798           frequent than we hoped, and user reported >1 minute waiting
 6799           for disk IO.
 6801         * 'machine-id-commit' functionality has been merged into
 6802           'machine-id-setup --commit'. The separate binary has been
 6803           removed.
 6805         * The WorkingDirectory= directive in unit files may now be set
 6806           to the special value '~'. In this case, the working
 6807           directory is set to the home directory of the user
 6808           configured in User=.
 6810         * "machinectl shell" will now open the shell in the home
 6811           directory of the selected user by default.
 6813         * The CrashChVT= configuration file setting is renamed to
 6814           CrashChangeVT=, following our usual logic of not
 6815           abbreviating unnecessarily. The old directive is still
 6816           supported for compat reasons. Also, this directive now takes
 6817           an integer value between 1 and 63, or a boolean value. The
 6818           formerly supported '-1' value for disabling stays around for
 6819           compat reasons.
 6821         * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
 6822           NoNewPrivileges=, TTYPath=, WorkingDirectory= and
 6823           RootDirectory= properties can now be set for transient
 6824           units.
 6826         * The systemd-analyze tool gained a new "set-log-target" verb
 6827           to change the logging target the system manager logs to
 6828           dynamically during runtime. This is similar to how
 6829           "systemd-analyze set-log-level" already changes the log
 6830           level.
 6832         * In nspawn /sys is now mounted as tmpfs, with only a selected
 6833           set of subdirectories mounted in from the real sysfs. This
 6834           enhances security slightly, and is useful for ensuring user
 6835           namespaces work correctly.
 6837         * Support for USB FunctionFS activation has been added. This
 6838           allows implementation of USB gadget services that are
 6839           activated as soon as they are requested, so that they don't
 6840           have to run continuously, similar to classic socket
 6841           activation.
 6843         * The "systemctl exit" command now optionally takes an
 6844           additional parameter that sets the exit code to return from
 6845           the systemd manager when exiting. This is only relevant when
 6846           running the systemd user instance, or when running the
 6847           system instance in a container.
 6849         * sd-bus gained the new API calls sd_bus_path_encode_many()
 6850           and sd_bus_path_decode_many() that allow easy encoding and
 6851           decoding of multiple identifier strings inside a D-Bus
 6852           object path. Another new call sd_bus_default_flush_close()
 6853           has been added to flush and close per-thread default
 6854           connections.
 6856         * systemd-cgtop gained support for a -M/--machine= switch to
 6857           show the control groups within a certain container only.
 6859         * "systemctl kill" gained support for an optional --fail
 6860           switch. If specified the requested operation will fail of no
 6861           processes have been killed, because the unit had no
 6862           processes attached, or similar.
 6864         * A new systemd.crash_reboot=1 kernel command line option has
 6865           been added that triggers a reboot after crashing. This can
 6866           also be set through CrashReboot= in systemd.conf.
 6868         * The RuntimeDirectory= setting now understands unit
 6869           specifiers like %i or %f.
 6871         * A new (still internal) library API sd-ipv4acd has been added,
 6872           that implements address conflict detection for IPv4. It's
 6873           based on code from sd-ipv4ll, and will be useful for
 6874           detecting DHCP address conflicts.
 6876         * File descriptors passed during socket activation may now be
 6877           named. A new API sd_listen_fds_with_names() is added to
 6878           access the names.  The default names may be overridden,
 6879           either in the .socket file using the FileDescriptorName=
 6880           parameter, or by passing FDNAME= when storing the file
 6881           descriptors using sd_notify().
 6883         * systemd-networkd gained support for:
 6885             - Setting the IPv6 Router Advertisement settings via
 6886               IPv6AcceptRouterAdvertisements= in .network files.
 6888             - Configuring the HelloTimeSec=, MaxAgeSec= and
 6889               ForwardDelaySec= bridge parameters in .netdev files.
 6891             - Configuring PreferredSource= for static routes in
 6892               .network files.
 6894         * The "ask-password" framework used to query for LUKS harddisk
 6895           passwords or SSL passwords during boot gained support for
 6896           caching passwords in the kernel keyring, if it is
 6897           available. This makes sure that the user only has to type in
 6898           a passphrase once if there are multiple objects to unlock
 6899           with the same one. Previously, such password caching was
 6900           available only when Plymouth was used; this moves the
 6901           caching logic into the systemd codebase itself. The
 6902           "systemd-ask-password" utility gained a new --keyname=
 6903           switch to control which kernel keyring key to use for
 6904           caching a password in. This functionality is also useful for
 6905           enabling display managers such as gdm to automatically
 6906           unlock the user's GNOME keyring if its passphrase, the
 6907           user's password and the harddisk password are the same, if
 6908           gdm-autologin is used.
 6910         * When downloading tar or raw images using "machinectl
 6911           pull-tar" or "machinectl pull-raw", a matching ".nspawn"
 6912           file is now also downloaded, if it is available and stored
 6913           next to the image file.
 6915         * Units of type ".socket" gained a new boolean setting
 6916           Writable= which is only useful in conjunction with
 6917           ListenSpecial=. If true, enables opening the specified
 6918           special file in O_RDWR mode rather than O_RDONLY mode.
 6920         * systemd-rfkill has been reworked to become a singleton
 6921           service that is activated through /dev/rfkill on each rfkill
 6922           state change and saves the settings to disk. This way,
 6923           systemd-rfkill is now compatible with devices that exist
 6924           only intermittendly, and even restores state if the previous
 6925           system shutdown was abrupt rather than clean.
 6927         * The journal daemon gained support for vacuuming old journal
 6928           files controlled by the number of files that shall remain,
 6929           in addition to the already existing control by size and by
 6930           date. This is useful as journal interleaving performance
 6931           degrades with too many separate journal files, and allows
 6932           putting an effective limit on them. The new setting defaults
 6933           to 100, but this may be changed by setting SystemMaxFiles=
 6934           and RuntimeMaxFiles= in journald.conf. Also, the
 6935           "journalctl" tool gained the new --vacuum-files= switch to
 6936           manually vacuum journal files to leave only the specified
 6937           number of files in place.
 6939         * udev will now create /dev/disk/by-path links for ATA devices
 6940           on kernels where that is supported.
 6942         * Galician, Serbian, Turkish and Korean translations were added.
 6944         Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
 6945         Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
 6946         (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
 6947         Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
 6948         Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
 6949         de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
 6950         Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
 6951         Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
 6952         Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
 6953         Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
 6954         Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
 6955         Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
 6956         Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
 6957         Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
 6958         Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
 6959         Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
 6960         Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
 6961         Zbigniew Jędrzejewski-Szmek, Марко М. Костић
 6963         — Berlin, 2015-10-07
 6965 CHANGES WITH 226:
 6967         * The DHCP implementation of systemd-networkd gained a set of
 6968           new features:
 6970           - The DHCP server now supports emitting DNS and NTP
 6971             information. It may be enabled and configured via
 6972             EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
 6973             and NTP information is enabled, but no servers are
 6974             configured, the corresponding uplink information (if there
 6975             is any) is propagated.
 6977           - Server and client now support transmission and reception
 6978             of timezone information. It can be configured via the
 6979             newly introduced network options UseTimezone=,
 6980             EmitTimezone=, and Timezone=.  Transmission of timezone
 6981             information is enabled between host and containers by
 6982             default now: the container will change its local timezone
 6983             to what the host has set.
 6985           - Lease timeouts can now be configured via
 6986             MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
 6988           - The DHCP server improved on the stability of
 6989             leases. Clients are more likely to get the same lease
 6990             information back, even if the server loses state.
 6992           - The DHCP server supports two new configuration options to
 6993             control the lease address pool metrics, PoolOffset= and
 6994             PoolSize=.
 6996         * The encapsulation limit of tunnels in systemd-networkd may
 6997           now be configured via 'EncapsulationLimit='. It allows
 6998           modifying the maximum additional levels of encapsulation
 6999           that are permitted to be prepended to a packet.
 7001         * systemd now supports the concept of user buses replacing
 7002           session buses, if used with dbus-1.10 (and enabled via dbus
 7003           --enable-user-session). It previously only supported this on
 7004           kdbus-enabled systems, and this release expands this to
 7005           'dbus-daemon' systems.
 7007         * systemd-networkd now supports predictable interface names
 7008           for virtio devices.
 7010         * systemd now optionally supports the new Linux kernel
 7011           "unified" control group hierarchy. If enabled via the kernel
 7012           command-line option 'systemd.unified_cgroup_hierarchy=1',
 7013           systemd will try to mount the unified cgroup hierarchy
 7014           directly on /sys/fs/cgroup. If not enabled, or not
 7015           available, systemd will fall back to the legacy cgroup
 7016           hierarchy setup, as before. Host system and containers can
 7017           mix and match legacy and unified hierarchies as they
 7018           wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
 7019           environment variable to individually select the hierarchy to
 7020           use for executed containers. By default, nspawn will use the
 7021           unified hierarchy for the containers if the host uses the
 7022           unified hierarchy, and the legacy hierarchy otherwise.
 7023           Please note that at this point the unified hierarchy is an
 7024           experimental kernel feature and is likely to change in one
 7025           of the next kernel releases.  Therefore, it should not be
 7026           enabled by default in downstream distributions yet. The
 7027           minimum required kernel version for the unified hierarchy to
 7028           work is 4.2. Note that when the unified hierarchy is used
 7029           for the first time delegated access to controllers is
 7030           safe. Because of this systemd-nspawn containers will get
 7031           access to controllers now, as will systemd user
 7032           sessions. This means containers and user sessions may now
 7033           manage their own resources, partitioning up what the system
 7034           grants them.
 7036         * A new special scope unit "init.scope" has been introduced
 7037           that encapsulates PID 1 of the system. It may be used to
 7038           determine resource usage and enforce resource limits on PID
 7039           1 itself. PID 1 hence moved out of the root of the control
 7040           group tree.
 7042         * The cgtop tool gained support for filtering out kernel
 7043           threads when counting tasks in a control group. Also, the
 7044           count of processes is now recursively summed up by
 7045           default. Two options -k and --recursive= have been added to
 7046           revert to old behaviour. The tool has also been updated to
 7047           work correctly in containers now.
 7049         * systemd-nspawn's --bind= and --bind-ro= options have been
 7050           extended to allow creation of non-recursive bind mounts.
 7052         * libsystemd gained two new calls sd_pid_get_cgroup() and
 7053           sd_peer_get_cgroup() which return the control group path of
 7054           a process or peer of a connected AF_UNIX socket. This
 7055           function call is particularly useful when implementing
 7056           delegated subtrees support in the control group hierarchy.
 7058         * The "sd-event" event loop API of libsystemd now supports
 7059           correct dequeuing of real-time signals, without losing
 7060           signal events.
 7062         * When systemd requests a polkit decision when managing units it
 7063           will now add additional fields to the request, including unit
 7064           name and desired operation. This enables more powerful polkit
 7065           policies, that make decisions depending on these parameters.
 7067         * nspawn learnt support for .nspawn settings files, that may
 7068           accompany the image files or directories of containers, and
 7069           may contain additional settings for the container. This is
 7070           an alternative to configuring container parameters via the
 7071           nspawn command line.
 7073         Contributions from: Cristian Rodríguez, Daniel Mack, David
 7074         Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
 7075         Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
 7076         Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
 7077         Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
 7078         Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
 7079         Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
 7080         Andersen, Tom Gundersen, Torstein Husebø
 7082         — Berlin, 2015-09-08
 7084 CHANGES WITH 225:
 7086         * machinectl gained a new verb 'shell' which opens a fresh
 7087           shell on the target container or the host. It is similar to
 7088           the existing 'login' command of machinectl, but spawns the
 7089           shell directly without prompting for username or
 7090           password. The pseudo machine '.host' now refers to the local
 7091           host and is used by default. Hence, 'machinectl shell' can
 7092           be used as replacement for 'su -' which spawns a session as
 7093           a fresh systemd unit in a way that is fully isolated from
 7094           the originating session.
 7096         * systemd-networkd learned to cope with private-zone DHCP
 7097           options and allows other programs to query the values.
 7099         * SELinux access control when enabling/disabling units is no
 7100           longer enforced with this release. The previous implementation
 7101           was incorrect, and a new corrected implementation is not yet
 7102           available. As unit file operations are still protected via
 7103           polkit and D-Bus policy this is not a security problem. Yet,
 7104           distributions which care about optimal SELinux support should
 7105           probably not stabilize on this release.
 7107         * sd-bus gained support for matches of type "arg0has=", that
 7108           test for membership of strings in string arrays sent in bus
 7109           messages.
 7111         * systemd-resolved now dumps the contents of its DNS and LLMNR
 7112           caches to the logs on reception of the SIGUSR1 signal. This
 7113           is useful to debug DNS behaviour.
 7115         * The coredumpctl tool gained a new --directory= option to
 7116           operate on journal files in a specific directory.
 7118         * "systemctl reboot" and related commands gained a new
 7119           "--message=" option which may be used to set a free-text
 7120           wall message when shutting down or rebooting the
 7121           system. This message is also logged, which is useful for
 7122           figuring out the reason for a reboot or shutdown a
 7123           posteriori.
 7125         * The "systemd-resolve-host" tool's -i switch now takes
 7126           network interface numbers as alternative to interface names.
 7128         * A new unit file setting for services has been introduced:
 7129           UtmpMode= allows configuration of how precisely systemd
 7130           handles utmp and wtmp entries for the service if this is
 7131           enabled. This allows writing services that appear similar to
 7132           user sessions in the output of the "w", "who", "last" and
 7133           "lastlog" tools.
 7135         * systemd-resolved will now locally synthesize DNS resource
 7136           records for the "localhost" and "gateway" domains as well as
 7137           the local hostname. This should ensure that clients querying
 7138           RRs via resolved will get similar results as those going via
 7139           NSS, if nss-myhostname is enabled.
 7141         Contributions from: Alastair Hughes, Alex Crawford, Daniel
 7142         Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
 7143         Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
 7144         Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
 7145         Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
 7146         Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
 7147         Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
 7148         Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
 7149         reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
 7150         Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
 7151         Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
 7152         WaLyong Cho, Zbigniew Jędrzejewski-Szmek
 7154         — Berlin, 2015-08-27
 7156 CHANGES WITH 224:
 7158         * The systemd-efi-boot-generator functionality was merged into
 7159           systemd-gpt-auto-generator.
 7161         * systemd-networkd now supports Group Policy for vxlan
 7162           devices. It can be enabled via the new boolean configuration
 7163           option called 'GroupPolicyExtension='.
 7165         Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
 7166         Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
 7167         Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
 7169         — Berlin, 2015-07-31
 7171 CHANGES WITH 223:
 7173         * The python-systemd code has been removed from the systemd repository.
 7174           A new repository has been created which accommodates the code from
 7175           now on, and we kindly ask distributions to create a separate package
 7176           for this: https://github.com/systemd/python-systemd
 7178         * The systemd daemon will now reload its main configuration
 7179           (/etc/systemd/system.conf) on daemon-reload.
 7181         * sd-dhcp now exposes vendor specific extensions via
 7182           sd_dhcp_lease_get_vendor_specific().
 7184         * systemd-networkd gained a number of new configuration options.
 7186           - A new boolean configuration option for TAP devices called
 7187             'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
 7188             device, thus allowing to send and receive GSO packets.
 7190           - A new tunnel configuration option called 'CopyDSCP='.
 7191             If enabled, the DSCP field of ip6 tunnels is copied into the
 7192             decapsulated packet.
 7194           - A set of boolean bridge configuration options were added.
 7195             'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
 7196             and 'UnicastFlood=' are now parsed by networkd and applied to the
 7197             respective bridge link device via the respective IFLA_BRPORT_*
 7198             netlink attribute.
 7200           - A new string configuration option to override the hostname sent
 7201             to a DHCP server, called 'Hostname='. If set and 'SendHostname='
 7202             is true, networkd will use the configured hostname instead of the
 7203             system hostname when sending DHCP requests.
 7205           - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
 7206             networkd will configure the IPv6 flow-label of the tunnel device
 7207             according to RFC2460.
 7209           - The 'macvtap' virtual network devices are now supported, similar to
 7210             the already supported 'macvlan' devices.
 7212         * systemd-resolved now implements RFC5452 to improve resilience against
 7213           cache poisoning. Additionally, source port randomization is enabled
 7214           by default to further protect against DNS spoofing attacks.
 7216         * nss-mymachines now supports translating UIDs and GIDs of running
 7217           containers with user-namespaces enabled. If a container 'foo'
 7218           translates a host uid 'UID' to the container uid 'TUID', then
 7219           nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
 7220           (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
 7221           mapped as 'vg-foo-TGID'.
 7223         Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
 7224         Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
 7225         HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
 7226         Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
 7227         Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
 7228         Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
 7229         Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
 7230         Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
 7231         Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
 7232         Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
 7234         — Berlin, 2015-07-29
 7236 CHANGES WITH 222:
 7238         * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
 7239           There are no known issues with current sysfs, and udev does not need
 7240           or should be used to work around such bugs.
 7242         * udev does no longer enable USB HID power management. Several reports
 7243           indicate, that some devices cannot handle that setting.
 7245         * The udev accelerometer helper was removed. The functionality
 7246           is now fully included in iio-sensor-proxy. But this means,
 7247           older iio-sensor-proxy versions will no longer provide
 7248           accelerometer/orientation data with this systemd version.
 7249           Please upgrade iio-sensor-proxy to version 1.0.
 7251         * networkd gained a new configuration option IPv6PrivacyExtensions=
 7252           which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
 7253           for Stateless Address") on selected networks.
 7255         * For the sake of fewer build-time dependencies and less code in the
 7256           main repository, the python bindings are about to be removed in the
 7257           next release. A new repository has been created which accommodates
 7258           the code from now on, and we kindly ask distributions to create a
 7259           separate package for this. The removal will take place in v223.
 7261             https://github.com/systemd/python-systemd
 7263         Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
 7264         Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
 7265         daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
 7266         Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
 7267         Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
 7268         (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
 7269         Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
 7270         Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
 7271         Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
 7272         Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
 7274         — Berlin, 2015-07-07
 7276 CHANGES WITH 221:
 7278         * The sd-bus.h and sd-event.h APIs have now been declared
 7279           stable and have been added to the official interface of
 7280           libsystemd.so. sd-bus implements an alternative D-Bus client
 7281           library, that is relatively easy to use, very efficient and
 7282           supports both classic D-Bus as well as kdbus as transport
 7283           backend. sd-event is a generic event loop abstraction that
 7284           is built around Linux epoll, but adds features such as event
 7285           prioritization or efficient timer handling. Both APIs are good
 7286           choices for C programs looking for a bus and/or event loop
 7287           implementation that is minimal and does not have to be
 7288           portable to other kernels.
 7290         * kdbus support is no longer compile-time optional. It is now
 7291           always built-in. However, it can still be disabled at
 7292           runtime using the kdbus=0 kernel command line setting, and
 7293           that setting may be changed to default to off, by specifying
 7294           --disable-kdbus at build-time. Note though that the kernel
 7295           command line setting has no effect if the kdbus.ko kernel
 7296           module is not installed, in which case kdbus is (obviously)
 7297           also disabled. We encourage all downstream distributions to
 7298           begin testing kdbus by adding it to the kernel images in the
 7299           development distributions, and leaving kdbus support in
 7300           systemd enabled.
 7302         * The minimal required util-linux version has been bumped to
 7303           2.26.
 7305         * Support for chkconfig (--enable-chkconfig) was removed in
 7306           favor of calling an abstraction tool
 7307           /lib/systemd/systemd-sysv-install. This needs to be
 7308           implemented for your distribution. See "SYSV INIT.D SCRIPTS"
 7309           in README for details.
 7311         * If there's a systemd unit and a SysV init script for the
 7312           same service name, and the user executes "systemctl enable"
 7313           for it (or a related call), then this will now enable both
 7314           (or execute the related operation on both), not just the
 7315           unit.
 7317         * The libudev API documentation has been converted from gtkdoc
 7318           into man pages.
 7320         * gudev has been removed from the systemd tree, it is now an
 7321           external project.
 7323         * The systemd-cgtop tool learnt a new --raw switch to generate
 7324           "raw" (machine parsable) output.
 7326         * networkd's IPForwarding= .network file setting learnt the
 7327           new setting "kernel", which ensures that networkd does not
 7328           change the IP forwarding sysctl from the default kernel
 7329           state.
 7331         * The systemd-logind bus API now exposes a new boolean
 7332           property "Docked" that reports whether logind considers the
 7333           system "docked", i.e. connected to a docking station or not.
 7335         Contributions from: Alex Crawford, Andreas Pokorny, Andrei
 7336         Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
 7337         Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
 7338         David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
 7339         Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
 7340         Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
 7341         Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
 7342         Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
 7343         Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
 7344         Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
 7345         Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
 7346         Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
 7347         Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
 7348         Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
 7349         Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
 7350         Fink, Zbigniew Jędrzejewski-Szmek
 7352         — Berlin, 2015-06-19
 7354 CHANGES WITH 220:
 7356         * The gudev library has been extracted into a separate repository
 7357           available at: https://git.gnome.org/browse/libgudev/
 7358           It is now managed as part of the Gnome project. Distributions
 7359           are recommended to pass --disable-gudev to systemd and use
 7360           gudev from the Gnome project instead. gudev is still included
 7361           in systemd, for now. It will be removed soon, though. Please
 7362           also see the announcement-thread on systemd-devel:
 7363           https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
 7365         * systemd now exposes a CPUUsageNSec= property for each
 7366           service unit on the bus, that contains the overall consumed
 7367           CPU time of a service (the sum of what each process of the
 7368           service consumed). This value is only available if
 7369           CPUAccounting= is turned on for a service, and is then shown
 7370           in the "systemctl status" output.
 7372         * Support for configuring alternative mappings of the old SysV
 7373           runlevels to systemd targets has been removed. They are now
 7374           hardcoded in a way that runlevels 2, 3, 4 all map to
 7375           multi-user.target and 5 to graphical.target (which
 7376           previously was already the default behaviour).
 7378         * The auto-mounter logic gained support for mount point
 7379           expiry, using a new TimeoutIdleSec= setting in .automount
 7380           units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
 7382         * The EFI System Partition (ESP) as mounted to /boot by
 7383           systemd-efi-boot-generator will now be unmounted
 7384           automatically after 2 minutes of not being used. This should
 7385           minimize the risk of ESP corruptions.
 7387         * New /etc/fstab options x-systemd.requires= and
 7388           x-systemd.requires-mounts-for= are now supported to express
 7389           additional dependencies for mounts. This is useful for
 7390           journaling file systems that support external journal
 7391           devices or overlay file systems that require underlying file
 7392           systems to be mounted.
 7394         * systemd does not support direct live-upgrades (via systemctl
 7395           daemon-reexec) from versions older than v44 anymore. As no
 7396           distribution we are aware of shipped such old versions in a
 7397           stable release this should not be problematic.
 7399         * When systemd forks off a new per-connection service instance
 7400           it will now set the $REMOTE_ADDR environment variable to the
 7401           remote IP address, and $REMOTE_PORT environment variable to
 7402           the remote IP port. This behaviour is similar to the
 7403           corresponding environment variables defined by CGI.
 7405         * systemd-networkd gained support for uplink failure
 7406           detection. The BindCarrier= option allows binding interface
 7407           configuration dynamically to the link sense of other
 7408           interfaces. This is useful to achieve behaviour like in
 7409           network switches.
 7411         * systemd-networkd gained support for configuring the DHCP
 7412           client identifier to use when requesting leases.
 7414         * systemd-networkd now has a per-network UseNTP= option to
 7415           configure whether NTP server information acquired via DHCP
 7416           is passed on to services like systemd-timesyncd.
 7418         * systemd-networkd gained support for vti6 tunnels.
 7420         * Note that systemd-networkd manages the sysctl variable
 7421           /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
 7422           it is configured for since v219. The variable controls IP
 7423           forwarding, and is a per-interface alternative to the global
 7424           /proc/sys/net/ipv[46]/ip_forward. This setting is
 7425           configurable in the IPForward= option, which defaults to
 7426           "no". This means if networkd is used for an interface it is
 7427           no longer sufficient to set the global sysctl option to turn
 7428           on IP forwarding! Instead, the .network file option
 7429           IPForward= needs to be turned on! Note that the
 7430           implementation of this behaviour was broken in v219 and has
 7431           been fixed in v220.
 7433         * Many bonding and vxlan options are now configurable in
 7434           systemd-networkd.
 7436         * systemd-nspawn gained a new --property= setting to set unit
 7437           properties for the container scope. This is useful for
 7438           setting resource parameters (e.g. "CPUShares=500") on
 7439           containers started from the command line.
 7441         * systemd-nspawn gained a new --private-users= switch to make
 7442           use of user namespacing available on recent Linux kernels.
 7444         * systemd-nspawn may now be called as part of a shell pipeline
 7445           in which case the pipes used for stdin and stdout are passed
 7446           directly to the process invoked in the container, without
 7447           indirection via a pseudo tty.
 7449         * systemd-nspawn gained a new switch to control the UNIX
 7450           signal to use when killing the init process of the container
 7451           when shutting down.
 7453         * systemd-nspawn gained a new --overlay= switch for mounting
 7454           overlay file systems into the container using the new kernel
 7455           overlayfs support.
 7457         * When a container image is imported via systemd-importd and
 7458           the host file system is not btrfs, a loopback block device
 7459           file is created in /var/lib/machines.raw with a btrfs file
 7460           system inside. It is then mounted to /var/lib/machines to
 7461           enable btrfs features for container management. The loopback
 7462           file and btrfs file system is grown as needed when container
 7463           images are imported via systemd-importd.
 7465         * systemd-machined/systemd-importd gained support for btrfs
 7466           quota, to enforce container disk space limits on disk. This
 7467           is exposed in "machinectl set-limit".
 7469         * systemd-importd now can import containers from local .tar,
 7470           .raw and .qcow2 images, and export them to .tar and .raw. It
 7471           can also import dkr v2 images now from the network (on top
 7472           of v1 as before).
 7474         * systemd-importd gained support for verifying downloaded
 7475           images with gpg2 (previously only gpg1 was supported).
 7477         * systemd-machined, systemd-logind, systemd: most bus calls are
 7478           now accessible to unprivileged processes via polkit. Also,
 7479           systemd-logind will now allow users to kill their own sessions
 7480           without further privileges or authorization.
 7482         * systemd-shutdownd has been removed. This service was
 7483           previously responsible for implementing scheduled shutdowns
 7484           as exposed in /usr/bin/shutdown's time parameter. This
 7485           functionality has now been moved into systemd-logind and is
 7486           accessible via a bus interface.
 7488         * "systemctl reboot" gained a new switch --firmware-setup that
 7489           can be used to reboot into the EFI firmware setup, if that
 7490           is available. systemd-logind now exposes an API on the bus
 7491           to trigger such reboots, in case graphical desktop UIs want
 7492           to cover this functionality.
 7494         * "systemctl enable", "systemctl disable" and "systemctl mask"
 7495           now support a new "--now" switch. If specified the units
 7496           that are enabled will also be started, and the ones
 7497           disabled/masked also stopped.
 7499         * The Gummiboot EFI boot loader tool has been merged into
 7500           systemd, and renamed to "systemd-boot". The bootctl tool has been
 7501           updated to support systemd-boot.
 7503         * An EFI kernel stub has been added that may be used to create
 7504           kernel EFI binaries that contain not only the actual kernel,
 7505           but also an initrd, boot splash, command line and OS release
 7506           information. This combined binary can then be signed as a
 7507           single image, so that the firmware can verify it all in one
 7508           step. systemd-boot has special support for EFI binaries created
 7509           like this and can extract OS release information from them
 7510           and show them in the boot menu. This functionality is useful
 7511           to implement cryptographically verified boot schemes.
 7513         * Optional support has been added to systemd-fsck to pass
 7514           fsck's progress report to an AF_UNIX socket in the file
 7515           system.
 7517         * udev will no longer create device symlinks for all block devices by
 7518           default. A deny list for excluding special block devices from this
 7519           logic has been turned into a allow list that requires picking block
 7520           devices explicitly that require device symlinks.
 7522         * A new (currently still internal) API sd-device.h has been
 7523           added to libsystemd. This modernized API is supposed to
 7524           replace libudev eventually. In fact, already much of libudev
 7525           is now just a wrapper around sd-device.h.
 7527         * A new hwdb database for storing metadata about pointing
 7528           stick devices has been added.
 7530         * systemd-tmpfiles gained support for setting file attributes
 7531           similar to the "chattr" tool with new 'h' and 'H' lines.
 7533         * systemd-journald will no longer unconditionally set the
 7534           btrfs NOCOW flag on new journal files. This is instead done
 7535           with tmpfiles snippet using the new 'h' line type. This
 7536           allows easy disabling of this logic, by masking the
 7537           journal-nocow.conf tmpfiles file.
 7539         * systemd-journald will now translate audit message types to
 7540           human readable identifiers when writing them to the
 7541           journal. This should improve readability of audit messages.
 7543         * The LUKS logic gained support for the offset= and skip=
 7544           options in /etc/crypttab, as previously implemented by
 7545           Debian.
 7547         * /usr/lib/os-release gained a new optional field VARIANT= for
 7548           distributions that support multiple variants (such as a
 7549           desktop edition, a server edition, …)
 7551         Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
 7552         Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
 7553         Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
 7554         Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
 7555         Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
 7556         Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
 7557         Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
 7558         Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
 7559         Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
 7560         Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
 7561         Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
 7562         Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
 7563         Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
 7564         Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
 7565         De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
 7566         Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
 7567         Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
 7568         Michael Biebl, Michael Marineau, Michael Olbrich, Michal
 7569         Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
 7570         Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
 7571         Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
 7572         Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
 7573         Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
 7574         Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
 7575         Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
 7576         Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
 7577         Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
 7579         — Berlin, 2015-05-22
 7581 CHANGES WITH 219:
 7583         * Introduce a new API "sd-hwdb.h" for querying the hardware
 7584           metadata database. With this minimal interface one can query
 7585           and enumerate the udev hwdb, decoupled from the old libudev
 7586           library. libudev's interface for this is now only a wrapper
 7587           around sd-hwdb. A new tool systemd-hwdb has been added to
 7588           interface with and update the database.
 7590         * When any of systemd's tools copies files (for example due to
 7591           tmpfiles' C lines) a btrfs reflink will attempted first,
 7592           before bytewise copying is done.
 7594         * systemd-nspawn gained a new --ephemeral switch. When
 7595           specified a btrfs snapshot is taken of the container's root
 7596           directory, and immediately removed when the container
 7597           terminates again. Thus, a container can be started whose
 7598           changes never alter the container's root directory, and are
 7599           lost on container termination. This switch can also be used
 7600           for starting a container off the root file system of the
 7601           host without affecting the host OS. This switch is only
 7602           available on btrfs file systems.
 7604         * systemd-nspawn gained a new --template= switch. It takes the
 7605           path to a container tree to use as template for the tree
 7606           specified via --directory=, should that directory be
 7607           missing. This allows instantiating containers dynamically,
 7608           on first run. This switch is only available on btrfs file
 7609           systems.
 7611         * When a .mount unit refers to a mount point on which multiple
 7612           mounts are stacked, and the .mount unit is stopped all of
 7613           the stacked mount points will now be unmounted until no
 7614           mount point remains.
 7616         * systemd now has an explicit notion of supported and
 7617           unsupported unit types. Jobs enqueued for unsupported unit
 7618           types will now fail with an "unsupported" error code. More
 7619           specifically .swap, .automount and .device units are not
 7620           supported in containers, .busname units are not supported on
 7621           non-kdbus systems. .swap and .automount are also not
 7622           supported if their respective kernel compile time options
 7623           are disabled.
 7625         * machinectl gained support for two new "copy-from" and
 7626           "copy-to" commands for copying files from a running
 7627           container to the host or vice versa.
 7629         * machinectl gained support for a new "bind" command to bind
 7630           mount host directories into local containers. This is
 7631           currently only supported for nspawn containers.
 7633         * networkd gained support for configuring bridge forwarding
 7634           database entries (fdb) from .network files.
 7636         * A new tiny daemon "systemd-importd" has been added that can
 7637           download container images in tar, raw, qcow2 or dkr formats,
 7638           and make them available locally in /var/lib/machines, so
 7639           that they can run as nspawn containers. The daemon can GPG
 7640           verify the downloads (not supported for dkr, since it has no
 7641           provisions for verifying downloads). It will transparently
 7642           decompress bz2, xz, gzip compressed downloads if necessary,
 7643           and restore sparse files on disk. The daemon uses privilege
 7644           separation to ensure the actual download logic runs with
 7645           fewer privileges than the daemon itself. machinectl has
 7646           gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
 7647           make the functionality of importd available to the
 7648           user. With this in place the Fedora and Ubuntu "Cloud"
 7649           images can be downloaded and booted as containers unmodified
 7650           (the Fedora images lack the appropriate GPG signature files
 7651           currently, so they cannot be verified, but this will change
 7652           soon, hopefully). Note that downloading images is currently
 7653           only fully supported on btrfs.
 7655         * machinectl is now able to list container images found in
 7656           /var/lib/machines, along with some metadata about sizes of
 7657           disk and similar. If the directory is located on btrfs and
 7658           quota is enabled, this includes quota display. A new command
 7659           "image-status" has been added that shows additional
 7660           information about images.
 7662         * machinectl is now able to clone container images
 7663           efficiently, if the underlying file system (btrfs) supports
 7664           it, with the new "machinectl clone" command. It also
 7665           gained commands for renaming and removing images, as well as
 7666           marking them read-only or read-write (supported also on
 7667           legacy file systems).
 7669         * networkd gained support for collecting LLDP network
 7670           announcements, from hardware that supports this. This is
 7671           shown in networkctl output.
 7673         * systemd-run gained support for a new -t (--pty) switch for
 7674           invoking a binary on a pty whose input and output is
 7675           connected to the invoking terminal. This allows executing
 7676           processes as system services while interactively
 7677           communicating with them via the terminal. Most interestingly
 7678           this is supported across container boundaries. Invoking
 7679           "systemd-run -t /bin/bash" is an alternative to running a
 7680           full login session, the difference being that the former
 7681           will not register a session, nor go through the PAM session
 7682           setup.
 7684         * tmpfiles gained support for a new "v" line type for creating
 7685           btrfs subvolumes. If the underlying file system is a legacy
 7686           file system, this automatically degrades to creating a
 7687           normal directory. Among others /var/lib/machines is now
 7688           created like this at boot, should it be missing.
 7690         * The directory /var/lib/containers/ has been deprecated and
 7691           been replaced by /var/lib/machines. The term "machines" has
 7692           been used in the systemd context as generic term for both
 7693           VMs and containers, and hence appears more appropriate for
 7694           this, as the directory can also contain raw images bootable
 7695           via qemu/kvm.
 7697         * systemd-nspawn when invoked with -M but without --directory=
 7698           or --image= is now capable of searching for the container
 7699           root directory, subvolume or disk image automatically, in
 7700           /var/lib/machines. systemd-nspawn@.service has been updated
 7701           to make use of this, thus allowing it to be used for raw
 7702           disk images, too.
 7704         * A new machines.target unit has been introduced that is
 7705           supposed to group all containers/VMs invoked as services on
 7706           the system. systemd-nspawn@.service has been updated to
 7707           integrate with that.
 7709         * machinectl gained a new "start" command, for invoking a
 7710           container as a service. "machinectl start foo" is mostly
 7711           equivalent to "systemctl start systemd-nspawn@foo.service",
 7712           but handles escaping in a nicer way.
 7714         * systemd-nspawn will now mount most of the cgroupfs tree
 7715           read-only into each container, with the exception of the
 7716           container's own subtree in the name=systemd hierarchy.
 7718         * journald now sets the special FS_NOCOW file flag for its
 7719           journal files. This should improve performance on btrfs, by
 7720           avoiding heavy fragmentation when journald's write-pattern
 7721           is used on COW file systems. It degrades btrfs' data
 7722           integrity guarantees for the files to the same levels as for
 7723           ext3/ext4 however. This should be OK though as journald does
 7724           its own data integrity checks and all its objects are
 7725           checksummed on disk. Also, journald should handle btrfs disk
 7726           full events a lot more gracefully now, by processing SIGBUS
 7727           errors, and not relying on fallocate() anymore.
 7729         * When journald detects that journal files it is writing to
 7730           have been deleted it will immediately start new journal
 7731           files.
 7733         * systemd now provides a way to store file descriptors
 7734           per-service in PID 1. This is useful for daemons to ensure
 7735           that fds they require are not lost during a daemon
 7736           restart. The fds are passed to the daemon on the next
 7737           invocation in the same way socket activation fds are
 7738           passed. This is now used by journald to ensure that the
 7739           various sockets connected to all the system's stdout/stderr
 7740           are not lost when journald is restarted. File descriptors
 7741           may be stored in PID 1 via the sd_pid_notify_with_fds() API,
 7742           an extension to sd_notify(). Note that a limit is enforced
 7743           on the number of fds a service can store in PID 1, and it
 7744           defaults to 0, so that no fds may be stored, unless this is
 7745           explicitly turned on.
 7747         * The default TERM variable to use for units connected to a
 7748           terminal, when no other value is explicitly is set is now
 7749           vt220 rather than vt102. This should be fairly safe still,
 7750           but allows PgUp/PgDn work.
 7752         * The /etc/crypttab option header= as known from Debian is now
 7753           supported.
 7755         * "loginctl user-status" and "loginctl session-status" will
 7756           now show the last 10 lines of log messages of the
 7757           user/session following the status output. Similar,
 7758           "machinectl status" will show the last 10 log lines
 7759           associated with a virtual machine or container
 7760           service. (Note that this is usually not the log messages
 7761           done in the VM/container itself, but simply what the
 7762           container manager logs. For nspawn this includes all console
 7763           output however.)
 7765         * "loginctl session-status" without further argument will now
 7766           show the status of the session of the caller. Similar,
 7767           "lock-session", "unlock-session", "activate",
 7768           "enable-linger", "disable-linger" may now be called without
 7769           session/user parameter in which case they apply to the
 7770           caller's session/user.
 7772         * An X11 session scriptlet is now shipped that uploads
 7773           $DISPLAY and $XAUTHORITY into the environment of the systemd
 7774           --user daemon if a session begins. This should improve
 7775           compatibility with X11 enabled applications run as systemd
 7776           user services.
 7778         * Generators are now subject to masking via /etc and /run, the
 7779           same way as unit files.
 7781         * networkd .network files gained support for configuring
 7782           per-link IPv4/IPv6 packet forwarding as well as IPv4
 7783           masquerading. This is by default turned on for veth links to
 7784           containers, as registered by systemd-nspawn. This means that
 7785           nspawn containers run with --network-veth will now get
 7786           automatic routed access to the host's networks without any
 7787           further configuration or setup, as long as networkd runs on
 7788           the host.
 7790         * systemd-nspawn gained the --port= (-p) switch to expose TCP
 7791           or UDP posts of a container on the host. With this in place
 7792           it is possible to run containers with private veth links
 7793           (--network-veth), and have their functionality exposed on
 7794           the host as if their services were running directly on the
 7795           host.
 7797         * systemd-nspawn's --network-veth switch now gained a short
 7798           version "-n", since with the changes above it is now truly
 7799           useful out-of-the-box. The systemd-nspawn@.service has been
 7800           updated to make use of it too by default.
 7802         * systemd-nspawn will now maintain a per-image R/W lock, to
 7803           ensure that the same image is not started more than once
 7804           writable. (It's OK to run an image multiple times
 7805           simultaneously in read-only mode.)
 7807         * systemd-nspawn's --image= option is now capable of
 7808           dissecting and booting MBR and GPT disk images that contain
 7809           only a single active Linux partition. Previously it
 7810           supported only GPT disk images with proper GPT type
 7811           IDs. This allows running cloud images from major
 7812           distributions directly with systemd-nspawn, without
 7813           modification.
 7815         * In addition to collecting mouse dpi data in the udev
 7816           hardware database, there's now support for collecting angle
 7817           information for mouse scroll wheels. The database is
 7818           supposed to guarantee similar scrolling behavior on mice
 7819           that it knows about. There's also support for collecting
 7820           information about Touchpad types.
 7822         * udev's input_id built-in will now also collect touch screen
 7823           dimension data and attach it to probed devices.
 7825         * /etc/os-release gained support for a Distribution Privacy
 7826           Policy link field.
 7828         * networkd gained support for creating "ipvlan", "gretap",
 7829           "ip6gre", "ip6gretap" and "ip6tnl" network devices.
 7831         * systemd-tmpfiles gained support for "a" lines for setting
 7832           ACLs on files.
 7834         * systemd-nspawn will now mount /tmp in the container to
 7835           tmpfs, automatically.
 7837         * systemd now exposes the memory.usage_in_bytes cgroup
 7838           attribute and shows it for each service in the "systemctl
 7839           status" output, if available.
 7841         * When the user presses Ctrl-Alt-Del more than 7x within 2s an
 7842           immediate reboot is triggered. This useful if shutdown is
 7843           hung and is unable to complete, to expedite the
 7844           operation. Note that this kind of reboot will still unmount
 7845           all file systems, and hence should not result in fsck being
 7846           run on next reboot.
 7848         * A .device unit for an optical block device will now be
 7849           considered active only when a medium is in the drive. Also,
 7850           mount units are now bound to their backing devices thus
 7851           triggering automatic unmounting when devices become
 7852           unavailable. With this in place systemd will now
 7853           automatically unmount left-over mounts when a CD-ROM is
 7854           ejected or an USB stick is yanked from the system.
 7856         * networkd-wait-online now has support for waiting for
 7857           specific interfaces only (with globbing), and for giving up
 7858           after a configurable timeout.
 7860         * networkd now exits when idle. It will be automatically
 7861           restarted as soon as interfaces show up, are removed or
 7862           change state. networkd will stay around as long as there is
 7863           at least one DHCP state machine or similar around, that keep
 7864           it non-idle.
 7866         * networkd may now configure IPv6 link-local addressing in
 7867           addition to IPv4 link-local addressing.
 7869         * The IPv6 "token" for use in SLAAC may now be configured for
 7870           each .network interface in networkd.
 7872         * Routes configured with networkd may now be assigned a scope
 7873           in .network files.
 7875         * networkd's [Match] sections now support globbing and lists
 7876           of multiple space-separated matches per item.
 7878         Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
 7879         Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
 7880         Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
 7881         Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
 7882         Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
 7883         Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
 7884         Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
 7885         Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
 7886         Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
 7887         Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
 7888         Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
 7889         Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
 7890         Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
 7891         Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
 7892         Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
 7893         Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
 7894         Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
 7895         Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
 7896         Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
 7897         Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
 7898         Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
 7899         Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
 7900         Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
 7901         Hoffmann, Zbigniew Jędrzejewski-Szmek
 7903         — Berlin, 2015-02-16
 7905 CHANGES WITH 218:
 7907         * When querying unit file enablement status (for example via
 7908           "systemctl is-enabled"), a new state "indirect" is now known
 7909           which indicates that a unit might not be enabled itself, but
 7910           another unit listed in its Also= setting might be.
 7912         * Similar to the various existing ConditionXYZ= settings for
 7913           units, there are now matching AssertXYZ= settings. While
 7914           failing conditions cause a unit to be skipped, but its job
 7915           to succeed, failing assertions declared like this will cause
 7916           a unit start operation and its job to fail.
 7918         * hostnamed now knows a new chassis type "embedded".
 7920         * systemctl gained a new "edit" command. When used on a unit
 7921           file, this allows extending unit files with .d/ drop-in
 7922           configuration snippets or editing the full file (after
 7923           copying it from /usr/lib to /etc). This will invoke the
 7924           user's editor (as configured with $EDITOR), and reload the
 7925           modified configuration after editing.
 7927         * "systemctl status" now shows the suggested enablement state
 7928           for a unit, as declared in the (usually vendor-supplied)
 7929           system preset files.
 7931         * nss-myhostname will now resolve the single-label hostname
 7932           "gateway" to the locally configured default IP routing
 7933           gateways, ordered by their metrics. This assigns a stable
 7934           name to the used gateways, regardless which ones are
 7935           currently configured. Note that the name will only be
 7936           resolved after all other name sources (if nss-myhostname is
 7937           configured properly) and should hence not negatively impact
 7938           systems that use the single-label hostname "gateway" in
 7939           other contexts.
 7941         * systemd-inhibit now allows filtering by mode when listing
 7942           inhibitors.
 7944         * Scope and service units gained a new "Delegate" boolean
 7945           property, which, when set, allows processes running inside the
 7946           unit to further partition resources. This is primarily
 7947           useful for systemd user instances as well as container
 7948           managers.
 7950         * journald will now pick up audit messages directly from
 7951           the kernel, and log them like any other log message. The
 7952           audit fields are split up and fully indexed. This means that
 7953           journalctl in many ways is now a (nicer!) alternative to
 7954           ausearch, the traditional audit client. Note that this
 7955           implements only a minimal audit client. If you want the
 7956           special audit modes like reboot-on-log-overflow, please use
 7957           the traditional auditd instead, which can be used in
 7958           parallel to journald.
 7960         * The ConditionSecurity= unit file option now understands the
 7961           special string "audit" to check whether auditing is
 7962           available.
 7964         * journalctl gained two new commands --vacuum-size= and
 7965           --vacuum-time= to delete old journal files until the
 7966           remaining ones take up no more than the specified size on disk,
 7967           or are not older than the specified time.
 7969         * A new, native PPPoE library has been added to sd-network,
 7970           systemd's library of light-weight networking protocols. This
 7971           library will be used in a future version of networkd to
 7972           enable PPPoE communication without an external pppd daemon.
 7974         * The busctl tool now understands a new "capture" verb that
 7975           works similar to "monitor", but writes a packet capture
 7976           trace to STDOUT that can be redirected to a file which is
 7977           compatible with libcap's capture file format. This can then
 7978           be loaded in Wireshark and similar tools to inspect bus
 7979           communication.
 7981         * The busctl tool now understands a new "tree" verb that shows
 7982           the object trees of a specific service on the bus, or of all
 7983           services.
 7985         * The busctl tool now understands a new "introspect" verb that
 7986           shows all interfaces and members of objects on the bus,
 7987           including their signature and values. This is particularly
 7988           useful to get more information about bus objects shown by
 7989           the new "busctl tree" command.
 7991         * The busctl tool now understands new verbs "call",
 7992           "set-property" and "get-property" for invoking bus method
 7993           calls, setting and getting bus object properties in a
 7994           friendly way.
 7996         * busctl gained a new --augment-creds= argument that controls
 7997           whether the tool shall augment credential information it
 7998           gets from the bus with data from /proc, in a possibly
 7999           race-ful way.
 8001         * nspawn's --link-journal= switch gained two new values
 8002           "try-guest" and "try-host" that work like "guest" and
 8003           "host", but do not fail if the host has no persistent
 8004           journaling enabled. -j is now equivalent to
 8005           --link-journal=try-guest.
 8007         * macvlan network devices created by nspawn will now have
 8008           stable MAC addresses.
 8010         * A new SmackProcessLabel= unit setting has been added, which
 8011           controls the SMACK security label processes forked off by
 8012           the respective unit shall use.
 8014         * If compiled with --enable-xkbcommon, systemd-localed will
 8015           verify x11 keymap settings by compiling the given keymap. It
 8016           will spew out warnings if the compilation fails. This
 8017           requires libxkbcommon to be installed.
 8019         * When a coredump is collected, a larger number of metadata
 8020           fields is now collected and included in the journal records
 8021           created for it. More specifically, control group membership,
 8022           environment variables, memory maps, working directory,
 8023           chroot directory, /proc/$PID/status, and a list of open file
 8024           descriptors is now stored in the log entry.
 8026         * The udev hwdb now contains DPI information for mice. For
 8027           details see:
 8029           http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
 8031         * All systemd programs that read standalone configuration
 8032           files in /etc now also support a corresponding series of
 8033           .conf.d configuration directories in /etc/, /run/,
 8034           /usr/local/lib/, /usr/lib/, and (if configured with
 8035           --enable-split-usr) /lib/.  In particular, the following
 8036           configuration files now have corresponding configuration
 8037           directories: system.conf user.conf, logind.conf,
 8038           journald.conf, sleep.conf, bootchart.conf, coredump.conf,
 8039           resolved.conf, timesyncd.conf, journal-remote.conf, and
 8040           journal-upload.conf.  Note that distributions should use the
 8041           configuration directories in /usr/lib/; the directories in
 8042           /etc/ are reserved for the system administrator.
 8044         * systemd-rfkill will no longer take the rfkill device name
 8045           into account when storing rfkill state on disk, as the name
 8046           might be dynamically assigned and not stable. Instead, the
 8047           ID_PATH udev variable combined with the rfkill type (wlan,
 8048           bluetooth, …) is used.
 8050         * A new service systemd-machine-id-commit.service has been
 8051           added. When used on systems where /etc is read-only during
 8052           boot, and /etc/machine-id is not initialized (but an empty
 8053           file), this service will copy the temporary machine ID
 8054           created as replacement into /etc after the system is fully
 8055           booted up. This is useful for systems that are freshly
 8056           installed with a non-initialized machine ID, but should get
 8057           a fixed machine ID for subsequent boots.
 8059         * networkd's .netdev files now provide a large set of
 8060           configuration parameters for VXLAN devices. Similarly, the
 8061           bridge port cost parameter is now configurable in .network
 8062           files. There's also new support for configuring IP source
 8063           routing. networkd .link files gained support for a new
 8064           OriginalName= match that is useful to match against the
 8065           original interface name the kernel assigned. .network files
 8066           may include MTU= and MACAddress= fields for altering the MTU
 8067           and MAC address while being connected to a specific network
 8068           interface.
 8070         * The LUKS logic gained supported for configuring
 8071           UUID-specific key files. There's also new support for naming
 8072           LUKS device from the kernel command line, using the new
 8073           luks.name= argument.
 8075         * Timer units may now be transiently created via the bus API
 8076           (this was previously already available for scope and service
 8077           units). In addition it is now possible to create multiple
 8078           transient units at the same time with a single bus call. The
 8079           "systemd-run" tool has been updated to make use of this for
 8080           running commands on a specified time, in at(1)-style.
 8082         * tmpfiles gained support for "t" lines, for assigning
 8083           extended attributes to files. Among other uses this may be
 8084           used to assign SMACK labels to files.
 8086         Contributions from: Alin Rauta, Alison Chaiken, Andrej
 8087         Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
 8088         Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
 8089         Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
 8090         Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
 8091         Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
 8092         Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
 8093         Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
 8094         Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
 8095         Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
 8096         Michael Biebl, Michael Chapman, Michael Marineau, Michal
 8097         Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
 8098         Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
 8099         Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
 8100         Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
 8101         Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
 8102         Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
 8103         Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
 8105         — Berlin, 2014-12-10
 8107 CHANGES WITH 217:
 8109         * journalctl gained the new options -t/--identifier= to match
 8110           on the syslog identifier (aka "tag"), as well as --utc to
 8111           show log timestamps in the UTC timezone. journalctl now also
 8112           accepts -n/--lines=all to disable line capping in a pager.
 8114         * journalctl gained a new switch, --flush, that synchronously
 8115           flushes logs from /run/log/journal to /var/log/journal if
 8116           persistent storage is enabled. systemd-journal-flush.service
 8117           now waits until the operation is complete.
 8119         * Services can notify the manager before they start a reload
 8120           (by sending RELOADING=1) or shutdown (by sending
 8121           STOPPING=1). This allows the manager to track and show the
 8122           internal state of daemons and closes a race condition when
 8123           the process is still running but has closed its D-Bus
 8124           connection.
 8126         * Services with Type=oneshot do not have to have any ExecStart
 8127           commands anymore.
 8129         * User units are now loaded also from
 8130           $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
 8131           /run/systemd/user directory that was already previously
 8132           supported, but is under the control of the user.
 8134         * Job timeouts (i.e. timeouts on the time a job that is
 8135           queued stays in the run queue) can now optionally result in
 8136           immediate reboot or power-off actions (JobTimeoutAction= and
 8137           JobTimeoutRebootArgument=). This is useful on ".target"
 8138           units, to limit the maximum time a target remains
 8139           undispatched in the run queue, and to trigger an emergency
 8140           operation in such a case. This is now used by default to
 8141           turn off the system if boot-up (as defined by everything in
 8142           basic.target) hangs and does not complete for at least
 8143           15min. Also, if power-off or reboot hang for at least 30min
 8144           an immediate power-off/reboot operation is triggered. This
 8145           functionality is particularly useful to increase reliability
 8146           on embedded devices, but also on laptops which might
 8147           accidentally get powered on when carried in a backpack and
 8148           whose boot stays stuck in a hard disk encryption passphrase
 8149           question.
 8151         * systemd-logind can be configured to also handle lid switch
 8152           events even when the machine is docked or multiple displays
 8153           are attached (HandleLidSwitchDocked= option).
 8155         * A helper binary and a service have been added which can be
 8156           used to resume from hibernation in the initramfs. A
 8157           generator will parse the resume= option on the kernel
 8158           command line to trigger resume.
 8160         * A user console daemon systemd-consoled has been
 8161           added. Currently, it is a preview, and will so far open a
 8162           single terminal on each session of the user marked as
 8163           Desktop=systemd-console.
 8165         * Route metrics can be specified for DHCP routes added by
 8166           systemd-networkd.
 8168         * The SELinux context of socket-activated services can be set
 8169           from the information provided by the networking stack
 8170           (SELinuxContextFromNet= option).
 8172         * Userspace firmware loading support has been removed and
 8173           the minimum supported kernel version is thus bumped to 3.7.
 8175         * Timeout for udev workers has been increased from 1 to 3
 8176           minutes, but a warning will be printed after 1 minute to
 8177           help diagnose kernel modules that take a long time to load.
 8179         * Udev rules can now remove tags on devices with TAG-="foobar".
 8181         * systemd's readahead implementation has been removed. In many
 8182           circumstances it didn't give expected benefits even for
 8183           rotational disk drives and was becoming less relevant in the
 8184           age of SSDs. As none of the developers has been using
 8185           rotating media anymore, and nobody stepped up to actively
 8186           maintain this component of systemd it has now been removed.
 8188         * Swap units can use Options= to specify discard options.
 8189           Discard options specified for swaps in /etc/fstab are now
 8190           respected.
 8192         * Docker containers are now detected as a separate type of
 8193           virtualization.
 8195         * The Password Agent protocol gained support for queries where
 8196           the user input is shown, useful e.g. for user names.
 8197           systemd-ask-password gained a new --echo option to turn that
 8198           on.
 8200         * The default sysctl.d/ snippets will now set:
 8202                 net.core.default_qdisc = fq_codel
 8204           This selects Fair Queuing Controlled Delay as the default
 8205           queuing discipline for network interfaces. fq_codel helps
 8206           fight the network bufferbloat problem. It is believed to be
 8207           a good default with no tuning required for most workloads.
 8208           Downstream distributions may override this choice. On 10Gbit
 8209           servers that do not do forwarding, "fq" may perform better.
 8210           Systems without a good clocksource should use "pfifo_fast".
 8212         * If kdbus is enabled during build a new option BusPolicy= is
 8213           available for service units, that allows locking all service
 8214           processes into a stricter bus policy, in order to limit
 8215           access to various bus services, or even hide most of them
 8216           from the service's view entirely.
 8218         * networkctl will now show the .network and .link file
 8219           networkd has applied to a specific interface.
 8221         * sd-login gained a new API call sd_session_get_desktop() to
 8222           query which desktop environment has been selected for a
 8223           session.
 8225         * UNIX utmp support is now compile-time optional to support
 8226           legacy-free systems.
 8228         * systemctl gained two new commands "add-wants" and
 8229           "add-requires" for pulling in units from specific targets
 8230           easily.
 8232         * If the word "rescue" is specified on the kernel command line
 8233           the system will now boot into rescue mode (aka
 8234           rescue.target), which was previously available only by
 8235           specifying "1" or "systemd.unit=rescue.target" on the kernel
 8236           command line. This new kernel command line option nicely
 8237           mirrors the already existing "emergency" kernel command line
 8238           option.
 8240         * New kernel command line options mount.usr=, mount.usrflags=,
 8241           mount.usrfstype= have been added that match root=, rootflags=,
 8242           rootfstype= but allow mounting a specific file system to
 8243           /usr.
 8245         * The $NOTIFY_SOCKET is now also passed to control processes of
 8246           services, not only the main process.
 8248         * This version reenables support for fsck's -l switch. This
 8249           means at least version v2.25 of util-linux is required for
 8250           operation, otherwise dead-locks on device nodes may
 8251           occur. Again: you need to update util-linux to at least
 8252           v2.25 when updating systemd to v217.
 8254         * The "multi-seat-x" tool has been removed from systemd, as
 8255           its functionality has been integrated into X servers 1.16,
 8256           and the tool is hence redundant. It is recommended to update
 8257           display managers invoking this tool to simply invoke X
 8258           directly from now on, again.
 8260         * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
 8261           message flag has been added for all of systemd's polkit
 8262           authenticated method calls has been added. In particular this
 8263           now allows optional interactive authorization via polkit for
 8264           many of PID1's privileged operations such as unit file
 8265           enabling and disabling.
 8267         * "udevadm hwdb --update" learnt a new switch "--usr" for
 8268           placing the rebuilt hardware database in /usr instead of
 8269           /etc. When used only hardware database entries stored in
 8270           /usr will be used, and any user database entries in /etc are
 8271           ignored. This functionality is useful for vendors to ship a
 8272           pre-built database on systems where local configuration is
 8273           unnecessary or unlikely.
 8275         * Calendar time specifications in .timer units now also
 8276           understand the strings "semi-annually", "quarterly" and
 8277           "minutely" as shortcuts (in addition to the preexisting
 8278           "annually", "hourly", …).
 8280         * systemd-tmpfiles will now correctly create files in /dev
 8281           at boot which are marked for creation only at boot. It is
 8282           recommended to always create static device nodes with 'c!'
 8283           and 'b!', so that they are created only at boot and not
 8284           overwritten at runtime.
 8286         * When the watchdog logic is used for a service (WatchdogSec=)
 8287           and the watchdog timeout is hit the service will now be
 8288           terminated with SIGABRT (instead of just SIGTERM), in order
 8289           to make sure a proper coredump and backtrace is
 8290           generated. This ensures that hanging services will result in
 8291           similar coredump/backtrace behaviour as services that hit a
 8292           segmentation fault.
 8294         Contributions from: Andreas Henriksson, Andrei Borzenkov,
 8295         Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
 8296         Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
 8297         Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
 8298         Herrmann, David Sommerseth, David Strauss, Emil Renner
 8299         Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
 8300         Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
 8301         Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
 8302         Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
 8303         Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
 8304         Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
 8305         Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
 8306         Michael Marineau, Michael Olbrich, Michael Scherer, Michal
 8307         Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
 8308         Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
 8309         Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
 8310         Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
 8311         Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
 8312         Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
 8313         Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
 8314         Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
 8315         Jędrzejewski-Szmek
 8317         — Berlin, 2014-10-28
 8319 CHANGES WITH 216:
 8321         * timedated no longer reads NTP implementation unit names from
 8322           /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
 8323           implementations should add a
 8325             Conflicts=systemd-timesyncd.service
 8327           to their unit files to take over and replace systemd's NTP
 8328           default functionality.
 8330         * systemd-sysusers gained a new line type "r" for configuring
 8331           which UID/GID ranges to allocate system users/groups
 8332           from. Lines of type "u" may now add an additional column
 8333           that specifies the home directory for the system user to be
 8334           created. Also, systemd-sysusers may now optionally read user
 8335           information from STDIN instead of a file. This is useful for
 8336           invoking it from RPM preinst scriptlets that need to create
 8337           users before the first RPM file is installed since these
 8338           files might need to be owned by them. A new
 8339           %sysusers_create_inline RPM macro has been introduced to do
 8340           just that. systemd-sysusers now updates the shadow files as
 8341           well as the user/group databases, which should enhance
 8342           compatibility with certain tools like grpck.
 8344         * A number of bus APIs of PID 1 now optionally consult polkit to
 8345           permit access for otherwise unprivileged clients under certain
 8346           conditions. Note that this currently doesn't support
 8347           interactive authentication yet, but this is expected to be
 8348           added eventually, too.
 8350         * /etc/machine-info now has new fields for configuring the
 8351           deployment environment of the machine, as well as the
 8352           location of the machine. hostnamectl has been updated with
 8353           new command to update these fields.
 8355         * systemd-timesyncd has been updated to automatically acquire
 8356           NTP server information from systemd-networkd, which might
 8357           have been discovered via DHCP.
 8359         * systemd-resolved now includes a caching DNS stub resolver
 8360           and a complete LLMNR name resolution implementation. A new
 8361           NSS module "nss-resolve" has been added which can be used
 8362           instead of glibc's own "nss-dns" to resolve hostnames via
 8363           systemd-resolved. Hostnames, addresses and arbitrary RRs may
 8364           be resolved via systemd-resolved D-Bus APIs. In contrast to
 8365           the glibc internal resolver systemd-resolved is aware of
 8366           multi-homed system, and keeps DNS server and caches separate
 8367           and per-interface. Queries are sent simultaneously on all
 8368           interfaces that have DNS servers configured, in order to
 8369           properly handle VPNs and local LANs which might resolve
 8370           separate sets of domain names. systemd-resolved may acquire
 8371           DNS server information from systemd-networkd automatically,
 8372           which in turn might have discovered them via DHCP. A tool
 8373           "systemd-resolve-host" has been added that may be used to
 8374           query the DNS logic in resolved. systemd-resolved implements
 8375           IDNA and automatically uses IDNA or UTF-8 encoding depending
 8376           on whether classic DNS or LLMNR is used as transport. In the
 8377           next releases we intend to add a DNSSEC and mDNS/DNS-SD
 8378           implementation to systemd-resolved.
 8380         * A new NSS module nss-mymachines has been added, that
 8381           automatically resolves the names of all local registered
 8382           containers to their respective IP addresses.
 8384         * A new client tool "networkctl" for systemd-networkd has been
 8385           added. It currently is entirely passive and will query
 8386           networking configuration from udev, rtnetlink and networkd,
 8387           and present it to the user in a very friendly
 8388           way. Eventually, we hope to extend it to become a full
 8389           control utility for networkd.
 8391         * .socket units gained a new DeferAcceptSec= setting that
 8392           controls the kernels' TCP_DEFER_ACCEPT sockopt for
 8393           TCP. Similarly, support for controlling TCP keep-alive
 8394           settings has been added (KeepAliveTimeSec=,
 8395           KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
 8396           turning off Nagle's algorithm on TCP has been added
 8397           (NoDelay=).
 8399         * logind learned a new session type "web", for use in projects
 8400           like Cockpit which register web clients as PAM sessions.
 8402         * timer units with at least one OnCalendar= setting will now
 8403           be started only after time-sync.target has been
 8404           reached. This way they will not elapse before the system
 8405           clock has been corrected by a local NTP client or
 8406           similar. This is particular useful on RTC-less embedded
 8407           machines, that come up with an invalid system clock.
 8409         * systemd-nspawn's --network-veth= switch should now result in
 8410           stable MAC addresses for both the outer and the inner side
 8411           of the link.
 8413         * systemd-nspawn gained a new --volatile= switch for running
 8414           container instances with /etc or /var unpopulated.
 8416         * The kdbus client code has been updated to use the new Linux
 8417           3.17 memfd subsystem instead of the old kdbus-specific one.
 8419         * systemd-networkd's DHCP client and server now support
 8420           FORCERENEW. There are also new configuration options to
 8421           configure the vendor client identifier and broadcast mode
 8422           for DHCP.
 8424         * systemd will no longer inform the kernel about the current
 8425           timezone, as this is necessarily incorrect and racy as the
 8426           kernel has no understanding of DST and similar
 8427           concepts. This hence means FAT timestamps will be always
 8428           considered UTC, similar to what Android is already
 8429           doing. Also, when the RTC is configured to the local time
 8430           (rather than UTC) systemd will never synchronize back to it,
 8431           as this might confuse Windows at a later boot.
 8433         * systemd-analyze gained a new command "verify" for offline
 8434           validation of unit files.
 8436         * systemd-networkd gained support for a couple of additional
 8437           settings for bonding networking setups. Also, the metric for
 8438           statically configured routes may now be configured. For
 8439           network interfaces where this is appropriate the peer IP
 8440           address may now be configured.
 8442         * systemd-networkd's DHCP client will no longer request
 8443           broadcasting by default, as this tripped up some networks.
 8444           For hardware where broadcast is required the feature should
 8445           be switched back on using RequestBroadcast=yes.
 8447         * systemd-networkd will now set up IPv4LL addresses (when
 8448           enabled) even if DHCP is configured successfully.
 8450         * udev will now default to respect network device names given
 8451           by the kernel when the kernel indicates that these are
 8452           predictable. This behavior can be tweaked by changing
 8453           NamePolicy= in the relevant .link file.
 8455         * A new library systemd-terminal has been added that
 8456           implements full TTY stream parsing and rendering. This
 8457           library is supposed to be used later on for implementing a
 8458           full userspace VT subsystem, replacing the current kernel
 8459           implementation.
 8461         * A new tool systemd-journal-upload has been added to push
 8462           journal data to a remote system running
 8463           systemd-journal-remote.
 8465         * journald will no longer forward all local data to another
 8466           running syslog daemon. This change has been made because
 8467           rsyslog (which appears to be the most commonly used syslog
 8468           implementation these days) no longer makes use of this, and
 8469           instead pulls the data out of the journal on its own. Since
 8470           forwarding the messages to a non-existent syslog server is
 8471           more expensive than we assumed we have now turned this
 8472           off. If you run a syslog server that is not a recent rsyslog
 8473           version, you have to turn this option on again
 8474           (ForwardToSyslog= in journald.conf).
 8476         * journald now optionally supports the LZ4 compressor for
 8477           larger journal fields. This compressor should perform much
 8478           better than XZ which was the previous default.
 8480         * machinectl now shows the IP addresses of local containers,
 8481           if it knows them, plus the interface name of the container.
 8483         * A new tool "systemd-escape" has been added that makes it
 8484           easy to escape strings to build unit names and similar.
 8486         * sd_notify() messages may now include a new ERRNO= field
 8487           which is parsed and collected by systemd and shown among the
 8488           "systemctl status" output for a service.
 8490         * A new component "systemd-firstboot" has been added that
 8491           queries the most basic systemd information (timezone,
 8492           hostname, root password) interactively on first
 8493           boot. Alternatively it may also be used to provision these
 8494           things offline on OS images installed into directories.
 8496         * The default sysctl.d/ snippets will now set
 8498                 net.ipv4.conf.default.promote_secondaries=1
 8500           This has the benefit of no flushing secondary IP addresses
 8501           when primary addresses are removed.
 8503         Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
 8504         Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
 8505         Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
 8506         Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
 8507         Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
 8508         B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
 8509         Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
 8510         Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
 8511         Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
 8512         Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
 8513         Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
 8514         Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
 8515         Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
 8516         Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
 8517         Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
 8519         — Berlin, 2014-08-19
 8521 CHANGES WITH 215:
 8523         * A new tool systemd-sysusers has been added. This tool
 8524           creates system users and groups in /etc/passwd and
 8525           /etc/group, based on static declarative system user/group
 8526           definitions in /usr/lib/sysusers.d/. This is useful to
 8527           enable factory resets and volatile systems that boot up with
 8528           an empty /etc directory, and thus need system users and
 8529           groups created during early boot. systemd now also ships
 8530           with two default sysusers.d/ files for the most basic
 8531           users and groups systemd and the core operating system
 8532           require.
 8534         * A new tmpfiles snippet has been added that rebuilds the
 8535           essential files in /etc on boot, should they be missing.
 8537         * A directive for ensuring automatic clean-up of
 8538           /var/cache/man/ has been removed from the default
 8539           configuration. This line should now be shipped by the man
 8540           implementation. The necessary change has been made to the
 8541           man-db implementation. Note that you need to update your man
 8542           implementation to one that ships this line, otherwise no
 8543           automatic clean-up of /var/cache/man will take place.
 8545         * A new condition ConditionNeedsUpdate= has been added that
 8546           may conditionalize services to only run when /etc or /var
 8547           are "older" than the vendor operating system resources in
 8548           /usr. This is useful for reconstructing or updating /etc
 8549           after an offline update of /usr or a factory reset, on the
 8550           next reboot. Services that want to run once after such an
 8551           update or reset should use this condition and order
 8552           themselves before the new systemd-update-done.service, which
 8553           will mark the two directories as fully updated. A number of
 8554           service files have been added making use of this, to rebuild
 8555           the udev hardware database, the journald message catalog and
 8556           dynamic loader cache (ldconfig). The systemd-sysusers tool
 8557           described above also makes use of this now. With this in
 8558           place it is now possible to start up a minimal operating
 8559           system with /etc empty cleanly. For more information on the
 8560           concepts involved see this recent blog story:
 8562           http://0pointer.de/blog/projects/stateless.html
 8564         * A new system group "input" has been introduced, and all
 8565           input device nodes get this group assigned. This is useful
 8566           for system-level software to get access to input devices. It
 8567           complements what is already done for "audio" and "video".
 8569         * systemd-networkd learnt minimal DHCPv4 server support in
 8570           addition to the existing DHCPv4 client support. It also
 8571           learnt DHCPv6 client and IPv6 Router Solicitation client
 8572           support. The DHCPv4 client gained support for static routes
 8573           passed in from the server. Note that the [DHCPv4] section
 8574           known in older systemd-networkd versions has been renamed to
 8575           [DHCP] and is now also used by the DHCPv6 client. Existing
 8576           .network files using settings of this section should be
 8577           updated, though compatibility is maintained. Optionally, the
 8578           client hostname may now be sent to the DHCP server.
 8580         * networkd gained support for vxlan virtual networks as well
 8581           as tun/tap and dummy devices.
 8583         * networkd gained support for automatic allocation of address
 8584           ranges for interfaces from a system-wide pool of
 8585           addresses. This is useful for dynamically managing a large
 8586           number of interfaces with a single network configuration
 8587           file. In particular this is useful to easily assign
 8588           appropriate IP addresses to the veth links of a large number
 8589           of nspawn instances.
 8591         * RPM macros for processing sysusers, sysctl and binfmt
 8592           drop-in snippets at package installation time have been
 8593           added.
 8595         * The /etc/os-release file should now be placed in
 8596           /usr/lib/os-release. The old location is automatically
 8597           created as symlink. /usr/lib is the more appropriate
 8598           location of this file, since it shall actually describe the
 8599           vendor operating system shipped in /usr, and not the
 8600           configuration stored in /etc.
 8602         * .mount units gained a new boolean SloppyOptions= setting
 8603           that maps to mount(8)'s -s option which enables permissive
 8604           parsing of unknown mount options.
 8606         * tmpfiles learnt a new "L+" directive which creates a symlink
 8607           but (unlike "L") deletes a pre-existing file first, should
 8608           it already exist and not already be the correct
 8609           symlink. Similarly, "b+", "c+" and "p+" directives have been
 8610           added as well, which create block and character devices, as
 8611           well as fifos in the filesystem, possibly removing any
 8612           pre-existing files of different types.
 8614         * For tmpfiles' "L", "L+", "C" and "C+" directives the final
 8615           'argument' field (which so far specified the source to
 8616           symlink/copy the files from) is now optional. If omitted the
 8617           same file os copied from /usr/share/factory/ suffixed by the
 8618           full destination path. This is useful for populating /etc
 8619           with essential files, by copying them from vendor defaults
 8620           shipped in /usr/share/factory/etc.
 8622         * A new command "systemctl preset-all" has been added that
 8623           applies the service preset settings to all installed unit
 8624           files. A new switch --preset-mode= has been added that
 8625           controls whether only enable or only disable operations
 8626           shall be executed.
 8628         * A new command "systemctl is-system-running" has been added
 8629           that allows checking the overall state of the system, for
 8630           example whether it is fully up and running.
 8632         * When the system boots up with an empty /etc, the equivalent
 8633           to "systemctl preset-all" is executed during early boot, to
 8634           make sure all default services are enabled after a factory
 8635           reset.
 8637         * systemd now contains a minimal preset file that enables the
 8638           most basic services systemd ships by default.
 8640         * Unit files' [Install] section gained a new DefaultInstance=
 8641           field for defining the default instance to create if a
 8642           template unit is enabled with no instance specified.
 8644         * A new passive target cryptsetup-pre.target has been added
 8645           that may be used by services that need to make they run and
 8646           finish before the first LUKS cryptographic device is set up.
 8648         * The /dev/loop-control and /dev/btrfs-control device nodes
 8649           are now owned by the "disk" group by default, opening up
 8650           access to this group.
 8652         * systemd-coredump will now automatically generate a
 8653           stack trace of all core dumps taking place on the system,
 8654           based on elfutils' libdw library. This stack trace is logged
 8655           to the journal.
 8657         * systemd-coredump may now optionally store coredumps directly
 8658           on disk (in /var/lib/systemd/coredump, possibly compressed),
 8659           instead of storing them unconditionally in the journal. This
 8660           mode is the new default. A new configuration file
 8661           /etc/systemd/coredump.conf has been added to configure this
 8662           and other parameters of systemd-coredump.
 8664         * coredumpctl gained a new "info" verb to show details about a
 8665           specific coredump. A new switch "-1" has also been added
 8666           that makes sure to only show information about the most
 8667           recent entry instead of all entries. Also, as the tool is
 8668           generally useful now the "systemd-" prefix of the binary
 8669           name has been removed. Distributions that want to maintain
 8670           compatibility with the old name should add a symlink from
 8671           the old name to the new name.
 8673         * journald's SplitMode= now defaults to "uid". This makes sure
 8674           that unprivileged users can access their own coredumps with
 8675           coredumpctl without restrictions.
 8677         * New kernel command line options "systemd.wants=" (for
 8678           pulling an additional unit during boot), "systemd.mask="
 8679           (for masking a specific unit for the boot), and
 8680           "systemd.debug-shell" (for enabling the debug shell on tty9)
 8681           have been added. This is implemented in the new generator
 8682           "systemd-debug-generator".
 8684         * systemd-nspawn will now by default filter a couple of
 8685           syscalls for containers, among them those required for
 8686           kernel module loading, direct x86 IO port access, swap
 8687           management, and kexec. Most importantly though
 8688           open_by_handle_at() is now prohibited for containers,
 8689           closing a hole similar to a recently discussed vulnerability
 8690           in docker regarding access to files on file hierarchies the
 8691           container should normally not have access to. Note that, for
 8692           nspawn, we generally make no security claims anyway (and
 8693           this is explicitly documented in the man page), so this is
 8694           just a fix for one of the most obvious problems.
 8696         * A new man page file-hierarchy(7) has been added that
 8697           contains a minimized, modernized version of the file system
 8698           layout systemd expects, similar in style to the FHS
 8699           specification or hier(5). A new tool systemd-path(1) has
 8700           been added to query many of these paths for the local
 8701           machine and user.
 8703         * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
 8704           longer done. Since the directory now has a per-user size
 8705           limit, and is cleaned on logout this appears unnecessary,
 8706           in particular since this now brings the lifecycle of this
 8707           directory closer in line with how IPC objects are handled.
 8709         * systemd.pc now exports a number of additional directories,
 8710           including $libdir (which is useful to identify the library
 8711           path for the primary architecture of the system), and a
 8712           couple of drop-in directories.
 8714         * udev's predictable network interface names now use the dev_port
 8715           sysfs attribute, introduced in linux 3.15 instead of dev_id to
 8716           distinguish between ports of the same PCI function. dev_id should
 8717           only be used for ports using the same HW address, hence the need
 8718           for dev_port.
 8720         * machined has been updated to export the OS version of a
 8721           container (read from /etc/os-release and
 8722           /usr/lib/os-release) on the bus. This is now shown in
 8723           "machinectl status" for a machine.
 8725         * A new service setting RestartForceExitStatus= has been
 8726           added. If configured to a set of exit signals or process
 8727           return values, the service will be restarted when the main
 8728           daemon process exits with any of them, regardless of the
 8729           Restart= setting.
 8731         * systemctl's -H switch for connecting to remote systemd
 8732           machines has been extended so that it may be used to
 8733           directly connect to a specific container on the
 8734           host. "systemctl -H root@foobar:waldi" will now connect as
 8735           user "root" to host "foobar", and then proceed directly to
 8736           the container named "waldi". Note that currently you have to
 8737           authenticate as user "root" for this to work, as entering
 8738           containers is a privileged operation.
 8740         Contributions from: Andreas Henriksson, Benjamin Steinwender,
 8741         Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
 8742         Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
 8743         Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
 8744         Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
 8745         Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
 8746         Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
 8747         Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
 8748         Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
 8749         Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
 8750         Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
 8751         Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
 8753         — Berlin, 2014-07-03
 8755 CHANGES WITH 214:
 8757         * As an experimental feature, udev now tries to lock the
 8758           disk device node (flock(LOCK_SH|LOCK_NB)) while it
 8759           executes events for the disk or any of its partitions.
 8760           Applications like partitioning programs can lock the
 8761           disk device node (flock(LOCK_EX)) and claim temporary
 8762           device ownership that way; udev will entirely skip all event
 8763           handling for this disk and its partitions. If the disk
 8764           was opened for writing, the close will trigger a partition
 8765           table rescan in udev's "watch" facility, and if needed
 8766           synthesize "change" events for the disk and all its partitions.
 8767           This is now unconditionally enabled, and if it turns out to
 8768           cause major problems, we might turn it on only for specific
 8769           devices, or might need to disable it entirely. Device Mapper
 8770           devices are excluded from this logic.
 8772         * We temporarily dropped the "-l" switch for fsck invocations,
 8773           since they collide with the flock() logic above. util-linux
 8774           upstream has been changed already to avoid this conflict,
 8775           and we will re-add "-l" as soon as util-linux with this
 8776           change has been released.
 8778         * The dependency on libattr has been removed. Since a long
 8779           time, the extended attribute calls have moved to glibc, and
 8780           libattr is thus unnecessary.
 8782         * Virtualization detection works without privileges now. This
 8783           means the systemd-detect-virt binary no longer requires
 8784           CAP_SYS_PTRACE file capabilities, and our daemons can run
 8785           with fewer privileges.
 8787         * systemd-networkd now runs under its own "systemd-network"
 8788           user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
 8789           CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
 8790           loses the ability to write to files owned by root this way.
 8792         * Similarly, systemd-resolved now runs under its own
 8793           "systemd-resolve" user with no capabilities remaining.
 8795         * Similarly, systemd-bus-proxyd now runs under its own
 8796           "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
 8798         * systemd-networkd gained support for setting up "veth"
 8799           virtual Ethernet devices for container connectivity, as well
 8800           as GRE and VTI tunnels.
 8802         * systemd-networkd will no longer automatically attempt to
 8803           manually load kernel modules necessary for certain tunnel
 8804           transports. Instead, it is assumed the kernel loads them
 8805           automatically when required. This only works correctly on
 8806           very new kernels. On older kernels, please consider adding
 8807           the kernel modules to /etc/modules-load.d/ as a work-around.
 8809         * The resolv.conf file systemd-resolved generates has been
 8810           moved to /run/systemd/resolve/. If you have a symlink from
 8811           /etc/resolv.conf, it might be necessary to correct it.
 8813         * Two new service settings, ProtectHome= and ProtectSystem=,
 8814           have been added. When enabled, they will make the user data
 8815           (such as /home) inaccessible or read-only and the system
 8816           (such as /usr) read-only, for specific services. This allows
 8817           very light-weight per-service sandboxing to avoid
 8818           modifications of user data or system files from
 8819           services. These two new switches have been enabled for all
 8820           of systemd's long-running services, where appropriate.
 8822         * Socket units gained new SocketUser= and SocketGroup=
 8823           settings to set the owner user and group of AF_UNIX sockets
 8824           and FIFOs in the file system.
 8826         * Socket units gained a new RemoveOnStop= setting. If enabled,
 8827           all FIFOS and sockets in the file system will be removed
 8828           when the specific socket unit is stopped.
 8830         * Socket units gained a new Symlinks= setting. It takes a list
 8831           of symlinks to create to file system sockets or FIFOs
 8832           created by the specific Unix sockets. This is useful to
 8833           manage symlinks to socket nodes with the same lifecycle as
 8834           the socket itself.
 8836         * The /dev/log socket and /dev/initctl FIFO have been moved to
 8837           /run, and have been replaced by symlinks. This allows
 8838           connecting to these facilities even if PrivateDevices=yes is
 8839           used for a service (which makes /dev/log itself unavailable,
 8840           but /run is left). This also has the benefit of ensuring
 8841           that /dev only contains device nodes, directories and
 8842           symlinks, and nothing else.
 8844         * sd-daemon gained two new calls sd_pid_notify() and
 8845           sd_pid_notifyf(). They are similar to sd_notify() and
 8846           sd_notifyf(), but allow overriding of the source PID of
 8847           notification messages if permissions permit this. This is
 8848           useful to send notify messages on behalf of a different
 8849           process (for example, the parent process). The
 8850           systemd-notify tool has been updated to make use of this
 8851           when sending messages (so that notification messages now
 8852           originate from the shell script invoking systemd-notify and
 8853           not the systemd-notify process itself. This should minimize
 8854           a race where systemd fails to associate notification
 8855           messages to services when the originating process already
 8856           vanished.
 8858         * A new "on-abnormal" setting for Restart= has been added. If
 8859           set, it will result in automatic restarts on all "abnormal"
 8860           reasons for a process to exit, which includes unclean
 8861           signals, core dumps, timeouts and watchdog timeouts, but
 8862           does not include clean and unclean exit codes or clean
 8863           signals. Restart=on-abnormal is an alternative for
 8864           Restart=on-failure for services that shall be able to
 8865           terminate and avoid restarts on certain errors, by
 8866           indicating so with an unclean exit code. Restart=on-failure
 8867           or Restart=on-abnormal is now the recommended setting for
 8868           all long-running services.
 8870         * If the InaccessibleDirectories= service setting points to a
 8871           mount point (or if there are any submounts contained within
 8872           it), it is now attempted to completely unmount it, to make
 8873           the file systems truly unavailable for the respective
 8874           service.
 8876         * The ReadOnlyDirectories= service setting and
 8877           systemd-nspawn's --read-only parameter are now recursively
 8878           applied to all submounts, too.
 8880         * Mount units may now be created transiently via the bus APIs.
 8882         * The support for SysV and LSB init scripts has been removed
 8883           from the systemd daemon itself. Instead, it is now
 8884           implemented as a generator that creates native systemd units
 8885           from these scripts when needed. This enables us to remove a
 8886           substantial amount of legacy code from PID 1, following the
 8887           fact that many distributions only ship a very small number
 8888           of LSB/SysV init scripts nowadays.
 8890         * Privileged Xen (dom0) domains are not considered
 8891           virtualization anymore by the virtualization detection
 8892           logic. After all, they generally have unrestricted access to
 8893           the hardware and usually are used to manage the unprivileged
 8894           (domU) domains.
 8896         * systemd-tmpfiles gained a new "C" line type, for copying
 8897           files or entire directories.
 8899         * systemd-tmpfiles "m" lines are now fully equivalent to "z"
 8900           lines. So far, they have been non-globbing versions of the
 8901           latter, and have thus been redundant. In future, it is
 8902           recommended to only use "z". "m" has hence been removed
 8903           from the documentation, even though it stays supported.
 8905         * A tmpfiles snippet to recreate the most basic structure in
 8906           /var has been added. This is enough to create the /var/run →
 8907           /run symlink and create a couple of structural
 8908           directories. This allows systems to boot up with an empty or
 8909           volatile /var. Of course, while with this change, the core OS
 8910           now is capable with dealing with a volatile /var, not all
 8911           user services are ready for it. However, we hope that sooner
 8912           or later, many service daemons will be changed upstream so
 8913           that they are able to automatically create their necessary
 8914           directories in /var at boot, should they be missing. This is
 8915           the first step to allow state-less systems that only require
 8916           the vendor image for /usr to boot.
 8918         * systemd-nspawn has gained a new --tmpfs= switch to mount an
 8919           empty tmpfs instance to a specific directory. This is
 8920           particularly useful for making use of the automatic
 8921           reconstruction of /var (see above), by passing --tmpfs=/var.
 8923         * Access modes specified in tmpfiles snippets may now be
 8924           prefixed with "~", which indicates that they shall be masked
 8925           by whether the existing file or directory is currently
 8926           writable, readable or executable at all. Also, if specified,
 8927           the sgid/suid/sticky bits will be masked for all
 8928           non-directories.
 8930         * A new passive target unit "network-pre.target" has been
 8931           added which is useful for services that shall run before any
 8932           network is configured, for example firewall scripts.
 8934         * The "floppy" group that previously owned the /dev/fd*
 8935           devices is no longer used. The "disk" group is now used
 8936           instead. Distributions should probably deprecate usage of
 8937           this group.
 8939         Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
 8940         King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
 8941         Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
 8942         Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
 8943         Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
 8944         Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
 8945         Jędrzejewski-Szmek
 8947         — Berlin, 2014-06-11
 8949 CHANGES WITH 213:
 8951         * A new "systemd-timesyncd" daemon has been added for
 8952           synchronizing the system clock across the network. It
 8953           implements an SNTP client. In contrast to NTP
 8954           implementations such as chrony or the NTP reference server,
 8955           this only implements a client side, and does not bother with
 8956           the full NTP complexity, focusing only on querying time from
 8957           one remote server and synchronizing the local clock to
 8958           it. Unless you intend to serve NTP to networked clients or
 8959           want to connect to local hardware clocks, this simple NTP
 8960           client should be more than appropriate for most
 8961           installations. The daemon runs with minimal privileges, and
 8962           has been hooked up with networkd to only operate when
 8963           network connectivity is available. The daemon saves the
 8964           current clock to disk every time a new NTP sync has been
 8965           acquired, and uses this to possibly correct the system clock
 8966           early at bootup, in order to accommodate for systems that
 8967           lack an RTC such as the Raspberry Pi and embedded devices,
 8968           and to make sure that time monotonically progresses on these
 8969           systems, even if it is not always correct. To make use of
 8970           this daemon, a new system user and group "systemd-timesync"
 8971           needs to be created on installation of systemd.
 8973         * The queue "seqnum" interface of libudev has been disabled, as
 8974           it was generally incompatible with device namespacing as
 8975           sequence numbers of devices go "missing" if the devices are
 8976           part of a different namespace.
 8978         * "systemctl list-timers" and "systemctl list-sockets" gained
 8979           a --recursive switch for showing units of these types also
 8980           for all local containers, similar in style to the already
 8981           supported --recursive switch for "systemctl list-units".
 8983         * A new RebootArgument= setting has been added for service
 8984           units, which may be used to specify a kernel reboot argument
 8985           to use when triggering reboots with StartLimitAction=.
 8987         * A new FailureAction= setting has been added for service
 8988           units which may be used to specify an operation to trigger
 8989           when a service fails. This works similarly to
 8990           StartLimitAction=, but unlike it, controls what is done
 8991           immediately rather than only after several attempts to
 8992           restart the service in question.
 8994         * hostnamed got updated to also expose the kernel name,
 8995           release, and version on the bus. This is useful for
 8996           executing commands like hostnamectl with the -H switch.
 8997           systemd-analyze makes use of this to properly display
 8998           details when running non-locally.
 9000         * The bootchart tool can now show cgroup information in the
 9001           graphs it generates.
 9003         * The CFS CPU quota cgroup attribute is now exposed for
 9004           services. The new CPUQuota= switch has been added for this
 9005           which takes a percentage value. Setting this will have the
 9006           result that a service may never get more CPU time than the
 9007           specified percentage, even if the machine is otherwise idle.
 9009         * systemd-networkd learned IPIP and SIT tunnel support.
 9011         * LSB init scripts exposing a dependency on $network will now
 9012           get a dependency on network-online.target rather than simply
 9013           network.target. This should bring LSB handling closer to
 9014           what it was on SysV systems.
 9016         * A new fsck.repair= kernel option has been added to control
 9017           how fsck shall deal with unclean file systems at boot.
 9019         * The (.ini) configuration file parser will now silently ignore
 9020           sections whose names begin with "X-". This may be used to maintain
 9021           application-specific extension sections in unit files.
 9023         * machined gained a new API to query the IP addresses of
 9024           registered containers. "machinectl status" has been updated
 9025           to show these addresses in its output.
 9027         * A new call sd_uid_get_display() has been added to the
 9028           sd-login APIs for querying the "primary" session of a
 9029           user. The "primary" session of the user is elected from the
 9030           user's sessions and generally a graphical session is
 9031           preferred over a text one.
 9033         * A minimal systemd-resolved daemon has been added. It
 9034           currently simply acts as a companion to systemd-networkd and
 9035           manages resolv.conf based on per-interface DNS
 9036           configuration, possibly supplied via DHCP. In the long run
 9037           we hope to extend this into a local DNSSEC enabled DNS and
 9038           mDNS cache.
 9040         * The systemd-networkd-wait-online tool is now enabled by
 9041           default. It will delay network-online.target until a network
 9042           connection has been configured. The tool primarily integrates
 9043           with networkd, but will also make a best effort to make sense
 9044           of network configuration performed in some other way.
 9046         * Two new service options StartupCPUShares= and
 9047           StartupBlockIOWeight= have been added that work similarly to
 9048           CPUShares= and BlockIOWeight= however only apply during
 9049           system startup. This is useful to prioritize certain services
 9050           differently during bootup than during normal runtime.
 9052         * hostnamed has been changed to prefer the statically
 9053           configured hostname in /etc/hostname (unless set to
 9054           'localhost' or empty) over any dynamic one supplied by
 9055           dhcp. With this change, the rules for picking the hostname
 9056           match more closely the rules of other configuration settings
 9057           where the local administrator's configuration in /etc always
 9058           overrides any other settings.
 9060         Contributions from: Ali H. Caliskan, Alison Chaiken, Bas van
 9061         den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
 9062         Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
 9063         David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
 9064         Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
 9065         Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
 9066         Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
 9067         Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
 9068         Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
 9069         Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
 9070         Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
 9071         Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
 9072         Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
 9073         Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
 9074         Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
 9075         Lindskog, WaLyong Cho, Will Woods, Zbigniew
 9076         Jędrzejewski-Szmek
 9078         — Beijing, 2014-05-28
 9080 CHANGES WITH 212:
 9082         * When restoring the screen brightness at boot, stay away from
 9083           the darkest setting or from the lowest 5% of the available
 9084           range, depending on which is the larger value of both. This
 9085           should effectively protect the user from rebooting into a
 9086           black screen, should the brightness have been set to minimum
 9087           by accident.
 9089         * sd-login gained a new sd_machine_get_class() call to
 9090           determine the class ("vm" or "container") of a machine
 9091           registered with machined.
 9093         * sd-login gained new calls
 9094           sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
 9095           to query the identity of the peer of a local AF_UNIX
 9096           connection. They operate similarly to their sd_pid_get_xyz()
 9097           counterparts.
 9099         * PID 1 will now maintain a system-wide system state engine
 9100           with the states "starting", "running", "degraded",
 9101           "maintenance", "stopping". These states are bound to system
 9102           startup, normal runtime, runtime with at least one failed
 9103           service, rescue/emergency mode and system shutdown. This
 9104           state is shown in the "systemctl status" output when no unit
 9105           name is passed. It is useful to determine system state, in
 9106           particularly when doing so for many systems or containers at
 9107           once.
 9109         * A new command "list-machines" has been added to "systemctl"
 9110           that lists all local OS containers and shows their system
 9111           state (see above), if systemd runs inside of them.
 9113         * systemctl gained a new "-r" switch to recursively enumerate
 9114           units on all local containers, when used with the
 9115           "list-unit" command (which is the default one that is
 9116           executed when no parameters are specified).
 9118         * The GPT automatic partition discovery logic will now honour
 9119           two GPT partition flags: one may be set on a partition to
 9120           cause it to be mounted read-only, and the other may be set
 9121           on a partition to ignore it during automatic discovery.
 9123         * Two new GPT type UUIDs have been added for automatic root
 9124           partition discovery, for 32-bit and 64-bit ARM. This is not
 9125           particularly useful for discovering the root directory on
 9126           these architectures during bare-metal boots (since UEFI is
 9127           not common there), but still very useful to allow booting of
 9128           ARM disk images in nspawn with the -i option.
 9130         * MAC addresses of interfaces created with nspawn's
 9131           --network-interface= switch will now be generated from the
 9132           machine name, and thus be stable between multiple invocations
 9133           of the container.
 9135         * logind will now automatically remove all IPC objects owned
 9136           by a user if she or he fully logs out. This makes sure that
 9137           users who are logged out cannot continue to consume IPC
 9138           resources. This covers SysV memory, semaphores and message
 9139           queues as well as POSIX shared memory and message
 9140           queues. Traditionally, SysV and POSIX IPC had no lifecycle
 9141           limits. With this functionality, that is corrected. This may
 9142           be turned off by using the RemoveIPC= switch of logind.conf.
 9144         * The systemd-machine-id-setup and tmpfiles tools gained a
 9145           --root= switch to operate on a specific root directory,
 9146           instead of /.
 9148         * journald can now forward logged messages to the TTYs of all
 9149           logged in users ("wall"). This is the default for all
 9150           emergency messages now.
 9152         * A new tool systemd-journal-remote has been added to stream
 9153           journal log messages across the network.
 9155         * /sys/fs/cgroup/ is now mounted read-only after all cgroup
 9156           controller trees are mounted into it. Note that the
 9157           directories mounted beneath it are not read-only. This is a
 9158           security measure and is particularly useful because glibc
 9159           actually includes a search logic to pick any tmpfs it can
 9160           find to implement shm_open() if /dev/shm is not available
 9161           (which it might very well be in namespaced setups).
 9163         * machinectl gained a new "poweroff" command to cleanly power
 9164           down a local OS container.
 9166         * The PrivateDevices= unit file setting will now also drop the
 9167           CAP_MKNOD capability from the capability bound set, and
 9168           imply DevicePolicy=closed.
 9170         * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
 9171           comprehensively on all long-running systemd services where
 9172           this is appropriate.
 9174         * systemd-udevd will now run in a disassociated mount
 9175           namespace. To mount directories from udev rules, make sure to
 9176           pull in mount units via SYSTEMD_WANTS properties.
 9178         * The kdbus support gained support for uploading policy into
 9179           the kernel. sd-bus gained support for creating "monitoring"
 9180           connections that can eavesdrop into all bus communication
 9181           for debugging purposes.
 9183         * Timestamps may now be specified in seconds since the UNIX
 9184           epoch Jan 1st, 1970 by specifying "@" followed by the value
 9185           in seconds.
 9187         * Native tcpwrap support in systemd has been removed. tcpwrap
 9188           is old code, not really maintained anymore and has serious
 9189           shortcomings, and better options such as firewalls
 9190           exist. For setups that require tcpwrap usage, please
 9191           consider invoking your socket-activated service via tcpd,
 9192           like on traditional inetd.
 9194         * A new system.conf configuration option
 9195           DefaultTimerAccuracySec= has been added that controls the
 9196           default AccuracySec= setting of .timer units.
 9198         * Timer units gained a new WakeSystem= switch. If enabled,
 9199           timers configured this way will cause the system to resume
 9200           from system suspend (if the system supports that, which most
 9201           do these days).
 9203         * Timer units gained a new Persistent= switch. If enabled,
 9204           timers configured this way will save to disk when they have
 9205           been last triggered. This information is then used on next
 9206           reboot to possible execute overdue timer events, that
 9207           could not take place because the system was powered off.
 9208           This enables simple anacron-like behaviour for timer units.
 9210         * systemctl's "list-timers" will now also list the time a
 9211           timer unit was last triggered in addition to the next time
 9212           it will be triggered.
 9214         * systemd-networkd will now assign predictable IPv4LL
 9215           addresses to its local interfaces.
 9217         Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
 9218         Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
 9219         Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
 9220         Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
 9221         Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
 9222         Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
 9223         Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
 9224         Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
 9225         Jędrzejewski-Szmek
 9227         — Berlin, 2014-03-25
 9229 CHANGES WITH 211:
 9231         * A new unit file setting RestrictAddressFamilies= has been
 9232           added to restrict which socket address families unit
 9233           processes gain access to. This takes address family names
 9234           like "AF_INET" or "AF_UNIX", and is useful to minimize the
 9235           attack surface of services via exotic protocol stacks. This
 9236           is built on seccomp system call filters.
 9238         * Two new unit file settings RuntimeDirectory= and
 9239           RuntimeDirectoryMode= have been added that may be used to
 9240           manage a per-daemon runtime directories below /run. This is
 9241           an alternative for setting up directory permissions with
 9242           tmpfiles snippets, and has the advantage that the runtime
 9243           directory's lifetime is bound to the daemon runtime and that
 9244           the daemon starts up with an empty directory each time. This
 9245           is particularly useful when writing services that drop
 9246           privileges using the User= or Group= setting.
 9248         * The DeviceAllow= unit setting now supports globbing for
 9249           matching against device group names.
 9251         * The systemd configuration file system.conf gained new
 9252           settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
 9253           DefaultMemoryAccounting= to globally turn on/off accounting
 9254           for specific resources (cgroups) for all units. These
 9255           settings may still be overridden individually in each unit
 9256           though.
 9258         * systemd-gpt-auto-generator is now able to discover /srv and
 9259           root partitions in addition to /home and swap partitions. It
 9260           also supports LUKS-encrypted partitions now. With this in
 9261           place, automatic discovery of partitions to mount following
 9262           the Discoverable Partitions Specification
 9263           (https://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
 9264           is now a lot more complete. This allows booting without
 9265           /etc/fstab and without root= on the kernel command line on
 9266           systems prepared appropriately.
 9268         * systemd-nspawn gained a new --image= switch which allows
 9269           booting up disk images and Linux installations on any block
 9270           device that follow the Discoverable Partitions Specification
 9271           (see above). This means that installations made with
 9272           appropriately updated installers may now be started and
 9273           deployed using container managers, completely
 9274           unmodified. (We hope that libvirt-lxc will add support for
 9275           this feature soon, too.)
 9277         * systemd-nspawn gained a new --network-macvlan= setting to
 9278           set up a private macvlan interface for the
 9279           container. Similarly, systemd-networkd gained a new
 9280           Kind=macvlan setting in .netdev files.
 9282         * systemd-networkd now supports configuring local addresses
 9283           using IPv4LL.
 9285         * A new tool systemd-network-wait-online has been added to
 9286           synchronously wait for network connectivity using
 9287           systemd-networkd.
 9289         * The sd-bus.h bus API gained a new sd_bus_track object for
 9290           tracking the lifecycle of bus peers. Note that sd-bus.h is
 9291           still not a public API though (unless you specify
 9292           --enable-kdbus on the configure command line, which however
 9293           voids your warranty and you get no API stability guarantee).
 9295         * The $XDG_RUNTIME_DIR runtime directories for each user are
 9296           now individual tmpfs instances, which has the benefit of
 9297           introducing separate pools for each user, with individual
 9298           size limits, and thus making sure that unprivileged clients
 9299           can no longer negatively impact the system or other users by
 9300           filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
 9301           RuntimeDirectorySize= has been introduced that allows
 9302           controlling the default size limit for all users. It
 9303           defaults to 10% of the available physical memory. This is no
 9304           replacement for quotas on tmpfs though (which the kernel
 9305           still does not support), as /dev/shm and /tmp are still
 9306           shared resources used by both the system and unprivileged
 9307           users.
 9309         * logind will now automatically turn off automatic suspending
 9310           on laptop lid close when more than one display is
 9311           connected. This was previously expected to be implemented
 9312           individually in desktop environments (such as GNOME),
 9313           however has been added to logind now, in order to fix a
 9314           boot-time race where a desktop environment might not have
 9315           been started yet and thus not been able to take an inhibitor
 9316           lock at the time where logind already suspends the system
 9317           due to a closed lid.
 9319         * logind will now wait at least 30s after each system
 9320           suspend/resume cycle, and 3min after system boot before
 9321           suspending the system due to a closed laptop lid. This
 9322           should give USB docking stations and similar enough time to
 9323           be probed and configured after system resume and boot in
 9324           order to then act as suspend blocker.
 9326         * systemd-run gained a new --property= setting which allows
 9327           initialization of resource control properties (and others)
 9328           for the created scope or service unit. Example: "systemd-run
 9329           --property=BlockIOWeight=10 updatedb" may be used to run
 9330           updatedb at a low block IO scheduling weight.
 9332         * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
 9333           now also work in --scope mode.
 9335         * When systemd is compiled with kdbus support, basic support
 9336           for enforced policies is now in place. (Note that enabling
 9337           kdbus still voids your warranty and no API compatibility
 9338           promises are made.)
 9340         Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
 9341         K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
 9342         Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
 9343         Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
 9344         Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
 9345         Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
 9346         Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
 9347         Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
 9348         Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
 9349         Zbigniew Jędrzejewski-Szmek
 9351         — Berlin, 2014-03-12
 9353 CHANGES WITH 210:
 9355         * systemd will now relabel /dev after loading the SMACK policy
 9356           according to SMACK rules.
 9358         * A new unit file option AppArmorProfile= has been added to
 9359           set the AppArmor profile for the processes of a unit.
 9361         * A new condition check ConditionArchitecture= has been added
 9362           to conditionalize units based on the system architecture, as
 9363           reported by uname()'s "machine" field.
 9365         * systemd-networkd now supports matching on the system
 9366           virtualization, architecture, kernel command line, hostname
 9367           and machine ID.
 9369         * logind is now a lot more aggressive when suspending the
 9370           machine due to a closed laptop lid. Instead of acting only
 9371           on the lid close action, it will continuously watch the lid
 9372           status and act on it. This is useful for laptops where the
 9373           power button is on the outside of the chassis so that it can
 9374           be reached without opening the lid (such as the Lenovo
 9375           Yoga). On those machines, logind will now immediately
 9376           re-suspend the machine if the power button has been
 9377           accidentally pressed while the laptop was suspended and in a
 9378           backpack or similar.
 9380         * logind will now watch SW_DOCK switches and inhibit reaction
 9381           to the lid switch if it is pressed. This means that logind
 9382           will not suspend the machine anymore if the lid is closed
 9383           and the system is docked, if the laptop supports SW_DOCK
 9384           notifications via the input layer. Note that ACPI docking
 9385           stations do not generate this currently. Also note that this
 9386           logic is usually not fully sufficient and Desktop
 9387           Environments should take a lid switch inhibitor lock when an
 9388           external display is connected, as systemd will not watch
 9389           this on its own.
 9391         * nspawn will now make use of the devices cgroup controller by
 9392           default, and only permit creation of and access to the usual
 9393           API device nodes like /dev/null or /dev/random, as well as
 9394           access to (but not creation of) the pty devices.
 9396         * We will now ship a default .network file for
 9397           systemd-networkd that automatically configures DHCP for
 9398           network interfaces created by nspawn's --network-veth or
 9399           --network-bridge= switches.
 9401         * systemd will now understand the usual M, K, G, T suffixes
 9402           according to SI conventions (i.e. to the base 1000) when
 9403           referring to throughput and hardware metrics. It will stay
 9404           with IEC conventions (i.e. to the base 1024) for software
 9405           metrics, according to what is customary according to
 9406           Wikipedia. We explicitly document which base applies for
 9407           each configuration option.
 9409         * The DeviceAllow= setting in unit files now supports a syntax to
 9410           allow-list an entire group of devices node majors at once, based on
 9411           the /proc/devices listing. For example, with the string "char-pts",
 9412           it is now possible to allow-list all current and future pseudo-TTYs
 9413           at once.
 9415         * sd-event learned a new "post" event source. Event sources of
 9416           this type are triggered by the dispatching of any event
 9417           source of a type that is not "post". This is useful for
 9418           implementing clean-up and check event sources that are
 9419           triggered by other work being done in the program.
 9421         * systemd-networkd is no longer statically enabled, but uses
 9422           the usual [Install] sections so that it can be
 9423           enabled/disabled using systemctl. It still is enabled by
 9424           default however.
 9426         * When creating a veth interface pair with systemd-nspawn, the
 9427           host side will now be prefixed with "vb-" if
 9428           --network-bridge= is used, and with "ve-" if --network-veth
 9429           is used. This way, it is easy to distinguish these cases on
 9430           the host, for example to apply different configuration to
 9431           them with systemd-networkd.
 9433         * The compatibility libraries for libsystemd-journal.so,
 9434           libsystem-id128.so, libsystemd-login.so and
 9435           libsystemd-daemon.so do not make use of IFUNC
 9436           anymore. Instead, we now build libsystemd.so multiple times
 9437           under these alternative names. This means that the footprint
 9438           is drastically increased, but given that these are
 9439           transitional compatibility libraries, this should not matter
 9440           much. This change has been made necessary to support the ARM
 9441           platform for these compatibility libraries, as the ARM
 9442           toolchain is not really at the same level as the toolchain
 9443           for other architectures like x86 and does not support
 9444           IFUNC. Please make sure to use --enable-compat-libs only
 9445           during a transitional period!
 9447         * The .include syntax has been deprecated and is not documented
 9448           anymore. Drop-in files in .d directories should be used instead.
 9450         Contributions from: Andreas Fuchs, Armin K., Colin Walters,
 9451         Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
 9452         Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
 9453         St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
 9454         Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
 9455         Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
 9456         Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
 9457         Zbigniew Jędrzejewski-Szmek
 9459         — Berlin, 2014-02-24
 9461 CHANGES WITH 209:
 9463         * A new component "systemd-networkd" has been added that can
 9464           be used to configure local network interfaces statically or
 9465           via DHCP. It is capable of bringing up bridges, VLANs, and
 9466           bonding. Currently, no hook-ups for interactive network
 9467           configuration are provided. Use this for your initrd,
 9468           container, embedded, or server setup if you need a simple,
 9469           yet powerful, network configuration solution. This
 9470           configuration subsystem is quite nifty, as it allows wildcard
 9471           hotplug matching in interfaces. For example, with a single
 9472           configuration snippet, you can configure that all Ethernet
 9473           interfaces showing up are automatically added to a bridge,
 9474           or similar. It supports link-sensing and more.
 9476         * A new tool "systemd-socket-proxyd" has been added which can
 9477           act as a bidirectional proxy for TCP sockets. This is
 9478           useful for adding socket activation support to services that
 9479           do not actually support socket activation, including virtual
 9480           machines and the like.
 9482         * Add a new tool to save/restore rfkill state on
 9483           shutdown/boot.
 9485         * Save/restore state of keyboard backlights in addition to
 9486           display backlights on shutdown/boot.
 9488         * udev learned a new SECLABEL{} construct to label device
 9489           nodes with a specific security label when they appear. For
 9490           now, only SECLABEL{selinux} is supported, but the syntax is
 9491           prepared for additional security frameworks.
 9493         * udev gained a new scheme to configure link-level attributes
 9494           from files in /etc/systemd/network/*.link. These files can
 9495           match against MAC address, device path, driver name and type,
 9496           and will apply attributes like the naming policy, link speed,
 9497           MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
 9498           address assignment policy (randomized, …).
 9500         * The configuration of network interface naming rules for
 9501           "permanent interface names" has changed: a new NamePolicy=
 9502           setting in the [Link] section of .link files determines the
 9503           priority of possible naming schemes (onboard, slot, MAC,
 9504           path). The default value of this setting is determined by
 9505           /usr/lib/net/links/99-default.link. Old
 9506           80-net-name-slot.rules udev configuration file has been
 9507           removed, so local configuration overriding this file should
 9508           be adapted to override 99-default.link instead.
 9510         * When the User= switch is used in a unit file, also
 9511           initialize $SHELL= based on the user database entry.
 9513         * systemd no longer depends on libdbus. All communication is
 9514           now done with sd-bus, systemd's low-level bus library
 9515           implementation.
 9517         * kdbus support has been added to PID 1 itself. When kdbus is
 9518           enabled, this causes PID 1 to set up the system bus and
 9519           enable support for a new ".busname" unit type that
 9520           encapsulates bus name activation on kdbus. It works a little
 9521           bit like ".socket" units, except for bus names. A new
 9522           generator has been added that converts classic dbus1 service
 9523           activation files automatically into native systemd .busname
 9524           and .service units.
 9526         * sd-bus: add a light-weight vtable implementation that allows
 9527           defining objects on the bus with a simple static const
 9528           vtable array of its methods, signals and properties.
 9530         * systemd will not generate or install static dbus
 9531           introspection data anymore to /usr/share/dbus-1/interfaces,
 9532           as the precise format of these files is unclear, and
 9533           nothing makes use of it.
 9535         * A proxy daemon is now provided to proxy clients connecting
 9536           via classic D-Bus AF_UNIX sockets to kdbus, to provide full
 9537           compatibility with classic D-Bus.
 9539         * A bus driver implementation has been added that supports the
 9540           classic D-Bus bus driver calls on kdbus, also for
 9541           compatibility purposes.
 9543         * A new API "sd-event.h" has been added that implements a
 9544           minimal event loop API built around epoll. It provides a
 9545           couple of features that direct epoll usage is lacking:
 9546           prioritization of events, scales to large numbers of timer
 9547           events, per-event timer slack (accuracy), system-wide
 9548           coalescing of timer events, exit handlers, watchdog
 9549           supervision support using systemd's sd_notify() API, child
 9550           process handling.
 9552         * A new API "sd-rntl.h" has been added that provides an API
 9553           around the route netlink interface of the kernel, similar in
 9554           style to "sd-bus.h".
 9556         * A new API "sd-dhcp-client.h" has been added that provides a
 9557           small DHCPv4 client-side implementation. This is used by
 9558           "systemd-networkd".
 9560         * There is a new kernel command line option
 9561           "systemd.restore_state=0|1". When set to "0", none of the
 9562           systemd tools will restore saved runtime state to hardware
 9563           devices. More specifically, the rfkill and backlight states
 9564           are not restored.
 9566         * The FsckPassNo= compatibility option in mount/service units
 9567           has been removed. The fstab generator will now add the
 9568           necessary dependencies automatically, and does not require
 9569           PID1's support for that anymore.
 9571         * journalctl gained a new switch, --list-boots, that lists
 9572           recent boots with their times and boot IDs.
 9574         * The various tools like systemctl, loginctl, timedatectl,
 9575           busctl, systemd-run, … have gained a new switch "-M" to
 9576           connect to a specific, local OS container (as direct
 9577           connection, without requiring SSH). This works on any
 9578           container that is registered with machined, such as those
 9579           created by libvirt-lxc or nspawn.
 9581         * systemd-run and systemd-analyze also gained support for "-H"
 9582           to connect to remote hosts via SSH. This is particularly
 9583           useful for systemd-run because it enables queuing of jobs
 9584           onto remote systems.
 9586         * machinectl gained a new command "login" to open a getty
 9587           login in any local container. This works with any container
 9588           that is registered with machined (such as those created by
 9589           libvirt-lxc or nspawn), and which runs systemd inside.
 9591         * machinectl gained a new "reboot" command that may be used to
 9592           trigger a reboot on a specific container that is registered
 9593           with machined. This works on any container that runs an init
 9594           system of some kind.
 9596         * systemctl gained a new "list-timers" command to print a nice
 9597           listing of installed timer units with the times they elapse
 9598           next.
 9600         * Alternative reboot() parameters may now be specified on the
 9601           "systemctl reboot" command line and are passed to the
 9602           reboot() system call.
 9604         * systemctl gained a new --job-mode= switch to configure the
 9605           mode to queue a job with. This is a more generic version of
 9606           --fail, --irreversible, and --ignore-dependencies, which are
 9607           still available but not advertised anymore.
 9609         * /etc/systemd/system.conf gained new settings to configure
 9610           various default timeouts of units, as well as the default
 9611           start limit interval and burst. These may still be overridden
 9612           within each Unit.
 9614         * PID1 will now export on the bus profile data of the security
 9615           policy upload process (such as the SELinux policy upload to
 9616           the kernel).
 9618         * journald: when forwarding logs to the console, include
 9619           timestamps (following the setting in
 9620           /sys/module/printk/parameters/time).
 9622         * OnCalendar= in timer units now understands the special
 9623           strings "yearly" and "annually". (Both are equivalent)
 9625         * The accuracy of timer units is now configurable with the new
 9626           AccuracySec= setting. It defaults to 1min.
 9628         * A new dependency type JoinsNamespaceOf= has been added that
 9629           allows running two services within the same /tmp and network
 9630           namespace, if PrivateNetwork= or PrivateTmp= are used.
 9632         * A new command "cat" has been added to systemctl. It outputs
 9633           the original unit file of a unit, and concatenates the
 9634           contents of additional "drop-in" unit file snippets, so that
 9635           the full configuration is shown.
 9637         * systemctl now supports globbing on the various "list-xyz"
 9638           commands, like "list-units" or "list-sockets", as well as on
 9639           those commands which take multiple unit names.
 9641         * journalctl's --unit= switch gained support for globbing.
 9643         * All systemd daemons now make use of the watchdog logic so
 9644           that systemd automatically notices when they hang.
 9646         * If the $container_ttys environment variable is set,
 9647           getty-generator will automatically spawn a getty for each
 9648           listed tty. This is useful for container managers to request
 9649           login gettys to be spawned on as many ttys as needed.
 9651         * %h, %s, %U specifier support is not available anymore when
 9652           used in unit files for PID 1. This is because NSS calls are
 9653           not safe from PID 1. They stay available for --user
 9654           instances of systemd, and as special case for the root user.
 9656         * loginctl gained a new "--no-legend" switch to turn off output
 9657           of the legend text.
 9659         * The "sd-login.h" API gained three new calls:
 9660           sd_session_is_remote(), sd_session_get_remote_user(),
 9661           sd_session_get_remote_host() to query information about
 9662           remote sessions.
 9664         * The udev hardware database now also carries vendor/product
 9665           information of SDIO devices.
 9667         * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
 9668           determine whether watchdog notifications are requested by
 9669           the system manager.
 9671         * Socket-activated per-connection services now include a
 9672           short description of the connection parameters in the
 9673           description.
 9675         * tmpfiles gained a new "--boot" option. When this is not used,
 9676           only lines where the command character is not suffixed with
 9677           "!" are executed. When this option is specified, those
 9678           options are executed too. This partitions tmpfiles
 9679           directives into those that can be safely executed at any
 9680           time, and those which should be run only at boot (for
 9681           example, a line that creates /run/nologin).
 9683         * A new API "sd-resolve.h" has been added which provides a simple
 9684           asynchronous wrapper around glibc NSS hostname resolution
 9685           calls, such as getaddrinfo(). In contrast to glibc's
 9686           getaddrinfo_a(), it does not use signals. In contrast to most
 9687           other asynchronous name resolution libraries, this one does
 9688           not reimplement DNS, but reuses NSS, so that alternate
 9689           hostname resolution systems continue to work, such as mDNS,
 9690           LDAP, etc. This API is based on libasyncns, but it has been
 9691           cleaned up for inclusion in systemd.
 9693         * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
 9694           "sd-daemon.h" are no longer found in individual libraries
 9695           libsystemd-journal.so, libsystemd-login.so,
 9696           libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
 9697           merged them into a single library, libsystemd.so, which
 9698           provides all symbols. The reason for this is cyclic
 9699           dependencies, as these libraries tend to use each other's
 9700           symbols. So far, we have managed to workaround that by linking
 9701           a copy of a good part of our code into each of these
 9702           libraries again and again, which, however, makes certain
 9703           things hard to do, like sharing static variables. Also, it
 9704           substantially increases footprint. With this change, there
 9705           is only one library for the basic APIs systemd
 9706           provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
 9707           "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
 9708           library as well, however are subject to the --enable-kdbus
 9709           switch (see below). Note that "sd-dhcp-client.h" is not part
 9710           of this library (this is because it only consumes, never
 9711           provides, services of/to other APIs). To make the transition
 9712           easy from the separate libraries to the unified one, we
 9713           provide the --enable-compat-libs compile-time switch which
 9714           will generate stub libraries that are compatible with the
 9715           old ones but redirect all calls to the new one.
 9717         * All of the kdbus logic and the new APIs "sd-bus.h",
 9718           "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
 9719           and "sd-utf8.h" are compile-time optional via the
 9720           "--enable-kdbus" switch, and they are not compiled in by
 9721           default. To make use of kdbus, you have to explicitly enable
 9722           the switch. Note however, that neither the kernel nor the
 9723           userspace API for all of this is considered stable yet. We
 9724           want to maintain the freedom to still change the APIs for
 9725           now. By specifying this build-time switch, you acknowledge
 9726           that you are aware of the instability of the current
 9727           APIs.
 9729         * Also, note that while kdbus is pretty much complete,
 9730           it lacks one thing: proper policy support. This means you
 9731           can build a fully working system with all features; however,
 9732           it will be highly insecure. Policy support will be added in
 9733           one of the next releases, at the same time that we will
 9734           declare the APIs stable.
 9736         * When the kernel command line argument "kdbus" is specified,
 9737           systemd will automatically load the kdbus.ko kernel module. At
 9738           this stage of development, it is only useful for testing kdbus
 9739           and should not be used in production. Note: if "--enable-kdbus"
 9740           is specified, and the kdbus.ko kernel module is available, and
 9741           "kdbus" is added to the kernel command line, the entire system
 9742           runs with kdbus instead of dbus-daemon, with the above mentioned
 9743           problem of missing the system policy enforcement. Also a future
 9744           version of kdbus.ko or a newer systemd will not be compatible with
 9745           each other, and will unlikely be able to boot the machine if only
 9746           one of them is updated.
 9748         * systemctl gained a new "import-environment" command which
 9749           uploads the caller's environment (or parts thereof) into the
 9750           service manager so that it is inherited by services started
 9751           by the manager. This is useful to upload variables like
 9752           $DISPLAY into the user service manager.
 9754         * A new PrivateDevices= switch has been added to service units
 9755           which allows running a service with a namespaced /dev
 9756           directory that does not contain any device nodes for
 9757           physical devices. More specifically, it only includes devices
 9758           such as /dev/null, /dev/urandom, and /dev/zero which are API
 9759           entry points.
 9761         * logind has been extended to support behaviour like VT
 9762           switching on seats that do not support a VT. This makes
 9763           multi-session available on seats that are not the first seat
 9764           (seat0), and on systems where kernel support for VTs has
 9765           been disabled at compile-time.
 9767         * If a process holds a delay lock for system sleep or shutdown
 9768           and fails to release it in time, we will now log its
 9769           identity. This makes it easier to identify processes that
 9770           cause slow suspends or power-offs.
 9772         * When parsing /etc/crypttab, support for a new key-slot=
 9773           option as supported by Debian is added. It allows indicating
 9774           which LUKS slot to use on disk, speeding up key loading.
 9776         * The sd_journal_sendv() API call has been checked and
 9777           officially declared to be async-signal-safe so that it may
 9778           be invoked from signal handlers for logging purposes.
 9780         * Boot-time status output is now enabled automatically after a
 9781           short timeout if boot does not progress, in order to give
 9782           the user an indication what she or he is waiting for.
 9784         * The boot-time output has been improved to show how much time
 9785           remains until jobs expire.
 9787         * The KillMode= switch in service units gained a new possible
 9788           value "mixed". If set, and the unit is shut down, then the
 9789           initial SIGTERM signal is sent only to the main daemon
 9790           process, while the following SIGKILL signal is sent to
 9791           all remaining processes of the service.
 9793         * When a scope unit is registered, a new property "Controller"
 9794           may be set. If set to a valid bus name, systemd will send a
 9795           RequestStop() signal to this name when it would like to shut
 9796           down the scope. This may be used to hook manager logic into
 9797           the shutdown logic of scope units. Also, scope units may now
 9798           be put in a special "abandoned" state, in which case the
 9799           manager process which created them takes no further
 9800           responsibilities for it.
 9802         * When reading unit files, systemd will now verify
 9803           the access mode of these files, and warn about certain
 9804           suspicious combinations. This has been added to make it
 9805           easier to track down packaging bugs where unit files are
 9806           marked executable or world-writable.
 9808         * systemd-nspawn gained a new "--setenv=" switch to set
 9809           container-wide environment variables. The similar option in
 9810           systemd-activate was renamed from "--environment=" to
 9811           "--setenv=" for consistency.
 9813         * systemd-nspawn has been updated to create a new kdbus domain
 9814           for each container that is invoked, thus allowing each
 9815           container to have its own set of system and user buses,
 9816           independent of the host.
 9818         * systemd-nspawn gained a new --drop-capability= switch to run
 9819           the container with less capabilities than the default. Both
 9820           --drop-capability= and --capability= now take the special
 9821           string "all" for dropping or keeping all capabilities.
 9823         * systemd-nspawn gained new switches for executing containers
 9824           with specific SELinux labels set.
 9826         * systemd-nspawn gained a new --quiet switch to not generate
 9827           any additional output but the container's own console
 9828           output.
 9830         * systemd-nspawn gained a new --share-system switch to run a
 9831           container without PID namespacing enabled.
 9833         * systemd-nspawn gained a new --register= switch to control
 9834           whether the container is registered with systemd-machined or
 9835           not. This is useful for containers that do not run full
 9836           OS images, but only specific apps.
 9838         * systemd-nspawn gained a new --keep-unit which may be used
 9839           when invoked as the only program from a service unit, and
 9840           results in registration of the unit service itself in
 9841           systemd-machined, instead of a newly opened scope unit.
 9843         * systemd-nspawn gained a new --network-interface= switch for
 9844           moving arbitrary interfaces to the container. The new
 9845           --network-veth switch creates a virtual Ethernet connection
 9846           between host and container. The new --network-bridge=
 9847           switch then allows assigning the host side of this virtual
 9848           Ethernet connection to a bridge device.
 9850         * systemd-nspawn gained a new --personality= switch for
 9851           setting the kernel personality for the container. This is
 9852           useful when running a 32-bit container on a 64-bit host. A
 9853           similar option Personality= is now also available for service
 9854           units to use.
 9856         * logind will now also track a "Desktop" identifier for each
 9857           session which encodes the desktop environment of it. This is
 9858           useful for desktop environments that want to identify
 9859           multiple running sessions of itself easily.
 9861         * A new SELinuxContext= setting for service units has been
 9862           added that allows setting a specific SELinux execution
 9863           context for a service.
 9865         * Most systemd client tools will now honour $SYSTEMD_LESS for
 9866           settings of the "less" pager. By default, these tools will
 9867           override $LESS to allow certain operations to work, such as
 9868           jump-to-the-end. With $SYSTEMD_LESS, it is possible to
 9869           influence this logic.
 9871         * systemd's "seccomp" hook-up has been changed to make use of
 9872           the libseccomp library instead of using its own
 9873           implementation. This has benefits for portability among
 9874           other things.
 9876         * For usage together with SystemCallFilter=, a new
 9877           SystemCallErrorNumber= setting has been introduced that
 9878           allows configuration of a system error number to be returned
 9879           on filtered system calls, instead of immediately killing the
 9880           process. Also, SystemCallArchitectures= has been added to
 9881           limit access to system calls of a particular architecture
 9882           (in order to turn off support for unused secondary
 9883           architectures). There is also a global
 9884           SystemCallArchitectures= setting in system.conf now to turn
 9885           off support for non-native system calls system-wide.
 9887         * systemd requires a kernel with a working name_to_handle_at(),
 9888           please see the kernel config requirements in the README file.
 9890         Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
 9891         Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
 9892         Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
 9893         Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
 9894         Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
 9895         David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
 9896         Elia Pinto, Florian Weimer, George McCollister, Goffredo
 9897         Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
 9898         Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
 9899         Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
 9900         Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
 9901         Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
 9902         Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
 9903         Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
 9904         Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
 9905         Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
 9906         Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
 9907         Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
 9908         Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
 9909         Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
 9910         Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
 9911         Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
 9912         Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
 9913         Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
 9915         — Berlin, 2014-02-20
 9917 CHANGES WITH 208:
 9919         * logind has gained support for facilitating privileged input
 9920           and drm device access for unprivileged clients. This work is
 9921           useful to allow Wayland display servers (and similar
 9922           programs, such as kmscon) to run under the user's ID and
 9923           access input and drm devices which are normally
 9924           protected. When this is used (and the kernel is new enough)
 9925           logind will "mute" IO on the file descriptors passed to
 9926           Wayland as long as it is in the background and "unmute" it
 9927           if it returns into the foreground. This allows secure
 9928           session switching without allowing background sessions to
 9929           eavesdrop on input and display data. This also introduces
 9930           session switching support if VT support is turned off in the
 9931           kernel, and on seats that are not seat0.
 9933         * A new kernel command line option luks.options= is understood
 9934           now which allows specifying LUKS options for usage for LUKS
 9935           encrypted partitions specified with luks.uuid=.
 9937         * tmpfiles.d(5) snippets may now use specifier expansion in
 9938           path names. More specifically %m, %b, %H, %v, are now
 9939           replaced by the local machine id, boot id, hostname, and
 9940           kernel version number.
 9942         * A new tmpfiles.d(5) command "m" has been introduced which
 9943           may be used to change the owner/group/access mode of a file
 9944           or directory if it exists, but do nothing if it does not.
 9946         * This release removes high-level support for the
 9947           MemorySoftLimit= cgroup setting. The underlying kernel
 9948           cgroup attribute memory.soft_limit= is currently badly
 9949           designed and likely to be removed from the kernel API in its
 9950           current form, hence we should not expose it for now.
 9952         * The memory.use_hierarchy cgroup attribute is now enabled for
 9953           all cgroups systemd creates in the memory cgroup
 9954           hierarchy. This option is likely to be come the built-in
 9955           default in the kernel anyway, and the non-hierarchical mode
 9956           never made much sense in the intrinsically hierarchical
 9957           cgroup system.
 9959         * A new field _SYSTEMD_SLICE= is logged along with all journal
 9960           messages containing the slice a message was generated
 9961           from. This is useful to allow easy per-customer filtering of
 9962           logs among other things.
 9964         * systemd-journald will no longer adjust the group of journal
 9965           files it creates to the "systemd-journal" group. Instead we
 9966           rely on the journal directory to be owned by the
 9967           "systemd-journal" group, and its setgid bit set, so that the
 9968           kernel file system layer will automatically enforce that
 9969           journal files inherit this group assignment. The reason for
 9970           this change is that we cannot allow NSS look-ups from
 9971           journald which would be necessary to resolve
 9972           "systemd-journal" to a numeric GID, because this might
 9973           create deadlocks if NSS involves synchronous queries to
 9974           other daemons (such as nscd, or sssd) which in turn are
 9975           logging clients of journald and might block on it, which
 9976           would then dead lock. A tmpfiles.d(5) snippet included in
 9977           systemd will make sure the setgid bit and group are
 9978           properly set on the journal directory if it exists on every
 9979           boot. However, we recommend adjusting it manually after
 9980           upgrades too (or from RPM scriptlets), so that the change is
 9981           not delayed until next reboot.
 9983         * Backlight and random seed files in /var/lib/ have moved into
 9984           the /var/lib/systemd/ directory, in order to centralize all
 9985           systemd generated files in one directory.
 9987         * Boot time performance measurements (as displayed by
 9988           "systemd-analyze" for example) will now read ACPI 5.0 FPDT
 9989           performance information if that's available to determine how
 9990           much time BIOS and boot loader initialization required. With
 9991           a sufficiently new BIOS you hence no longer need to boot
 9992           with Gummiboot to get access to such information.
 9994         Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
 9995         Cristian Rodríguez, Dave Reisner, David Herrmann, David
 9996         Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
 9997         feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
 9998         Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
 9999         Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
10000         Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
10002         — Berlin, 2013-10-02
10004 CHANGES WITH 207:
10006         * The Restart= option for services now understands a new
10007           on-watchdog setting, which will restart the service
10008           automatically if the service stops sending out watchdog keep
10009           alive messages (as configured with WatchdogSec=).
10011         * The getty generator (which is responsible for bringing up a
10012           getty on configured serial consoles) will no longer only
10013           start a getty on the primary kernel console but on all
10014           others, too. This makes the order in which console= is
10015           specified on the kernel command line less important.
10017         * libsystemd-logind gained a new sd_session_get_vt() call to
10018           retrieve the VT number of a session.
10020         * If the option "tries=0" is set for an entry of /etc/crypttab
10021           its passphrase is queried indefinitely instead of any
10022           maximum number of tries.
10024         * If a service with a configure PID file terminates its PID
10025           file will now be removed automatically if it still exists
10026           afterwards. This should put an end to stale PID files.
10028         * systemd-run will now also take relative binary path names
10029           for execution and no longer insists on absolute paths.
10031         * InaccessibleDirectories= and ReadOnlyDirectories= now take
10032           paths that are optionally prefixed with "-" to indicate that
10033           it should not be considered a failure if they do not exist.
10035         * journalctl -o (and similar commands) now understands a new
10036           output mode "short-precise", it is similar to "short" but
10037           shows timestamps with usec accuracy.
10039         * The option "discard" (as known from Debian) is now
10040           synonymous to "allow-discards" in /etc/crypttab. In fact,
10041           "discard" is preferred now (since it is easier to remember
10042           and type).
10044         * Some licensing clean-ups were made, so that more code is now
10045           LGPL-2.1 licensed than before.
10047         * A minimal tool to save/restore the display backlight
10048           brightness across reboots has been added. It will store the
10049           backlight setting as late as possible at shutdown, and
10050           restore it as early as possible during reboot.
10052         * A logic to automatically discover and enable home and swap
10053           partitions on GPT disks has been added. With this in place
10054           /etc/fstab becomes optional for many setups as systemd can
10055           discover certain partitions located on the root disk
10056           automatically. Home partitions are recognized under their
10057           GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
10058           partitions are recognized under their GPT type ID
10059           0657fd6da4ab43c484e50933c84b4f4f.
10061         * systemd will no longer pass any environment from the kernel
10062           or initrd to system services. If you want to set an
10063           environment for all services, do so via the kernel command
10064           line systemd.setenv= assignment.
10066         * The systemd-sysctl tool no longer natively reads the file
10067           /etc/sysctl.conf. If desired, the file should be symlinked
10068           from /etc/sysctl.d/99-sysctl.conf. Apart from providing
10069           legacy support by a symlink rather than built-in code, it
10070           also makes the otherwise hidden order of application of the
10071           different files visible. (Note that this partly reverts to a
10072           pre-198 application order of sysctl knobs!)
10074         * The "systemctl set-log-level" and "systemctl dump" commands
10075           have been moved to systemd-analyze.
10077         * systemd-run learned the new --remain-after-exit switch,
10078           which causes the scope unit not to be cleaned up
10079           automatically after the process terminated.
10081         * tmpfiles learned a new --exclude-prefix= switch to exclude
10082           certain paths from operation.
10084         * journald will now automatically flush all messages to disk
10085           as soon as a message at the log level CRIT, ALERT or EMERG
10086           is received.
10088         Contributions from: Andrew Cook, Brandon Philips, Christian
10089         Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
10090         Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
10091         McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
10092         Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
10093         Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
10094         Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
10095         Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
10096         Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
10097         Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
10098         Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
10099         Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
10100         William Giokas, Zbigniew Jędrzejewski-Szmek
10102         — Berlin, 2013-09-13
10104 CHANGES WITH 206:
10106         * The documentation has been updated to cover the various new
10107           concepts introduced with 205.
10109         * Unit files now understand the new %v specifier which
10110           resolves to the kernel version string as returned by "uname
10111           -r".
10113         * systemctl now supports filtering the unit list output by
10114           load state, active state and sub state, using the new
10115           --state= parameter.
10117         * "systemctl status" will now show the results of the
10118           condition checks (like ConditionPathExists= and similar) of
10119           the last start attempts of the unit. They are also logged to
10120           the journal.
10122         * "journalctl -b" may now be used to look for boot output of a
10123           specific boot. Try "journalctl -b -1" for the previous boot,
10124           but the syntax is substantially more powerful.
10126         * "journalctl --show-cursor" has been added which prints the
10127           cursor string the last shown log line. This may then be used
10128           with the new "journalctl --after-cursor=" switch to continue
10129           browsing logs from that point on.
10131         * "journalctl --force" may now be used to force regeneration
10132           of an FSS key.
10134         * Creation of "dead" device nodes has been moved from udev
10135           into kmod and tmpfiles. Previously, udev would read the kmod
10136           databases to pre-generate dead device nodes based on meta
10137           information contained in kernel modules, so that these would
10138           be auto-loaded on access rather then at boot. As this
10139           does not really have much to do with the exposing actual
10140           kernel devices to userspace this has always been slightly
10141           alien in the udev codebase. Following the new scheme kmod
10142           will now generate a runtime snippet for tmpfiles from the
10143           module meta information and it now is tmpfiles' job to the
10144           create the nodes. This also allows overriding access and
10145           other parameters for the nodes using the usual tmpfiles
10146           facilities. As side effect this allows us to remove the
10147           CAP_SYS_MKNOD capability bit from udevd entirely.
10149         * logind's device ACLs may now be applied to these "dead"
10150           devices nodes too, thus finally allowing managed access to
10151           devices such as /dev/snd/sequencer without loading the
10152           backing module right-away.
10154         * A new RPM macro has been added that may be used to apply
10155           tmpfiles configuration during package installation.
10157         * systemd-detect-virt and ConditionVirtualization= now can
10158           detect User-Mode-Linux machines (UML).
10160         * journald will now implicitly log the effective capabilities
10161           set of processes in the message metadata.
10163         * systemd-cryptsetup has gained support for TrueCrypt volumes.
10165         * The initrd interface has been simplified (more specifically,
10166           support for passing performance data via environment
10167           variables and fsck results via files in /run has been
10168           removed). These features were non-essential, and are