]> git.ipfire.org Git - thirdparty/systemd.git/blame - NEWS
Merge pull request #11351 from yuwata/ethtool-port-table
[thirdparty/systemd.git] / NEWS
CommitLineData
d657c51f 1systemd System and Service Manager
220a21d3 2
32673162 3CHANGES WITH 240:
fcb97512 4
e68a35a7
ZJS
5 * NoNewPrivileges=yes has been set for all long-running services
6 implemented by systemd. Previously, this was problematic due to
7 SELinux (as this would also prohibit the transition from PID1's label
8 to the service's label). This restriction has since been lifted, but
9 an SELinux policy update is required.
10 (See e.g. https://github.com/fedora-selinux/selinux-policy/pull/234.)
11
aa2437e2
YW
12 * DynamicUser=yes is dropped from systemd-networkd.service,
13 systemd-resolved.service and systemd-timesyncd.service, which was
14 enabled in v239 for systemd-networkd.service and systemd-resolved.service,
15 and since v236 for systemd-timesyncd.service. The users and groups
16 systemd-network, systemd-resolve and systemd-timesync are created
17 by systemd-sysusers again. Distributors or system administrators
18 may need to create these users and groups if they not exist (or need
19 to re-enable DynamicUser= for those units) while upgrading systemd.
20
b1a082cd
ZJS
21 * When unit files are loaded from disk, previously systemd would
22 sometimes (depending on the unit loading order) load units from the
23 target path of symlinks in .wants/ or .requires/ directories of other
24 units. This meant that unit could be loaded from different paths
25 depending on whether the unit was requested explicitly or as a
26 dependency of another unit, not honouring the priority of directories
27 in search path. It also meant that it was possible to successfully
28 load and start units which are not found in the unit search path, as
29 long as they were requested as a dependency and linked to from
30 .wants/ or .requires/. The target paths of those symlinks are not
31 used for loading units anymore and the unit file must be found in
32 the search path.
33
fcb97512 34 * A new service type has been added: Type=exec. It's very similar to
421e3b45 35 Type=simple but ensures the service manager will wait for both fork()
fcb97512
LP
36 and execve() of the main service binary to complete before proceeding
37 with follow-up units. This is primarily useful so that the manager
38 propagates any errors in the preparation phase of service execution
39 back to the job that requested the unit to be started. For example,
40 consider a service that has ExecStart= set to a file system binary
421e3b45
ZJS
41 that doesn't exist. With Type=simple starting the unit would be
42 considered instantly successful, as only fork() has to complete
43 successfully and the manager does not wait for execve(), and hence
44 its failure is seen "too late". With the new Type=exec service type
45 starting the unit will fail, as the manager will wait for the
46 execve() and notice its failure, which is then propagated back to the
47 start job.
fcb97512
LP
48
49 NOTE: with the next release 241 of systemd we intend to change the
50 systemd-run tool to default to Type=exec for transient services
51 started by it. This should be mostly safe, but in specific corner
52 cases might result in problems, as the systemd-run tool will then
6b1ab752 53 block on NSS calls (such as user name look-ups due to User=) done
fcb97512
LP
54 between the fork() and execve(), which under specific circumstances
55 might cause problems. It is recommended to specify "-p Type=simple"
56 explicitly in the few cases where this applies. For regular,
57 non-transient services (i.e. those defined with unit files on disk)
58 we will continue to default to Type=simple.
59
0972c1ae
LP
60 * The Linux kernel's current default RLIMIT_NOFILE resource limit for
61 userspace processes is set to 1024 (soft) and 4096
62 (hard). Previously, systemd passed this on unmodified to all
63 processes it forked off. With this systemd release the hard limit
0abf9492 64 systemd passes on is increased to 512K, overriding the kernel's
0972c1ae
LP
65 defaults and substantially increasing the number of simultaneous file
66 descriptors unprivileged userspace processes can allocate. Note that
67 the soft limit remains at 1024 for compatibility reasons: the
68 traditional UNIX select() call cannot deal with file descriptors >=
69 1024 and increasing the soft limit globally might thus result in
70 programs unexpectedly allocating a high file descriptor and thus
71 failing abnormally when attempting to use it with select() (of
72 course, programs shouldn't use select() anymore, and prefer
73 poll()/epoll, but the call unfortunately remains undeservedly popular
74 at this time). This change reflects the fact that file descriptor
75 handling in the Linux kernel has been optimized in more recent
76 kernels and allocating large numbers of them should be much cheaper
77 both in memory and in performance than it used to be. Programs that
78 want to take benefit of the increased limit have to "opt-in" into
421e3b45
ZJS
79 high file descriptors explicitly by raising their soft limit. Of
80 course, when they do that they must acknowledge that they cannot use
81 select() anymore (and neither can any shared library they use — or
82 any shared library used by any shared library they use and so on).
83 Which default hard limit is most appropriate is of course hard to
84 decide. However, given reports that ~300K file descriptors are used
85 in real-life applications we believe 512K is sufficiently high as new
86 default for now. Note that there are also reports that using very
87 high hard limits (e.g. 1G) is problematic: some software allocates
88 large arrays with one element for each potential file descriptor
89 (Java, …) — a high hard limit thus triggers excessively large memory
90 allocations in these applications. Hopefully, the new default of 512K
91 is a good middle ground: higher than what real-life applications
92 currently need, and low enough for avoid triggering excessively large
93 allocations in problematic software. (And yes, somebody should fix
94 Java.)
0972c1ae 95
a8b627aa
LP
96 * The fs.nr_open and fs.file-max sysctls are now automatically bumped
97 to the highest possible values, as separate accounting of file
98 descriptors is no longer necessary, as memcg tracks them correctly as
99 part of the memory accounting anyway. Thus, from the four limits on
100 file descriptors currently enforced (fs.file-max, fs.nr_open,
101 RLIMIT_NOFILE hard, RLIMIT_NOFILE soft) we turn off the first two,
102 and keep only the latter two. A set of build-time options
103 (-Dbump-proc-sys-fs-file-max=no and -Dbump-proc-sys-fs-nr-open=no)
104 has been added to revert this change in behaviour, which might be
105 an option for systems that turn off memcg in the kernel.
106
4f7dc24f
LP
107 * When no /etc/locale.conf file exists (and hence no locale settings
108 are in place), systemd will now use the "C.UTF-8" locale by default,
109 and set LANG= to it. This locale is supported by various
110 distributions including Fedora, with clear indications that upstream
111 glibc is going to make it available too. This locale enables UTF-8
112 mode by default, which appears appropriate for 2018.
113
230450d4
LR
114 * The "net.ipv4.conf.all.rp_filter" sysctl will now be set to 2 by
115 default. This effectively switches the RFC3704 Reverse Path filtering
116 from Strict mode to Loose mode. This is more appropriate for hosts
117 that have multiple links with routes to the same networks (e.g.
118 a client with a Wi-Fi and Ethernet both connected to the internet).
119
6b1ab752 120 Consult the kernel documentation for details on this sysctl:
230450d4
LR
121 https://www.kernel.org/doc/Documentation/networking/ip-sysctl.txt
122
23305a29
CD
123 * CPUAccounting=yes no longer enables the CPU controller when using
124 kernel 4.15+ and the unified cgroup hierarchy, as required accounting
125 statistics are now provided independently from the CPU controller.
126
6b1ab752 127 * Support for disabling a particular cgroup controller within a sub-tree
a8467688
CD
128 has been added through the DisableControllers= directive.
129
8f044cf9
CD
130 * cgroup_no_v1=all on the kernel command line now also implies
131 using the unified cgroup hierarchy, unless one explicitly passes
132 systemd.unified_cgroup_hierarchy=0 on the kernel command line.
133
6b1ab752
LP
134 * The new "MemoryMin=" unit file property may now be used to set the
135 memory usage protection limit of processes invoked by the unit. This
4e1dfa45 136 controls the cgroup v2 memory.min attribute. Similarly, the new
6b1ab752 137 "IODeviceLatencyTargetSec=" property has been added, wrapping the new
4e1dfa45 138 cgroup v2 io.latency cgroup property for configuring per-service I/O
6b1ab752
LP
139 latency.
140
4e1dfa45
CD
141 * systemd now supports the cgroup v2 devices BPF logic, as counterpart
142 to the cgroup v1 "devices" cgroup controller.
6b1ab752
LP
143
144 * systemd-escape now is able to combine --unescape with --template. It
145 also learnt a new option --instance for extracting and unescaping the
146 instance part of a unit name.
147
148 * sd-bus now provides the sd_bus_message_readv() which is similar to
149 sd_bus_message_read() but takes a va_list object. The pair
150 sd_bus_set_method_call_timeout() and sd_bus_get_method_call_timeout()
421e3b45 151 has been added for configuring the default method call timeout to
6b1ab752
LP
152 use. sd_bus_error_move() may be used to efficiently move the contents
153 from one sd_bus_error structure to another, invalidating the
154 source. sd_bus_set_close_on_exit() and sd_bus_get_close_on_exit() may
155 be used to control whether a bus connection object is automatically
156 flushed when an sd-event loop is exited.
157
158 * When processing classic BSD syslog log messages, journald will now
159 save the original time-stamp string supplied in the new
160 SYSLOG_TIMESTAMP= journal field. This permits consumers to
161 reconstruct the original BSD syslog message more correctly.
162
163 * StandardOutput=/StandardError= in service files gained support for
164 new "append:…" parameters, for connecting STDOUT/STDERR of a service
165 to a file, and appending to it.
166
167 * The signal to use as last step of killing of unit processes is now
168 configurable. Previously it was hard-coded to SIGKILL, which may now
169 be overridden with the new KillSignal= setting. Note that this is the
46b028f2 170 signal used when regular termination (i.e. SIGTERM) does not suffice.
421e3b45
ZJS
171 Similarly, the signal used when aborting a program in case of a
172 watchdog timeout may now be configured too (WatchdogSignal=).
6b1ab752
LP
173
174 * The XDG_SESSION_DESKTOP environment variable may now be configured in
175 the pam_systemd argument line, using the new desktop= switch. This is
176 useful to initialize it properly from a display manager without
177 having to touch C code.
178
421e3b45
ZJS
179 * Most configuration options that previously accepted percentage values
180 now also accept permille values with the '‰' suffix (instead of '%').
6b1ab752 181
6b1ab752
LP
182 * systemd-resolved may now optionally use OpenSSL instead of GnuTLS for
183 DNS-over-TLS.
184
185 * systemd-resolved's configuration file resolved.conf gained a new
186 option ReadEtcHosts= which may be used to turn off processing and
187 honoring /etc/hosts entries.
188
189 * The "--wait" switch may now be passed to "systemctl
190 is-system-running", in which case the tool will synchronously wait
191 until the system finished start-up.
192
193 * hostnamed gained a new bus call to determine the DMI product UUID.
194
195 * On x86-64 systemd will now prefer using the RDRAND processor
196 instruction over /dev/urandom whenever it requires randomness that
197 neither has to be crypto-grade nor should be reproducible. This
198 should substantially reduce the amount of entropy systemd requests
199 from the kernel during initialization on such systems, though not
200 reduce it to zero. (Why not zero? systemd still needs to allocate
201 UUIDs and such uniquely, which require high-quality randomness.)
202
203 * networkd gained support for Foo-Over-UDP, ERSPAN and ISATAP
204 tunnels. It also gained a new option ForceDHCPv6PDOtherInformation=
205 for forcing the "Other Information" bit in IPv6 RA messages. The
d6131be9 206 bonding logic gained four new options AdActorSystemPriority=,
6b1ab752 207 AdUserPortKey=, AdActorSystem= for configuring various 802.3ad
d6131be9
YW
208 aspects, and DynamicTransmitLoadBalancing= for enabling dynamic
209 shuffling of flows. The tunnel logic gained a new
210 IPv6RapidDeploymentPrefix= option for configuring IPv6 Rapid
211 Deployment. The policy rule logic gained four new options IPProtocol=,
212 SourcePort= and DestinationPort=, InvertRule=. The bridge logic gained
213 support for the MulticastToUnicast= option. networkd also gained
214 support for configuring static IPv4 ARP or IPv6 neighbor entries.
6b1ab752
LP
215
216 * .preset files (as read by 'systemctl preset') may now be used to
217 instantiate services.
218
219 * /etc/crypttab now understands the sector-size= option to configure
220 the sector size for an encrypted partition.
221
222 * Key material for encrypted disks may now be placed on a formatted
421e3b45
ZJS
223 medium, and referenced from /etc/crypttab by the UUID of the file
224 system, followed by "=" suffixed by the path to the key file.
6b1ab752
LP
225
226 * The "collect" udev component has been removed without replacement, as
421e3b45 227 it is neither used nor maintained.
6b1ab752
LP
228
229 * When the RuntimeDirectory=, StateDirectory=, CacheDirectory=,
230 LogsDirectory=, ConfigurationDirectory= settings are used in a
231 service the executed processes will now receive a set of environment
421e3b45
ZJS
232 variables containing the full paths of these directories.
233 Specifically, RUNTIME_DIRECTORY=, STATE_DIRECTORY, CACHE_DIRECTORY,
234 LOGS_DIRECTORY, CONFIGURATION_DIRECTORY are now set if these options
235 are used. Note that these options may be used multiple times per
236 service in which case the resulting paths will be concatenated and
237 separated by colons.
6b1ab752
LP
238
239 * Predictable interface naming has been extended to cover InfiniBand
240 NICs. They will be exposed with an "ib" prefix.
241
242 * tmpfiles.d/ line types may now be suffixed with a '-' character, in
243 which case the respective line failing is ignored.
244
245 * .link files may now be used to configure the equivalent to the
246 "ethtool advertise" commands.
247
248 * The sd-device.h and sd-hwdb.h APIs are now exported, as an
249 alternative to libudev.h. Previously, the latter was just an internal
250 wrapper around the former, but now these two APIs are exposed
251 directly.
252
253 * sd-id128.h gained a new function sd_id128_get_boot_app_specific()
254 which calculates an app-specific boot ID similar to how
255 sd_id128_get_machine_app_specific() generates an app-specific machine
256 ID.
257
258 * A new tool systemd-id128 has been added that can be used to determine
259 and generate various 128bit IDs.
260
261 * /etc/os-release gained two new standardized fields DOCUMENTATION_URL=
262 and LOGO=.
263
264 * systemd-hibernate-resume-generator will now honor the "noresume"
265 kernel command line option, in which case it will bypass resuming
266 from any hibernated image.
267
268 * The systemd-sleep.conf configuration file gained new options
269 AllowSuspend=, AllowHibernation=, AllowSuspendThenHibernate=,
270 AllowHybridSleep= for prohibiting specific sleep modes even if the
421e3b45 271 kernel exports them.
6b1ab752
LP
272
273 * portablectl is now officially supported and has thus moved to
274 /usr/bin/.
275
276 * bootctl learnt the two new commands "set-default" and "set-oneshot"
277 for setting the default boot loader item to boot to (either
278 persistently or only for the next boot). This is currently only
279 compatible with sd-boot, but may be implemented on other boot loaders
280 too, that follow the boot loader interface. The updated interface is
281 now documented here:
282
283 https://systemd.io/BOOT_LOADER_INTERFACE
284
285 * A new kernel command line option systemd.early_core_pattern= is now
286 understood which may be used to influence the core_pattern PID 1
287 installs during early boot.
288
289 * busctl learnt two new options -j and --json= for outputting method
290 call replies, properties and monitoring output in JSON.
291
292 * journalctl's JSON output now supports simple ANSI coloring as well as
293 a new "json-seq" mode for generating RFC7464 output.
294
295 * Unit files now support the %g/%G specifiers that resolve to the UNIX
296 group/GID of the service manager runs as, similar to the existing
297 %u/%U specifiers that resolve to the UNIX user/UID.
298
299 * systemd-logind learnt a new global configuration option
300 UserStopDelaySec= that may be set in logind.conf. It specifies how
301 long the systemd --user instance shall remain started after a user
302 logs out. This is useful to speed up repetitive re-connections of the
303 same user, as it means the user's service manager doesn't have to be
304 stopped/restarted on each iteration, but can be reused between
305 subsequent options. This setting defaults to 10s. systemd-logind also
306 exports two new properties on its Manager D-Bus objects indicating
421e3b45
ZJS
307 whether the system's lid is currently closed, and whether the system
308 is on AC power.
6b1ab752
LP
309
310 * systemd gained support for a generic boot counting logic, which
311 generically permits automatic reverting to older boot loader entries
312 if newer updated ones don't work. The boot loader side is implemented
313 in sd-boot, but is kept open for other boot loaders too. For details
314 see:
315
316 https://systemd.io/AUTOMATIC_BOOT_ASSESSMENT
317
318 * The SuccessAction=/FailureAction= unit file settings now learnt two
319 new parameters: "exit" and "exit-force", which result in immediate
320 exiting of the service manager, and are only useful in systemd --user
321 and container environments.
322
323 * Unit files gained support for a pair of options
324 FailureActionExitStatus=/SuccessActionExitStatus= for configuring the
325 exit status to use as service manager exit status when
326 SuccessAction=/FailureAction= is set to exit or exit-force.
327
328 * A pair of LogRateLimitIntervalSec=/LogRateLimitBurst= per-service
329 options may now be used to configure the log rate limiting applied by
330 journald per-service.
331
332 * systemd-analyze gained a new verb "timespan" for parsing and
333 normalizing time span values (i.e. strings like "5min 7s 8us").
334
335 * systemd-analyze also gained a new verb "security" for analyzing the
336 security and sand-boxing settings of services in order to determine an
337 "exposure level" for them, indicating whether a service would benefit
338 from more sand-boxing options turned on for them.
339
340 * "systemd-analyze syscall-filter" will now also show system calls
341 supported by the local kernel but not included in any of the defined
342 groups.
343
344 * .nspawn files now understand the Ephemeral= setting, matching the
345 --ephemeral command line switch.
346
347 * sd-event gained the new APIs sd_event_source_get_floating() and
348 sd_event_source_set_floating() for controlling whether a specific
349 event source is "floating", i.e. destroyed along with the even loop
350 object itself.
351
352 * Unit objects on D-Bus gained a new "Refs" property that lists all
421e3b45
ZJS
353 clients that currently have a reference on the unit (to ensure it is
354 not unloaded).
6b1ab752
LP
355
356 * The JoinControllers= option in system.conf is no longer supported, as
357 it didn't work correctly, is hard to support properly, is legacy (as
4e1dfa45 358 the concept only exists on cgroup v1) and apparently wasn't used.
6b1ab752
LP
359
360 * Journal messages that are generated whenever a unit enters the failed
421e3b45
ZJS
361 state are now tagged with a unique MESSAGE_ID. Similarly, messages
362 generated whenever a service process exits are now made recognizable,
363 too. A taged message is also emitted whenever a unit enters the
364 "dead" state on success.
6b1ab752
LP
365
366 * systemd-run gained a new switch --working-directory= for configuring
367 the working directory of the service to start. A shortcut -d is
368 equivalent, setting the working directory of the service to the
369 current working directory of the invoking program. The new --shell
370 (or just -S) option has been added for invoking the $SHELL of the
371 caller as a service, and implies --pty --same-dir --wait --collect
421e3b45 372 --service-type=exec. Or in other words, "systemd-run -S" is now the
6b1ab752
LP
373 quickest way to quickly get an interactive in a fully clean and
374 well-defined system service context.
375
376 * machinectl gained a new verb "import-fs" for importing an OS tree
377 from a directory. Moreover, when a directory or tarball is imported
378 and single top-level directory found with the OS itself below the OS
379 tree is automatically mangled and moved one level up.
380
421e3b45
ZJS
381 * systemd-importd will no longer set up an implicit btrfs loop-back
382 file system on /var/lib/machines. If one is already set up, it will
383 continue to be used.
6b1ab752
LP
384
385 * A new generator "systemd-run-generator" has been added. It will
386 synthesize a unit from one or more program command lines included in
387 the kernel command line. This is very useful in container managers
388 for example:
389
390 # systemd-nspawn -i someimage.raw -b systemd.run='"some command line"'
391
392 This will run "systemd-nspawn" on an image, invoke the specified
421e3b45
ZJS
393 command line and immediately shut down the container again, returning
394 the command line's exit code.
6b1ab752 395
421e3b45 396 * The block device locking logic is now documented:
6b1ab752
LP
397
398 https://systemd.io/BLOCK_DEVICE_LOCKING
399
400 * loginctl and machinectl now optionally output the various tables in
401 JSON using the --output= switch. It is our intention to add similar
402 support to systemctl and all other commands.
403
404 * udevadm's query and trigger verb now optionally take a .device unit
405 name as argument.
406
407 * systemd-udevd's network naming logic now understands a new
421e3b45 408 net.naming-scheme= kernel command line switch, which may be used to
6b1ab752
LP
409 pick a specific version of the naming scheme. This helps stabilizing
410 interface names even as systemd/udev are updated and the naming logic
411 is improved.
412
67081438
LP
413 * sd-id128.h learnt two new auxiliary helpers: sd_id128_is_allf() and
414 SD_ID128_ALLF to test if a 128bit ID is set to all 0xFF bytes, and to
415 initialize one to all 0xFF.
416
144d7f1d
LP
417 * After loading the SELinux policy systemd will now recursively relabel
418 all files and directories listed in
419 /run/systemd/relabel-extra.d/*.relabel (which should be simple
420 newline separated lists of paths) in addition to the ones it already
421 implicitly relabels in /run, /dev and /sys. After the relabelling is
422 completed the *.relabel files (and /run/systemd/relabel-extra.d/) are
423 removed. This is useful to permit initrds (i.e. code running before
424 the SELinux policy is in effect) to generate files in the host
425 filesystem safely and ensure that the correct label is applied during
426 the transition to the host OS.
427
98a7b55a
LP
428 * KERNEL API BREAKAGE: Linux kernel 4.18 changed behaviour regarding
429 mknod() handling in user namespaces. Previously mknod() would always
430 fail with EPERM in user namespaces. Since 4.18 mknod() will succeed
431 but device nodes generated that way cannot be opened, and attempts to
432 open them result in EPERM. This breaks the "graceful fallback" logic
433 in systemd's PrivateDevices= sand-boxing option. This option is
434 implemented defensively, so that when systemd detects it runs in a
435 restricted environment (such as a user namespace, or an environment
436 where mknod() is blocked through seccomp or absence of CAP_SYS_MKNOD)
437 where device nodes cannot be created the effect of PrivateDevices= is
438 bypassed (following the logic that 2nd-level sand-boxing is not
439 essential if the system systemd runs in is itself already sand-boxed
440 as a whole). This logic breaks with 4.18 in container managers where
441 user namespacing is used: suddenly PrivateDevices= succeeds setting
442 up a private /dev/ file system containing devices nodes — but when
443 these are opened they don't work.
444
445 At this point is is recommended that container managers utilizing
446 user namespaces that intend to run systemd in the payload explicitly
447 block mknod() with seccomp or similar, so that the graceful fallback
448 logic works again.
449
450 We are very sorry for the breakage and the requirement to change
451 container configurations for newer kernels. It's purely caused by an
452 incompatible kernel change. The relevant kernel developers have been
453 notified about this userspace breakage quickly, but they chose to
454 ignore it.
455
6b1ab752
LP
456 Contributions from: afg, Alan Jenkins, Aleksei Timofeyev, Alexander
457 Filippov, Alexander Kurtz, Alexey Bogdanenko, Andreas Henriksson,
458 Andrew Jorgensen, Anita Zhang, apnix-uk, Arkan49, Arseny Maslennikov,
459 asavah, Asbjørn Apeland, aszlig, Bastien Nocera, Ben Boeckel, Benedikt
c37e2358
LP
460 Morbach, Benjamin Berg, Bruce Zhang, Carlo Caione, Cedric Viou, Chen
461 Qi, Chris Chiu, Chris Down, Chris Morin, Christian Rebischke, Claudius
462 Ellsel, Colin Guthrie, dana, Daniel, Daniele Medri, Daniel Kahn
463 Gillmor, Daniel Rusek, Daniel van Vugt, Dariusz Gadomski, Dave Reisner,
464 David Anderson, Davide Cavalca, David Leeds, David Malcolm, David
1742aae2
ZJS
465 Strauss, David Tardon, Dimitri John Ledkov, Dmitry Torokhov, dj-kaktus,
466 Dongsu Park, Elias Probst, Emil Soleyman, Erik Kooistra, Ervin Peters,
467 Evgeni Golov, Evgeny Vereshchagin, Fabrice Fontaine, Faheel Ahmad,
468 Faizal Luthfi, Felix Yan, Filipe Brandenburger, Franck Bui, Frank
469 Schaefer, Frantisek Sumsal, Gautier Husson, Gianluca Boiano, Giuseppe
470 Scrivano, glitsj16, Hans de Goede, Harald Hoyer, Harry Mallon, Harshit
471 Jain, Helmut Grohne, Henry Tung, Hui Yiqun, imayoda, Insun Pyo, Iwan
472 Timmer, Jan Janssen, Jan Pokorný, Jan Synacek, Jason A. Donenfeld,
473 javitoom, Jérémy Nouhaud, Jeremy Su, Jiuyang Liu, João Paulo Rechi
474 Vita, Joe Hershberger, Joe Rayhawk, Joerg Behrmann, Joerg Steffens,
475 Jonas Dorel, Jon Ringle, Josh Soref, Julian Andres Klode, Jun Bo Bi,
476 Jürg Billeter, Keith Busch, Khem Raj, Kirill Marinushkin, Larry
477 Bernstone, Lennart Poettering, Lion Yang, Li Song, Lorenz
478 Hübschle-Schneider, Lubomir Rintel, Lucas Werkmeister, Ludwin Janvier,
479 Lukáš Nykrýn, Luke Shumaker, mal, Marc-Antoine Perennou, Marcin
480 Skarbek, Marco Trevisan (Treviño), Marian Cepok, Mario Hros, Marko
481 Myllynen, Markus Grimm, Martin Pitt, Martin Sobotka, Martin Wilck,
482 Mathieu Trudel-Lapierre, Matthew Leeds, Michael Biebl, Michael Olbrich,
483 Michael 'pbone' Pobega, Michael Scherer, Michal Koutný, Michal
484 Sekletar, Michal Soltys, Mike Gilbert, Mike Palmer, Muhammet Kara, Neal
485 Gompa, Neil Brown, Network Silence, Niklas Tibbling, Nikolas Nyby,
486 Nogisaka Sadata, Oliver Smith, Patrik Flykt, Pavel Hrdina, Paweł
487 Szewczyk, Peter Hutterer, Piotr Drąg, Ray Strode, Reinhold Mueller,
488 Renaud Métrich, Roman Gushchin, Ronny Chevalier, Rubén Suárez Alvarez,
489 Ruixin Bao, RussianNeuroMancer, Ryutaroh Matsumoto, Saleem Rashid, Sam
490 Morris, Samuel Morris, Sandy Carter, scootergrisen, Sébastien Bacher,
491 Sergey Ptashnick, Shawn Landden, Shengyao Xue, Shih-Yuan Lee
492 (FourDollars), Silvio Knizek, Sjoerd Simons, Stasiek Michalski, Stephen
493 Gallagher, Steven Allen, Steve Ramage, Susant Sahani, Sven Joachim,
494 Sylvain Plantefève, Tanu Kaskinen, Tejun Heo, Thiago Macieira, Thomas
495 Blume, Thomas Haller, Thomas H. P. Andersen, Tim Ruffing, TJ, Tobias
496 Jungel, Todd Walton, Tommi Rantala, Tomsod M, Tony Novak, Tore
497 Anderson, Trevonn, Victor Laskurain, Victor Tapia, Violet Halo, Vojtech
498 Trefny, welaq, William A. Kennington III, William Douglas, Wyatt Ward,
499 Xiang Fan, Xi Ruoyao, Xuanwo, Yann E. Morin, YmrDtnJu, Yu Watanabe,
500 Zbigniew Jędrzejewski-Szmek, Zhang Xianwei, Zsolt Dollenstein
501
502 — Warsaw, 2018-12-21
6b1ab752 503
e8498f82 504CHANGES WITH 239:
019cb3ab
SH
505
506 * NETWORK INTERFACE DEVICE NAMING CHANGES: systemd-udevd's "net_id"
d69f5282
ZJS
507 builtin will name network interfaces differently than in previous
508 versions for virtual network interfaces created with SR-IOV and NPAR
509 and for devices where the PCI network controller device does not have
510 a slot number associated.
511
512 SR-IOV virtual devices are now named based on the name of the parent
513 interface, with a suffix of "v<N>", where <N> is the virtual device
514 number. Previously those virtual devices were named as if completely
515 independent.
516
517 The ninth and later NPAR virtual devices will be named following the
518 scheme used for the first eight NPAR partitions. Previously those
519 devices were not renamed and the kernel default (eth<n>) was used.
520
521 "net_id" will also generate names for PCI devices where the PCI
522 network controller device does not have an associated slot number
523 itself, but one of its parents does. Previously those devices were
524 not renamed and the kernel default (eth<n>) was used.
019cb3ab 525
6e2d744b
YW
526 * AF_INET and AF_INET6 are dropped from RestrictAddressFamilies= in
527 systemd-logind.service. Since v235, IPAddressDeny=any has been set to
fe903cf4
LP
528 the unit. So, it is expected that the default behavior of
529 systemd-logind is not changed. However, if distribution packagers or
530 administrators disabled or modified IPAddressDeny= setting by a
531 drop-in config file, then it may be necessary to update the file to
532 re-enable AF_INET and AF_INET6 to support network user name services,
533 e.g. NIS.
534
535 * When the RestrictNamespaces= unit property is specified multiple
536 times, then the specified types are merged now. Previously, only the
537 last assignment was used. So, if distribution packagers or
538 administrators modified the setting by a drop-in config file, then it
539 may be necessary to update the file.
e0eee477 540
1fc83d09
LP
541 * When OnFailure= is used in combination with Restart= on a service
542 unit, then the specified units will no longer be triggered on
543 failures that result in restarting. Previously, the specified units
544 would be activated each time the unit failed, even when the unit was
545 going to be restarted automatically. This behaviour contradicted the
546 documentation. With this release the code is adjusted to match the
547 documentation.
548
41a4c3ec
LP
549 * systemd-tmpfiles will now print a notice whenever it encounters
550 tmpfiles.d/ lines referencing the /var/run/ directory. It will
551 recommend reworking them to use the /run/ directory instead (for
5cadf58e
ZJS
552 which /var/run/ is simply a symlinked compatibility alias). This way
553 systemd-tmpfiles can properly detect line conflicts and merge lines
554 referencing the same file by two paths, without having to access
555 them.
41a4c3ec 556
ce55bd5e
ZJS
557 * systemctl disable/unmask/preset/preset-all cannot be used with
558 --runtime. Previously this was allowed, but resulted in unintuitive
fe903cf4
LP
559 behaviour that wasn't useful. systemctl disable/unmask will now undo
560 both runtime and persistent enablement/masking, i.e. it will remove
561 any relevant symlinks both in /run and /etc.
ce55bd5e 562
e01d9e21
LP
563 * Note that all long-running system services shipped with systemd will
564 now default to a system call whitelist (rather than a blacklist, as
565 before). In particular, systemd-udevd will now enforce one too. For
566 most cases this should be safe, however downstream distributions
567 which disabled sandboxing of systemd-udevd (specifically the
568 MountFlags= setting), might want to disable this security feature
569 too, as the default whitelisting will prohibit all mount, swap,
570 reboot and clock changing operations from udev rules.
571
5cadf58e
ZJS
572 * sd-boot acquired new loader configuration settings to optionally turn
573 off Windows and MacOS boot partition discovery as well as
574 reboot-into-firmware menu items. It is also able to pick a better
575 screen resolution for HiDPI systems, and now provides loader
41a4c3ec
LP
576 configuration settings to change the resolution explicitly.
577
c9299be2
IT
578 * systemd-resolved now supports DNS-over-TLS. It's still
579 turned off by default, use DNSOverTLS=opportunistic to turn it on in
c086ce8c
LP
580 resolved.conf. We intend to make this the default as soon as couple
581 of additional techniques for optimizing the initial latency caused by
582 establishing a TLS/TCP connection are implemented.
583
73c718a9
YW
584 * systemd-resolved.service and systemd-networkd.service now set
585 DynamicUser=yes. The users systemd-resolve and systemd-network are
abc291aa
LP
586 not created by systemd-sysusers anymore.
587
588 NOTE: This has a chance of breaking nss-ldap and similar NSS modules
589 that embedd a network facing module into any process using getpwuid()
590 or related call: the dynamic allocation of the user ID for
591 systemd-resolved.service means the service manager has to check NSS
592 if the user name is already taken when forking off the service. Since
593 the user in the common case won't be defined in /etc/passwd the
594 lookup is likely to trigger nss-ldap which in turn might use NSS to
595 ask systemd-resolved for hostname lookups. This will hence result in
596 a deadlock: a user name lookup in order to start
597 systemd-resolved.service will result in a host name lookup for which
598 systemd-resolved.service needs to be started already. There are
599 multiple ways to work around this problem: pre-allocate the
600 "systemd-resolve" user on such systems, so that nss-ldap won't be
601 triggered; or use a different NSS package that doesn't do networking
602 in-process but provides a local asynchronous name cache; or configure
603 the NSS package to avoid lookups for UIDs in the range `pkg-config
604 systemd --variable=dynamicuidmin` … `pkg-config systemd
605 --variable=dynamicuidmax`, so that it does not consider itself
606 authoritative for the same UID range systemd allocates dynamic users
607 from.
73c718a9 608
41a4c3ec
LP
609 * The systemd-resolve tool has been renamed to resolvectl (it also
610 remains available under the old name, for compatibility), and its
611 interface is now verb-based, similar in style to the other <xyz>ctl
5cadf58e
ZJS
612 tools, such as systemctl or loginctl.
613
75da262a
LP
614 * The resolvectl/systemd-resolve tool also provides 'resolvconf'
615 compatibility. It may be symlinked under the 'resolvconf' name, in
5cadf58e
ZJS
616 which case it will take arguments and input compatible with the
617 Debian and FreeBSD resolvconf tool.
41a4c3ec
LP
618
619 * Support for suspend-then-hibernate has been added, i.e. a sleep mode
3f9a0a52 620 where the system initially suspends, and after a timeout resumes and
41a4c3ec
LP
621 hibernates again.
622
623 * networkd's ClientIdentifier= now accepts a new option "duid-only". If
624 set the client will only send a DUID as client identifier.
625
626 * The nss-systemd glibc NSS module will now enumerate dynamic users and
627 groups in effect. Previously, it could resolve UIDs/GIDs to user
628 names/groups and vice versa, but did not support enumeration.
629
630 * journald's Compress= configuration setting now optionally accepts a
631 byte threshold value. All journal objects larger than this threshold
632 will be compressed, smaller ones will not. Previously this threshold
633 was not configurable and set to 512.
634
5cadf58e
ZJS
635 * A new system.conf setting NoNewPrivileges= is now available which may
636 be used to turn off acquisition of new privileges system-wide
637 (i.e. set Linux' PR_SET_NO_NEW_PRIVS for PID 1 itself, and thus also
638 for all its children). Note that turning this option on means setuid
639 binaries and file system capabilities lose their special powers.
640 While turning on this option is a big step towards a more secure
641 system, doing so is likely to break numerous pre-existing UNIX tools,
642 in particular su and sudo.
41a4c3ec
LP
643
644 * A new service systemd-time-sync-wait.service has been added. If
645 enabled it will delay the time-sync.target unit at boot until time
bc99dac5 646 synchronization has been received from the network. This
41a4c3ec
LP
647 functionality is useful on systems lacking a local RTC or where it is
648 acceptable that the boot process shall be delayed by external network
649 services.
650
651 * When hibernating, systemd will now inform the kernel of the image
652 write offset, on kernels new enough to support this. This means swap
653 files should work for hibernation now.
654
5cadf58e
ZJS
655 * When loading unit files, systemd will now look for drop-in unit files
656 extensions in additional places. Previously, for a unit file name
41a4c3ec
LP
657 "foo-bar-baz.service" it would look for dropin files in
658 "foo-bar-baz.service.d/*.conf". Now, it will also look in
659 "foo-bar-.service.d/*.conf" and "foo-.service.d/", i.e. at the
660 service name truncated after all inner dashes. This scheme allows
661 writing drop-ins easily that apply to a whole set of unit files at
662 once. It's particularly useful for mount and slice units (as their
5cadf58e
ZJS
663 naming is prefix based), but is also useful for service and other
664 units, for packages that install multiple unit files at once,
41a4c3ec 665 following a strict naming regime of beginning the unit file name with
5cadf58e
ZJS
666 the package's name. Two new specifiers are now supported in unit
667 files to match this: %j and %J are replaced by the part of the unit
668 name following the last dash.
669
670 * Unit files and other configuration files that support specifier
88099359 671 expansion now understand another three new specifiers: %T and %V will
5cadf58e 672 resolve to /tmp and /var/tmp respectively, or whatever temporary
88099359
ZJS
673 directory has been set for the calling user. %E will expand to either
674 /etc (for system units) or $XDG_CONFIG_HOME (for user units).
41a4c3ec
LP
675
676 * The ExecStart= lines of unit files are no longer required to
677 reference absolute paths. If non-absolute paths are specified the
678 specified binary name is searched within the service manager's
5cadf58e
ZJS
679 built-in $PATH, which may be queried with 'systemd-path
680 search-binaries-default'. It's generally recommended to continue to
681 use absolute paths for all binaries specified in unit files.
41a4c3ec 682
c7f93e28
ZJS
683 * Units gained a new load state "bad-setting", which is used when a
684 unit file was loaded, but contained fatal errors which prevent it
ba1dc1a1
LP
685 from being started (for example, a service unit has been defined
686 lacking both ExecStart= and ExecStop= lines).
c7f93e28 687
41a4c3ec
LP
688 * coredumpctl's "gdb" verb has been renamed to "debug", in order to
689 support alternative debuggers, for example lldb. The old name
690 continues to be available however, for compatibility reasons. Use the
5cadf58e
ZJS
691 new --debugger= switch or the $SYSTEMD_DEBUGGER environment variable
692 to pick an alternative debugger instead of the default gdb.
41a4c3ec
LP
693
694 * systemctl and the other tools will now output escape sequences that
695 generate proper clickable hyperlinks in various terminal emulators
696 where useful (for example, in the "systemctl status" output you can
697 now click on the unit file name to quickly open it in the
698 editor/viewer of your choice). Note that not all terminal emulators
699 support this functionality yet, but many do. Unfortunately, the
700 "less" pager doesn't support this yet, hence this functionality is
701 currently automatically turned off when a pager is started (which
702 happens quite often due to auto-paging). We hope to remove this
5cadf58e
ZJS
703 limitation as soon as "less" learns these escape sequences. This new
704 behaviour may also be turned off explicitly with the $SYSTEMD_URLIFY
705 environment variable. For details on these escape sequences see:
41a4c3ec
LP
706 https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda
707
708 * networkd's .network files now support a new IPv6MTUBytes= option for
709 setting the MTU used by IPv6 explicitly as well as a new MTUBytes=
710 option in the [Route] section to configure the MTU to use for
711 specific routes. It also gained support for configuration of the DHCP
712 "UserClass" option through the new UserClass= setting. It gained
713 three new options in the new [CAN] section for configuring CAN
714 networks. The MULTICAST and ALLMULTI interface flags may now be
715 controlled explicitly with the new Multicast= and AllMulticast=
716 settings.
717
718 * networkd will now automatically make use of the kernel's route
719 expiration feature, if it is available.
720
5cadf58e
ZJS
721 * udevd's .link files now support setting the number of receive and
722 transmit channels, using the RxChannels=, TxChannels=,
723 OtherChannels=, CombinedChannels= settings.
724
725 * Support for UDPSegmentationOffload= has been removed, given its
726 limited support in hardware, and waning software support.
41a4c3ec
LP
727
728 * networkd's .netdev files now support creating "netdevsim" interfaces.
729
730 * PID 1 learnt a new bus call GetUnitByControlGroup() which may be used
731 to query the unit belonging to a specific kernel control group.
732
5cadf58e 733 * systemd-analyze gained a new verb "cat-config", which may be used to
41a4c3ec
LP
734 dump the contents of any configuration file, with all its matching
735 drop-in files added in, and honouring the usual search and masking
736 logic applied to systemd configuration files. For example use
737 "systemd-analyze cat-config systemd/system.conf" to get the complete
738 system configuration file of systemd how it would be loaded by PID 1
5cadf58e
ZJS
739 itself. Similar to this, various tools such as systemd-tmpfiles or
740 systemd-sysusers, gained a new option "--cat-config", which does the
41a4c3ec
LP
741 corresponding operation for their own configuration settings. For
742 example, "systemd-tmpfiles --cat-config" will now output the full
743 list of tmpfiles.d/ lines in place.
744
704ae536
YW
745 * timedatectl gained three new verbs: "show" shows bus properties of
746 systemd-timedated, "timesync-status" shows the current NTP
747 synchronization state of systemd-timesyncd, and "show-timesync"
748 shows bus properties of systemd-timesyncd.
41a4c3ec
LP
749
750 * systemd-timesyncd gained a bus interface on which it exposes details
751 about its state.
752
73c718a9
YW
753 * A new environment variable $SYSTEMD_TIMEDATED_NTP_SERVICES is now
754 understood by systemd-timedated. It takes a colon-separated list of
755 unit names of NTP client services. The list is used by
756 "timedatectl set-ntp".
757
41a4c3ec
LP
758 * systemd-nspawn gained a new --rlimit= switch for setting initial
759 resource limits for the container payload. There's a new switch
5cadf58e 760 --hostname= to explicitly override the container's hostname. A new
41a4c3ec
LP
761 --no-new-privileges= switch may be used to control the
762 PR_SET_NO_NEW_PRIVS flag for the container payload. A new
763 --oom-score-adjust= switch controls the OOM scoring adjustment value
764 for the payload. The new --cpu-affinity= switch controls the CPU
765 affinity of the container payload. The new --resolv-conf= switch
766 allows more detailed control of /etc/resolv.conf handling of the
5cadf58e 767 container. Similarly, the new --timezone= switch allows more detailed
41a4c3ec
LP
768 control of /etc/localtime handling of the container.
769
5cadf58e 770 * systemd-detect-virt gained a new --list switch, which will print a
41a4c3ec
LP
771 list of all currently known VM and container environments.
772
5cadf58e 773 * Support for "Portable Services" has been added, see
41a4c3ec 774 doc/PORTABLE_SERVICES.md for details. Currently, the support is still
5cadf58e
ZJS
775 experimental, but this is expected to change soon. Reflecting this
776 experimental state, the "portablectl" binary is not installed into
41a4c3ec
LP
777 /usr/bin yet. The binary has to be called with the full path
778 /usr/lib/systemd/portablectl instead.
779
780 * journalctl's and systemctl's -o switch now knows a new log output
781 mode "with-unit". The output it generates is very similar to the
782 regular "short" mode, but displays the unit name instead of the
783 syslog tag for each log line. Also, the date is shown with timezone
784 information. This mode is probably more useful than the classic
785 "short" output mode for most purposes, except where pixel-perfect
786 compatibility with classic /var/log/messages formatting is required.
787
788 * A new --dump-bus-properties switch has been added to the systemd
789 binary, which may be used to dump all supported D-Bus properties.
c7f93e28
ZJS
790 (Options which are still supported, but are deprecated, are *not*
791 shown.)
41a4c3ec 792
41a4c3ec
LP
793 * sd-bus gained a set of new calls:
794 sd_bus_slot_set_floating()/sd_bus_slot_get_floating() may be used to
795 enable/disable the "floating" state of a bus slot object,
796 i.e. whether the slot object pins the bus it is allocated for into
797 memory or if the bus slot object gets disconnected when the bus goes
798 away. sd_bus_open_with_description(),
799 sd_bus_open_user_with_description(),
800 sd_bus_open_system_with_description() may be used to allocate bus
801 objects and set their description string already during allocation.
802
803 * sd-event gained support for watching inotify events from the event
804 loop, in an efficient way, sharing inotify handles between multiple
805 users. For this a new function sd_event_add_inotify() has been added.
806
807 * sd-event and sd-bus gained support for calling special user-supplied
808 destructor functions for userdata pointers associated with
c7f93e28
ZJS
809 sd_event_source, sd_bus_slot, and sd_bus_track objects. For this new
810 functions sd_bus_slot_set_destroy_callback,
811 sd_bus_slot_get_destroy_callback, sd_bus_track_set_destroy_callback,
812 sd_bus_track_get_destroy_callback,
813 sd_event_source_set_destroy_callback,
814 sd_event_source_get_destroy_callback have been added.
41a4c3ec
LP
815
816 * The "net.ipv4.tcp_ecn" sysctl will now be turned on by default.
817
818 * PID 1 will now automatically reschedule .timer units whenever the
5cadf58e 819 local timezone changes. (They previously got rescheduled
41a4c3ec
LP
820 automatically when the system clock changed.)
821
822 * New documentation has been added to document cgroups delegation,
823 portable services and the various code quality tools we have set up:
824
a8a27374
SK
825 https://github.com/systemd/systemd/blob/master/docs/CGROUP_DELEGATION.md
826 https://github.com/systemd/systemd/blob/master/docs/PORTABLE_SERVICES.md
827 https://github.com/systemd/systemd/blob/master/docs/CODE_QUALITY.md
41a4c3ec 828
d6906108
LP
829 * The Boot Loader Specification has been added to the source tree.
830
a8a27374 831 https://github.com/systemd/systemd/blob/master/docs/BOOT_LOADER_SPECIFICATION.md
d6906108
LP
832
833 While moving it into our source tree we have updated it and further
834 changes are now accepted through the usual github PR workflow.
835
41a4c3ec
LP
836 * pam_systemd will now look for PAM userdata fields systemd.memory_max,
837 systemd.tasks_max, systemd.cpu_weight, systemd.io_weight set by
838 earlier PAM modules. The data in these fields is used to initialize
839 the session scope's resource properties. Thus external PAM modules
840 may now configure per-session limits, for example sourced from
841 external user databases.
842
843 * socket units with Accept=yes will now maintain a "refused" counter in
844 addition to the existing "accepted" counter, counting connections
845 refused due to the enforced limits.
846
847 * The "systemd-path search-binaries-default" command may now be use to
848 query the default, built-in $PATH PID 1 will pass to the services it
849 manages.
850
c49a7cbd
LP
851 * A new unit file setting PrivateMounts= has been added. It's a boolean
852 option. If enabled the unit's processes are invoked in their own file
853 system namespace. Note that this behaviour is also implied if any
854 other file system namespacing options (such as PrivateTmp=,
855 PrivateDevices=, ProtectSystem=, …) are used. This option is hence
856 primarily useful for services that do not use any of the other file
857 system namespacing options. One such service is systemd-udevd.service
858 wher this is now used by default.
859
57ab451e
ZJS
860 * ConditionSecurity= gained a new value "uefi-secureboot" that is true
861 when the system is booted in UEFI "secure mode".
862
c7668c1c
LP
863 * A new unit "system-update-pre.target" is added, which defines an
864 optional synchronization point for offline system updates, as
865 implemented by the pre-existing "system-update.target" unit. It
866 allows ordering services before the service that executes the actual
867 update process in a generic way.
868
41a4c3ec 869 Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale,
ec53d48c 870 Alexander Kurtz, Alex Gartrell, Anssi Hannula, Arnaud Rebillout, Brian
bb6f071f
LP
871 J. Murrell, Bruno Vernay, Chris Lamb, Chris Lesiak, Christian Brauner,
872 Christian Hesse, Christian Rebischke, Colin Guthrie, Daniel Dao, Daniel
873 Lin, Danylo Korostil, Davide Cavalca, David Tardon, Dimitri John
874 Ledkov, Dmitriy Geels, Douglas Christman, Elia Geretto, emelenas, Emil
875 Velikov, Evgeny Vereshchagin, Felipe Sateler, Feng Sun, Filipe
876 Brandenburger, Franck Bui, futpib, Giuseppe Scrivano, Guillem Jover,
877 guixxx, Hannes Reinecke, Hans de Goede, Harald Hoyer, Henrique Dante de
878 Almeida, Hiram van Paassen, Ian Miell, Igor Gnatenko, Ivan Shapovalov,
879 Iwan Timmer, James Cowgill, Jan Janssen, Jan Synacek, Jared Kazimir,
880 Jérémy Rosen, João Paulo Rechi Vita, Joost Heitbrink, Jui-Chi Ricky
881 Liang, Jürg Billeter, Kai-Heng Feng, Karol Augustin, Kay Sievers,
882 Krzysztof Nowicki, Lauri Tirkkonen, Lennart Poettering, Leonard König,
883 Long Li, Luca Boccassi, Lucas Werkmeister, Marcel Hoppe, Marc
884 Kleine-Budde, Mario Limonciello, Martin Jansa, Martin Wilck, Mathieu
885 Malaterre, Matteo F. Vescovi, Matthew McGinn, Matthias-Christian Ott,
886 Michael Biebl, Michael Olbrich, Michael Prokop, Michal Koutný, Michal
887 Sekletar, Mike Gilbert, Mikhail Kasimov, Milan Broz, Milan Pässler,
888 Mladen Pejaković, Muhammet Kara, Nicolas Boichat, Omer Katz, Paride
889 Legovini, Paul Menzel, Paul Milliken, Pavel Hrdina, Peter A. Bigot,
890 Peter D'Hoye, Peter Hutterer, Peter Jones, Philip Sequeira, Philip
61d0025d 891 Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de Araujo,
bb6f071f
LP
892 Ronny Chevalier, Rosen Penev, Rubén Suárez Alvarez, Ryan Gonzalez,
893 Salvo Tomaselli, Sebastian Reichel, Sergey Ptashnick, Sergio Lindo
894 Mansilla, Stefan Schweter, Stephen Hemminger, Stuart Hayes, Susant
895 Sahani, Sylvain Plantefève, Thomas H. P. Andersen, Tobias Jungel,
896 Tomasz Torcz, Vito Caputo, Will Dietz, Will Thompson, Wim van Mourik,
897 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
41a4c3ec 898
e8498f82 899 — Berlin, 2018-06-22
41a4c3ec 900
c657bff1 901CHANGES WITH 238:
e0c46a73
LP
902
903 * The MemoryAccounting= unit property now defaults to on. After
904 discussions with the upstream control group maintainers we learnt
905 that the negative impact of cgroup memory accounting on current
906 kernels is finally relatively minimal, so that it should be safe to
444d5863
ZJS
907 enable this by default without affecting system performance. Besides
908 memory accounting only task accounting is turned on by default, all
909 other forms of resource accounting (CPU, IO, IP) remain off for now,
910 because it's not clear yet that their impact is small enough to move
911 from opt-in to opt-out. We recommend downstreams to leave memory
07a35e84 912 accounting on by default if kernel 4.14 or higher is primarily
444d5863
ZJS
913 used. On very resource constrained systems or when support for old
914 kernels is a necessity, -Dmemory-accounting-default=false can be used
915 to revert this change.
e0c46a73 916
313c32c3
ZJS
917 * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update,
918 %udev_rules_update) and the journal catalog (%journal_catalog_update)
919 from the upgrade scriptlets of individual packages now do nothing.
920 Transfiletriggers have been added which will perform those updates
921 once at the end of the transaction.
922
923 Similar transfiletriggers have been added to execute any sysctl.d
924 and binfmt.d rules. Thus, it should be unnecessary to provide any
925 scriptlets to execute this configuration from package installation
926 scripts.
927
928 * systemd-sysusers gained a mode where the configuration to execute is
929 specified on the command line, but this configuration is not executed
930 directly, but instead it is merged with the configuration on disk,
931 and the result is executed. This is useful for package installation
932 scripts which want to create the user before installing any files on
933 disk (in case some of those files are owned by that user), while
934 still allowing local admin overrides.
935
07a35e84 936 This functionality is exposed to rpm scriptlets through a new
313c32c3
ZJS
937 %sysusers_create_package macro. Old %sysusers_create and
938 %sysusers_create_inline macros are deprecated.
939
940 A transfiletrigger for sysusers.d configuration is now installed,
07a35e84 941 which means that it should be unnecessary to call systemd-sysusers from
313c32c3
ZJS
942 package installation scripts, unless the package installs any files
943 owned by those newly-created users, in which case
944 %sysusers_create_package should be used.
945
946 * Analogous change has been done for systemd-tmpfiles: it gained a mode
947 where the command-line configuration is merged with the configuration
948 on disk. This is exposed as the new %tmpfiles_create_package macro,
949 and %tmpfiles_create is deprecated. A transfiletrigger is installed
950 for tmpfiles.d, hence it should be unnecessary to call systemd-tmpfiles
951 from package installation scripts.
952
953 * sysusers.d configuration for a user may now also specify the group
954 number, in addition to the user number ("u username 123:456"), or
955 without the user number ("u username -:456").
956
957 * Configution items for systemd-sysusers can now be specified as
958 positional arguments when the new --inline switch is used.
959
960 * The login shell of users created through sysusers.d may now be
961 specified (previously, it was always /bin/sh for root and
962 /sbin/nologin for other users).
963
964 * systemd-analyze gained a new --global switch to look at global user
965 configuration. It also gained a unit-paths verb to list the unit load
966 paths that are compiled into systemd (which can be used with
967 --systemd, --user, or --global).
968
969 * udevadm trigger gained a new --settle/-w option to wait for any
970 triggered events to finish (but just those, and not any other events
971 which are triggered meanwhile).
972
973 * The action that systemd-logind takes when the lid is closed and the
974 machine is connected to external power can now be configured using
975 HandleLidSwitchExternalPower= in logind.conf. Previously, this action
976 was determined by HandleLidSwitch=, and, for backwards compatibility,
977 is still is, if HandleLidSwitchExternalPower= is not explicitly set.
978
979 * journalctl will periodically call sd_journal_process() to make it
980 resilient against inotify queue overruns when journal files are
981 rotated very quickly.
982
983 * Two new functions in libsystemd — sd_bus_get_n_queued_read and
984 sd_bus_get_n_queued_write — may be used to check the number of
985 pending bus messages.
986
987 * systemd gained a new
988 org.freedesktop.systemd1.Manager.AttachProcessesToUnit dbus call
989 which can be used to migrate foreign processes to scope and service
990 units. The primary user for this new API is systemd itself: the
991 systemd --user instance uses this call of the systemd --system
992 instance to migrate processes if it itself gets the request to
993 migrate processes and the kernel refuses this due to access
994 restrictions. Thanks to this "systemd-run --scope --user …" works
4e1dfa45 995 again in pure cgroup v2 environments when invoked from the user
313c32c3
ZJS
996 session scope.
997
998 * A new TemporaryFileSystem= setting can be used to mask out part of
999 the real file system tree with tmpfs mounts. This may be combined
1000 with BindPaths= and BindReadOnlyPaths= to hide files or directories
1001 not relevant to the unit, while still allowing some paths lower in
1002 the tree to be accessed.
1003
1004 ProtectHome=tmpfs may now be used to hide user home and runtime
1005 directories from units, in a way that is mostly equivalent to
1006 "TemporaryFileSystem=/home /run/user /root".
1007
1008 * Non-service units are now started with KeyringMode=shared by default.
1009 This means that mount and swapon and other mount tools have access
1010 to keys in the main keyring.
1011
1012 * /sys/fs/bpf is now mounted automatically.
1013
1014 * QNX virtualization is now detected by systemd-detect-virt and may
1015 be used in ConditionVirtualization=.
1016
1017 * IPAccounting= may now be enabled also for slice units.
1018
1019 * A new -Dsplit-bin= build configuration switch may be used to specify
1020 whether bin and sbin directories are merged, or if they should be
1021 included separately in $PATH and various listings of executable
1022 directories. The build configuration scripts will try to autodetect
1023 the proper values of -Dsplit-usr= and -Dsplit-bin= based on build
1024 system, but distributions are encouraged to configure this
1025 explicitly.
1026
1027 * A new -Dok-color= build configuration switch may be used to change
1028 the colour of "OK" status messages.
1029
1030 * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with
1031 PrivateNetwork=yes was buggy in previous versions of systemd. This
1032 means that after the upgrade and daemon-reexec, any such units must
1033 be restarted.
1034
1035 * INCOMPATIBILITY: as announced in the NEWS for 237, systemd-tmpfiles
1036 will not exclude read-only files owned by root from cleanup.
1037
c657bff1
ZJS
1038 Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet,
1039 Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337,
1040 Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo
1041 de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel
1042 Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny
1043 Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib,
1044 Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer,
1045 Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld,
1046 Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas
1047 Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt,
1048 MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren,
1049 Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert
1050 Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
1051 Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain
1052 Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe,
1053 Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić)
1054
1055 — Warsaw, 2018-03-05
1056
82c8e3e6 1057CHANGES WITH 237:
2b0c59ba
MP
1058
1059 * Some keyboards come with a zoom see-saw or rocker which until now got
1060 mapped to the Linux "zoomin/out" keys in hwdb. However, these
1061 keycodes are not recognized by any major desktop. They now produce
1062 Up/Down key events so that they can be used for scrolling.
1063
49e87292
LP
1064 * INCOMPATIBILITY: systemd-tmpfiles' "f" lines changed behaviour
1065 slightly: previously, if an argument was specified for lines of this
1066 type (i.e. the right-most column was set) this string was appended to
1067 existing files each time systemd-tmpfiles was run. This behaviour was
1068 different from what the documentation said, and not particularly
1069 useful, as repeated systemd-tmpfiles invocations would not be
1070 idempotent and grow such files without bounds. With this release
15c5594b
ZJS
1071 behaviour has been altered to match what the documentation says:
1072 lines of this type only have an effect if the indicated files don't
1073 exist yet, and only then the argument string is written to the file.
49e87292 1074
82c8e3e6
LP
1075 * FUTURE INCOMPATIBILITY: In systemd v238 we intend to slightly change
1076 systemd-tmpfiles behaviour: previously, read-only files owned by root
1077 were always excluded from the file "aging" algorithm (i.e. the
1078 automatic clean-up of directories like /tmp based on
1079 atime/mtime/ctime). We intend to drop this restriction, and age files
1080 by default even when owned by root and read-only. This behaviour was
1081 inherited from older tools, but there have been requests to remove
1082 it, and it's not obvious why this restriction was made in the first
1083 place. Please speak up now, if you are aware of software that reqires
1084 this behaviour, otherwise we'll remove the restriction in v238.
1085
95894b91
LP
1086 * A new environment variable $SYSTEMD_OFFLINE is now understood by
1087 systemctl. It takes a boolean argument. If on, systemctl assumes it
1088 operates on an "offline" OS tree, and will not attempt to talk to the
1089 service manager. Previously, this mode was implicitly enabled if a
1090 chroot() environment was detected, and this new environment variable
1091 now provides explicit control.
1092
1a0cd2c7
ZJS
1093 * .path and .socket units may now be created transiently, too.
1094 Previously only service, mount, automount and timer units were
95894b91
LP
1095 supported as transient units. The systemd-run tool has been updated
1096 to expose this new functionality, you may hence use it now to bind
1097 arbitrary commands to path or socket activation on-the-fly from the
1a0cd2c7
ZJS
1098 command line. Moreover, almost all properties are now exposed for the
1099 unit types that already supported transient operation.
95894b91
LP
1100
1101 * The systemd-mount command gained support for a new --owner= parameter
1102 which takes a user name, which is then resolved and included in uid=
1103 and gid= mount options string of the file system to mount.
1104
1105 * A new unit condition ConditionControlGroupController= has been added
1106 that checks whether a specific cgroup controller is available.
1107
1108 * Unit files, udev's .link files, and systemd-networkd's .netdev and
1109 .network files all gained support for a new condition
1110 ConditionKernelVersion= for checking against specific kernel
1111 versions.
1112
1113 * In systemd-networkd, the [IPVLAN] section in .netdev files gained
6cddc792 1114 support for configuring device flags in the Flags= setting. In the
95894b91
LP
1115 same files, the [Tunnel] section gained support for configuring
1116 AllowLocalRemote=. The [Route] section in .network files gained
1117 support for configuring InitialCongestionWindow=,
1118 InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now
1119 understands RapidCommit=.
1120
1121 * systemd-networkd's DHCPv6 support gained support for Prefix
1122 Delegation.
1123
1124 * sd-bus gained support for a new "watch-bind" feature. When this
1125 feature is enabled, an sd_bus connection may be set up to connect to
1126 an AF_UNIX socket in the file system as soon as it is created. This
1127 functionality is useful for writing early-boot services that
1128 automatically connect to the system bus as soon as it is started,
1129 without ugly time-based polling. systemd-networkd and
1130 systemd-resolved have been updated to make use of this
1131 functionality. busctl exposes this functionality in a new
1132 --watch-bind= command line switch.
1133
1134 * sd-bus will now optionally synthesize a local "Connected" signal as
1135 soon as a D-Bus connection is set up fully. This message mirrors the
1136 already existing "Disconnected" signal which is synthesized when the
1137 connection is terminated. This signal is generally useful but
1138 particularly handy in combination with the "watch-bind" feature
1139 described above. Synthesizing of this message has to be requested
1140 explicitly through the new API call sd_bus_set_connected_signal(). In
1141 addition a new call sd_bus_is_ready() has been added that checks
caf2a2d8 1142 whether a connection is fully set up (i.e. between the "Connected" and
95894b91
LP
1143 "Disconnected" signals).
1144
1145 * sd-bus gained two new calls sd_bus_request_name_async() and
1146 sd_bus_release_name_async() for asynchronously registering bus
1147 names. Similar, there is now sd_bus_add_match_async() for installing
1148 a signal match asynchronously. All of systemd's own services have
1149 been updated to make use of these calls. Doing these operations
1150 asynchronously has two benefits: it reduces the risk of deadlocks in
1151 case of cyclic dependencies between bus services, and it speeds up
1152 service initialization since synchronization points for bus
1153 round-trips are removed.
1154
1155 * sd-bus gained two new calls sd_bus_match_signal() and
1156 sd_bus_match_signal_async(), which are similar to sd_bus_add_match()
1157 and sd_bus_add_match_async() but instead of taking a D-Bus match
1158 string take match fields as normal function parameters.
1159
1160 * sd-bus gained two new calls sd_bus_set_sender() and
1161 sd_bus_message_set_sender() for setting the sender name of outgoing
1162 messages (either for all outgoing messages or for just one specific
1163 one). These calls are only useful in direct connections as on
1164 brokered connections the broker fills in the sender anyway,
1165 overwriting whatever the client filled in.
1166
1167 * sd-event gained a new pseudo-handle that may be specified on all API
1168 calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When
1169 used this refers to the default event loop object of the calling
1170 thread. Note however that this does not implicitly allocate one —
6cddc792
CR
1171 which has to be done prior by using sd_event_default(). Similarly
1172 sd-bus gained three new pseudo-handles SD_BUS_DEFAULT,
95894b91
LP
1173 SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer
1174 to the default bus of the specified type of the calling thread. Here
1175 too this does not implicitly allocate bus connection objects, this
1176 has to be done prior with sd_bus_default() and friends.
1177
1178 * sd-event gained a new call pair
6cddc792
CR
1179 sd_event_source_{get|set}_io_fd_own(). This may be used to request
1180 automatic closure of the file descriptor an IO event source watches
95894b91
LP
1181 when the event source is destroyed.
1182
1183 * systemd-networkd gained support for natively configuring WireGuard
1184 connections.
1185
6cddc792
CR
1186 * In previous versions systemd synthesized user records both for the
1187 "nobody" (UID 65534) and "root" (UID 0) users in nss-systemd and
1188 internally. In order to simplify distribution-wide renames of the
95894b91
LP
1189 "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a
1190 new transitional flag file has been added: if
1191 /etc/systemd/dont-synthesize-nobody exists synthesizing of the 65534
1192 user and group record within the systemd codebase is disabled.
1193
1194 * systemd-notify gained a new --uid= option for selecting the source
1195 user/UID to use for notification messages sent to the service
1196 manager.
1197
31751f7e 1198 * journalctl gained a new --grep= option to list only entries in which
e6501af8
ZJS
1199 the message matches a certain pattern. By default matching is case
1200 insensitive if the pattern is lowercase, and case sensitive
1201 otherwise. Option --case-sensitive=yes|no can be used to override
1202 this an specify case sensitivity or case insensitivity.
1203
56a29112 1204 * There's now a "systemd-analyze service-watchdogs" command for printing
508058c9 1205 the current state of the service runtime watchdog, and optionally
56a29112 1206 enabling or disabling the per-service watchdogs system-wide if given a
508058c9
LP
1207 boolean argument (i.e. the concept you configure in WatchdogSec=), for
1208 debugging purposes. There's also a kernel command line option
56a29112 1209 systemd.service_watchdogs= for controlling the same.
508058c9
LP
1210
1211 * Two new "log-level" and "log-target" options for systemd-analyze were
bc99dac5 1212 added that merge the now deprecated get-log-level, set-log-level and
508058c9
LP
1213 get-log-target, set-log-target pairs. The deprecated options are still
1214 understood for backwards compatibility. The two new options print the
1215 current value when no arguments are given, and set them when a
56a29112 1216 level/target is given as an argument.
95894b91 1217
508058c9
LP
1218 * sysusers.d's "u" lines now optionally accept both a UID and a GID
1219 specification, separated by a ":" character, in order to create users
1220 where UID and GID do not match.
1221
95894b91 1222 Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov,
508058c9
LP
1223 Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman
1224 Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton
1225 Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov,
1226 Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui,
1227 Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian
1228 Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander
1229 Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen,
1230 Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg
1231 Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering,
1232 Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt,
1233 Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy,
1234 Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał
1235 Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf
1236 Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer,
1237 Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer,
1238 Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani,
1239 Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz
1240 Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary
1241 Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян
1242 Палаузов
1243
1244 — Brno, 2018-01-28
2b0c59ba 1245
a1b2c92d 1246CHANGES WITH 236:
195b943d 1247
89780840
ZJS
1248 * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
1249 in v235 has been extended to also set the dummy.ko module option
1250 numdummies=0, preventing the kernel from automatically creating
1251 dummy0. All dummy interfaces must now be explicitly created.
195b943d 1252
3925496a
LP
1253 * Unknown '%' specifiers in configuration files are now rejected. This
1254 applies to units and tmpfiles.d configuration. Any percent characters
1255 that are followed by a letter or digit that are not supposed to be
1256 interpreted as the beginning of a specifier should be escaped by
1257 doubling ("%%"). (So "size=5%" is still accepted, as well as
1258 "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
1259 valid specifiers today.)
751223fe 1260
e6b2d948 1261 * systemd-resolved now maintains a new dynamic
89780840
ZJS
1262 /run/systemd/resolve/stub-resolv.conf compatibility file. It is
1263 recommended to make /etc/resolv.conf a symlink to it. This file
1264 points at the systemd-resolved stub DNS 127.0.0.53 resolver and
1265 includes dynamically acquired search domains, achieving more correct
1266 DNS resolution by software that bypasses local DNS APIs such as NSS.
e6b2d948 1267
67eb5b38
LP
1268 * The "uaccess" udev tag has been dropped from /dev/kvm and
1269 /dev/dri/renderD*. These devices now have the 0666 permissions by
1270 default (but this may be changed at build-time). /dev/dri/renderD*
1271 will now be owned by the "render" group along with /dev/kfd.
1272
89780840
ZJS
1273 * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
1274 systemd-journal-gatewayd.service and
1275 systemd-journal-upload.service. This means "nss-systemd" must be
1276 enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
1277 services are resolved properly.
67eb5b38 1278
3925496a
LP
1279 * In /etc/fstab two new mount options are now understood:
1280 x-systemd.makefs and x-systemd.growfs. The former has the effect that
1281 the configured file system is formatted before it is mounted, the
1282 latter that the file system is resized to the full block device size
1283 after it is mounted (i.e. if the file system is smaller than the
1284 partition it resides on, it's grown). This is similar to the fsck
1285 logic in /etc/fstab, and pulls in systemd-makefs@.service and
1286 systemd-growfs@.service as necessary, similar to
1287 systemd-fsck@.service. Resizing is currently only supported on ext4
1288 and btrfs.
1289
67eb5b38
LP
1290 * In systemd-networkd, the IPv6 RA logic now optionally may announce
1291 DNS server and domain information.
1292
1293 * Support for the LUKS2 on-disk format for encrypted partitions has
1294 been added. This requires libcryptsetup2 during compilation and
1295 runtime.
1296
89780840 1297 * The systemd --user instance will now signal "readiness" when its
67eb5b38
LP
1298 basic.target unit has been reached, instead of when the run queue ran
1299 empty for the first time.
1300
8ea2dcb0
ZJS
1301 * Tmpfiles.d with user configuration are now also supported.
1302 systemd-tmpfiles gained a new --user switch, and snippets placed in
1303 ~/.config/user-tmpfiles.d/ and corresponding directories will be
1304 executed by systemd-tmpfiles --user running in the new
1305 systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
1306 running in the user session.
1307
1308 * Unit files and tmpfiles.d snippets learnt three new % specifiers:
1309 %S resolves to the top-level state directory (/var/lib for the system
1310 instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
1311 top-level cache directory (/var/cache for the system instance,
1312 $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
1313 logs directory (/var/log for the system instance,
67eb5b38 1314 $XDG_CONFIG_HOME/log/ for the user instance). This matches the
8ea2dcb0 1315 existing %t specifier, that resolves to the top-level runtime
67eb5b38
LP
1316 directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
1317 user instance).
1318
1319 * journalctl learnt a new parameter --output-fields= for limiting the
1320 set of journal fields to output in verbose and JSON output modes.
1321
1322 * systemd-timesyncd's configuration file gained a new option
89780840
ZJS
1323 RootDistanceMaxSec= for setting the maximum root distance of servers
1324 it'll use, as well as the new options PollIntervalMinSec= and
1325 PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
67eb5b38
LP
1326
1327 * bootctl gained a new command "list" for listing all available boot
89780840 1328 menu items on systems that follow the boot loader specification.
67eb5b38
LP
1329
1330 * systemctl gained a new --dry-run switch that shows what would be done
1331 instead of doing it, and is currently supported by the shutdown and
1332 sleep verbs.
1333
e9ad86d5 1334 * ConditionSecurity= can now detect the TOMOYO security module.
67eb5b38
LP
1335
1336 * Unit file [Install] sections are now also respected in unit drop-in
89780840 1337 files. This is intended to be used by drop-ins under /usr/lib/.
67eb5b38 1338
89780840 1339 * systemd-firstboot may now also set the initial keyboard mapping.
67eb5b38 1340
89780840
ZJS
1341 * Udev "changed" events for devices which are exposed as systemd
1342 .device units are now propagated to units specified in
1343 ReloadPropagatedFrom= as reload requests.
67eb5b38 1344
89780840
ZJS
1345 * If a udev device has a SYSTEMD_WANTS= property containing a systemd
1346 unit template name (i.e. a name in the form of 'foobar@.service',
1347 without the instance component between the '@' and - the '.'), then
1348 the escaped sysfs path of the device is automatically used as the
1349 instance.
67eb5b38
LP
1350
1351 * SystemCallFilter= in unit files has been extended so that an "errno"
1352 can be specified individually for each system call. Example:
1353 SystemCallFilter=~uname:EILSEQ.
1354
1355 * The cgroup delegation logic has been substantially updated. Delegate=
1356 now optionally takes a list of controllers (instead of a boolean, as
1357 before), which lists the controllers to delegate at least.
1358
89780840 1359 * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
67eb5b38 1360
89780840
ZJS
1361 * A new LogLevelMax= setting configures the maximum log level any
1362 process of the service may log at (i.e. anything with a lesser
1363 priority than what is specified is automatically dropped). A new
1364 LogExtraFields= setting allows configuration of additional journal
1365 fields to attach to all log records generated by any of the unit's
1366 processes.
67eb5b38 1367
89780840
ZJS
1368 * New StandardInputData= and StandardInputText= settings along with the
1369 new option StandardInput=data may be used to configure textual or
1370 binary data that shall be passed to the executed service process via
1371 standard input, encoded in-line in the unit file.
67eb5b38
LP
1372
1373 * StandardInput=, StandardOutput= and StandardError= may now be used to
1374 connect stdin/stdout/stderr of executed processes directly with a
1375 file or AF_UNIX socket in the file system, using the new "file:" option.
1376
89780840
ZJS
1377 * A new unit file option CollectMode= has been added, that allows
1378 tweaking the garbage collection logic for units. It may be used to
1379 tell systemd to garbage collect units that have failed automatically
1380 (normally it only GCs units that exited successfully). systemd-run
1381 and systemd-mount expose this new functionality with a new -G option.
1382
67eb5b38
LP
1383 * "machinectl bind" may now be used to bind mount non-directories
1384 (i.e. regularfiles, devices, fifos, sockets).
1385
1386 * systemd-analyze gained a new verb "calendar" for validating and
1387 testing calendar time specifications to use for OnCalendar= in timer
1388 units. Besides validating the expression it will calculate the next
1389 time the specified expression would elapse.
1390
1391 * In addition to the pre-existing FailureAction= unit file setting
89780840
ZJS
1392 there's now SuccessAction=, for configuring a shutdown action to
1393 execute when a unit completes successfully. This is useful in
1394 particular inside containers that shall terminate after some workload
1395 has been completed. Also, both options are now supported for all unit
1396 types, not just services.
67eb5b38
LP
1397
1398 * networkds's IP rule support gained two new options
dd014eeb 1399 IncomingInterface= and OutgoingInterface= for configuring the incoming
67eb5b38
LP
1400 and outgoing interfaces of configured rules. systemd-networkd also
1401 gained support for "vxcan" network devices.
1402
1403 * networkd gained a new setting RequiredForOnline=, taking a
1404 boolean. If set, systemd-wait-online will take it into consideration
1405 when determining that the system is up, otherwise it will ignore the
1406 interface for this purpose.
1407
1408 * The sd_notify() protocol gained support for a new operation: with
1409 FDSTOREREMOVE=1 file descriptors may be removed from the per-service
1410 store again, ahead of POLLHUP or POLLERR when they are removed
1411 anyway.
1412
f09eb768
LP
1413 * A new document doc/UIDS-GIDS.md has been added to the source tree,
1414 that documents the UID/GID range and assignment assumptions and
3925496a
LP
1415 requirements of systemd.
1416
1417 * The watchdog device PID 1 will ping may now be configured through the
1418 WatchdogDevice= configuration file setting, or by setting the
1419 systemd.watchdog_service= kernel commandline option.
1420
1421 * systemd-resolved's gained support for registering DNS-SD services on
1422 the local network using MulticastDNS. Services may either be
1423 registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
1424 the same dir below /run, /usr/lib), or through its D-Bus API.
1425
a327431b
DB
1426 * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
1427 extend the effective start, runtime, and stop time. The service must
1428 continue to send EXTEND_TIMEOUT_USEC within the period specified to
1429 prevent the service manager from making the service as timedout.
1430
ea2a3c9e
LP
1431 * systemd-resolved's DNSSEC support gained support for RFC 8080
1432 (Ed25519 keys and signatures).
1433
a1b2c92d
LP
1434 * The systemd-resolve command line tool gained a new set of options
1435 --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
1436 --set-nta= and --revert to configure per-interface DNS configuration
1437 dynamically during runtime. It's useful for pushing DNS information
1438 into systemd-resolved from DNS hook scripts that various interface
1439 managing software supports (such as pppd).
1440
1441 * systemd-nspawn gained a new --network-namespace-path= command line
1442 option, which may be used to make a container join an existing
1443 network namespace, by specifying a path to a "netns" file.
1444
3925496a
LP
1445 Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
1446 Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
1447 Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
ea2a3c9e 1448 Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
a1b2c92d
LP
1449 John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
1450 Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
1451 Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
1452 Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
1453 Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
1454 Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
1455 Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
1456 Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
1457 Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
1458 Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
1459 Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
1460 Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
1461 Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
1462 Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
1463 Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
1464 Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
1465 Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
1466 Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
1467 Jędrzejewski-Szmek, Zeal Jagannatha
67eb5b38 1468
ea2a3c9e 1469 — Berlin, 2017-12-14
3754abc5 1470
582faeb4
DJL
1471CHANGES WITH 235:
1472
2bcbffd6
LP
1473 * INCOMPATIBILITY: systemd-logind.service and other long-running
1474 services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
1475 communication with the outside. This generally improves security of
1476 the system, and is in almost all cases a safe and good choice, as
23d37367 1477 these services do not and should not provide any network-facing
2bcbffd6
LP
1478 functionality. However, systemd-logind uses the glibc NSS API to
1479 query the user database. This creates problems on systems where NSS
1480 is set up to directly consult network services for user database
1481 lookups. In particular, this creates incompatibilities with the
1482 "nss-nis" module, which attempts to directly contact the NIS/YP
1483 network servers it is configured for, and will now consistently
1484 fail. In such cases, it is possible to turn off IP sandboxing for
1485 systemd-logind.service (set IPAddressDeny= in its [Service] section
1486 to the empty string, via a .d/ unit file drop-in). Downstream
1487 distributions might want to update their nss-nis packaging to include
1488 such a drop-in snippet, accordingly, to hide this incompatibility
1489 from the user. Another option is to make use of glibc's nscd service
1490 to proxy such network requests through a privilege-separated, minimal
1491 local caching daemon, or to switch to more modern technologies such
1492 sssd, whose NSS hook-ups generally do not involve direct network
1493 access. In general, we think it's definitely time to question the
1494 implementation choices of nss-nis, i.e. whether it's a good idea
1495 today to embed a network-facing loadable module into all local
1496 processes that need to query the user database, including the most
1497 trivial and benign ones, such as "ls". For more details about
1498 IPAddressDeny= see below.
1499
fccf5419
LP
1500 * A new modprobe.d drop-in is now shipped by default that sets the
1501 bonding module option max_bonds=0. This overrides the kernel default,
1502 to avoid conflicts and ambiguity as to whether or not bond0 should be
1503 managed by systemd-networkd or not. This resolves multiple issues
1504 with bond0 properties not being applied, when bond0 is configured
1505 with systemd-networkd. Distributors may choose to not package this,
1506 however in that case users will be prevented from correctly managing
1507 bond0 interface using systemd-networkd.
582faeb4 1508
ef5a8cb1 1509 * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
21723f53
ZJS
1510 which print the logging level and target of the system manager. They
1511 complement the existing "set-log-level" and "set-log-target" verbs
1512 used to change those values.
ef5a8cb1 1513
fccf5419
LP
1514 * journald.conf gained a new boolean setting ReadKMsg= which defaults
1515 to on. If turned off kernel log messages will not be read by
21723f53
ZJS
1516 systemd-journald or included in the logs. It also gained a new
1517 setting LineMax= for configuring the maximum line length in
1518 STDOUT/STDERR log streams. The new default for this value is 48K, up
1519 from the previous hardcoded 2048.
fccf5419 1520
21723f53
ZJS
1521 * A new unit setting RuntimeDirectoryPreserve= has been added, which
1522 allows more detailed control of what to do with a runtime directory
1523 configured with RuntimeDirectory= (i.e. a directory below /run or
1524 $XDG_RUNTIME_DIR) after a unit is stopped.
fccf5419
LP
1525
1526 * The RuntimeDirectory= setting for units gained support for creating
1527 deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
1528 one top-level directory.
1529
1530 * Units gained new options StateDirectory=, CacheDirectory=,
1531 LogsDirectory= and ConfigurationDirectory= which are closely related
1532 to RuntimeDirectory= but manage per-service directories below
21723f53 1533 /var/lib, /var/cache, /var/log and /etc. By making use of them it is
fccf5419
LP
1534 possible to write unit files which when activated automatically gain
1535 properly owned service specific directories in these locations, thus
1536 making unit files self-contained and increasing compatibility with
1537 stateless systems and factory reset where /etc or /var are
1538 unpopulated at boot. Matching these new settings there's also
1539 StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
1540 ConfigurationDirectoryMode= for configuring the access mode of these
75dfbbac
LP
1541 directories. These settings are particularly useful in combination
1542 with DynamicUser=yes as they provide secure, properly-owned,
1543 writable, and stateful locations for storage, excluded from the
1544 sandbox that such services live in otherwise.
fccf5419
LP
1545
1546 * Automake support has been removed from this release. systemd is now
1547 Meson-only.
1548
1549 * systemd-journald will now aggressively cache client metadata during
1550 runtime, speeding up log write performance under pressure. This comes
1551 at a small price though: as much of the metadata is read
1552 asynchronously from /proc/ (and isn't implicitly attached to log
1553 datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
1554 metadata stored alongside a log entry might be slightly
1555 out-of-date. Previously it could only be slightly newer than the log
1556 message. The time window is small however, and given that the kernel
1557 is unlikely to be improved anytime soon in this regard, this appears
1558 acceptable to us.
1559
1560 * nss-myhostname/systemd-resolved will now by default synthesize an
1561 A/AAAA resource record for the "_gateway" hostname, pointing to the
1562 current default IP gateway. Previously it did that for the "gateway"
1563 name, hampering adoption, as some distributions wanted to leave that
1564 host name open for local use. The old behaviour may still be
1565 requested at build time.
1566
1567 * systemd-networkd's [Address] section in .network files gained a new
1568 Scope= setting for configuring the IP address scope. The [Network]
1569 section gained a new boolean setting ConfigureWithoutCarrier= that
1570 tells systemd-networkd to ignore link sensing when configuring the
1571 device. The [DHCP] section gained a new Anonymize= boolean option for
1572 turning on a number of options suggested in RFC 7844. A new
1573 [RoutingPolicyRule] section has been added for configuring the IP
1574 routing policy. The [Route] section has gained support for a new
1575 Type= setting which permits configuring
1576 blackhole/unreachable/prohibit routes.
1577
1578 * The [VRF] section in .netdev files gained a new Table= setting for
1579 configuring the routing table to use. The [Tunnel] section gained a
1580 new Independent= boolean field for configuring tunnels independent of
1581 an underlying network interface. The [Bridge] section gained a new
1582 GroupForwardMask= option for configuration of propagation of link
1583 local frames between bridge ports.
1584
1585 * The WakeOnLan= setting in .link files gained support for a number of
1586 new modes. A new TCP6SegmentationOffload= setting has been added for
1587 configuring TCP/IPv6 hardware segmentation offload.
1588
1589 * The IPv6 RA sender implementation may now optionally send out RDNSS
21723f53 1590 and RDNSSL records to supply DNS configuration to peers.
fccf5419
LP
1591
1592 * systemd-nspawn gained support for a new --system-call-filter= command
21723f53
ZJS
1593 line option for adding and removing entries in the default system
1594 call filter it applies. Moreover systemd-nspawn has been changed to
fccf5419
LP
1595 implement a system call whitelist instead of a blacklist.
1596
1597 * systemd-run gained support for a new --pipe command line option. If
1598 used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
1599 are directly passed on to the activated transient service
21723f53
ZJS
1600 executable. This allows invoking arbitrary processes as systemd
1601 services (for example to take benefit of dependency management,
1602 accounting management, resource management or log management that is
1603 done automatically for services) — while still allowing them to be
fccf5419
LP
1604 integrated in a classic UNIX shell pipeline.
1605
1606 * When a service sends RELOAD=1 via sd_notify() and reload propagation
1607 using ReloadPropagationTo= is configured, a reload is now propagated
1608 to configured units. (Previously this was only done on explicitly
1609 requested reloads, using "systemctl reload" or an equivalent
1610 command.)
1611
1612 * For each service unit a restart counter is now kept: it is increased
1613 each time the service is restarted due to Restart=, and may be
1614 queried using "systemctl show -p NRestarts …".
1615
44898c53
LP
1616 * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
1617 @signal and @timer have been added, for usage with SystemCallFilter=
fccf5419
LP
1618 in unit files and the new --system-call-filter= command line option
1619 of systemd-nspawn (see above).
1620
1621 * ExecStart= lines in unit files gained two new modifiers: when a
1622 command line is prefixed with "!" the command will be executed as
1623 configured, except for the credentials applied by
1624 setuid()/setgid()/setgroups(). It is very similar to the pre-existing
1625 "+", but does still apply namespacing options unlike "+". There's
1626 also "!!" now, which is mostly identical, but becomes a NOP on
1627 systems that support ambient capabilities. This is useful to write
1628 unit files that work with ambient capabilities where possible but
1629 automatically fall back to traditional privilege dropping mechanisms
1630 on systems where this is not supported.
1631
1632 * ListenNetlink= settings in socket units now support RDMA netlink
1633 sockets.
1634
1635 * A new unit file setting LockPersonality= has been added which permits
1636 locking down the chosen execution domain ("personality") of a service
1637 during runtime.
1638
1639 * A new special target "getty-pre.target" has been added, which is
1640 ordered before all text logins, and may be used to order services
21723f53 1641 before textual logins acquire access to the console.
fccf5419
LP
1642
1643 * systemd will now attempt to load the virtio-rng.ko kernel module very
1644 early on if a VM environment supporting this is detected. This should
1645 improve entropy during early boot in virtualized environments.
1646
1647 * A _netdev option is now supported in /etc/crypttab that operates in a
1648 similar way as the same option in /etc/fstab: it permits configuring
21723f53
ZJS
1649 encrypted devices that need to be ordered after the network is up.
1650 Following this logic, two new special targets
fccf5419 1651 remote-cryptsetup-pre.target and remote-cryptsetup.target have been
21723f53
ZJS
1652 added that are to cryptsetup.target what remote-fs.target and
1653 remote-fs-pre.target are to local-fs.target.
fccf5419
LP
1654
1655 * Service units gained a new UnsetEnvironment= setting which permits
21723f53
ZJS
1656 unsetting specific environment variables for services that are
1657 normally passed to it (for example in order to mask out locale
fccf5419
LP
1658 settings for specific services that can't deal with it).
1659
1660 * Units acquired a new boolean option IPAccounting=. When turned on, IP
1661 traffic accounting (packet count as well as byte count) is done for
1662 the service, and shown as part of "systemctl status" or "systemd-run
1663 --wait".
1664
1665 * Service units acquired two new options IPAddressAllow= and
1666 IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
1667 for configuring a simple IP access control list for all sockets of
1668 the unit. These options are available also on .slice and .socket
1669 units, permitting flexible access list configuration for individual
1670 services as well as groups of services (as defined by a slice unit),
1671 including system-wide. Note that IP ACLs configured this way are
1672 enforced on every single IPv4 and IPv6 socket created by any process
1673 of the service unit, and apply to ingress as well as egress traffic.
1674
21723f53 1675 * If CPUAccounting= or IPAccounting= is turned on for a unit a new
608f70e6 1676 structured log message is generated each time the unit is stopped,
fccf5419
LP
1677 containing information about the consumed resources of this
1678 invocation.
1679
1680 * A new setting KeyringMode= has been added to unit files, which may be
1681 used to control how the kernel keyring is set up for executed
1682 processes.
1683
e06fafb2
LP
1684 * "systemctl poweroff", "systemctl reboot", "systemctl halt",
1685 "systemctl kexec" and "systemctl exit" are now always asynchronous in
1686 behaviour (that is: these commands return immediately after the
21723f53
ZJS
1687 operation was enqueued instead of waiting for the operation to
1688 complete). Previously, "systemctl poweroff" and "systemctl reboot"
e06fafb2
LP
1689 were asynchronous on systems using systemd-logind (i.e. almost
1690 always, and like they were on sysvinit), and the other three commands
1691 were unconditionally synchronous. With this release this is cleaned
1692 up, and callers will see the same asynchronous behaviour on all
1693 systems for all five operations.
1694
1695 * systemd-logind gained new Halt() and CanHalt() bus calls for halting
1696 the system.
1697
fccf5419
LP
1698 * .timer units now accept calendar specifications in other timezones
1699 than UTC or the local timezone.
1700
f6e64b78 1701 * The tmpfiles snippet var.conf has been changed to create
21723f53
ZJS
1702 /var/log/btmp with access mode 0660 instead of 0600. It was owned by
1703 the "utmp" group already, and it appears to be generally understood
1704 that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
1705 databases. Previously this was implemented correctly for all these
1706 databases excepts btmp, which has been opened up like this now
1707 too. Note that while the other databases are world-readable
1708 (i.e. 0644), btmp is not and remains more restrictive.
f6e64b78 1709
d55b0463
LP
1710 * The systemd-resolve tool gained a new --reset-server-features
1711 switch. When invoked like this systemd-resolved will forget
1712 everything it learnt about the features supported by the configured
1713 upstream DNS servers, and restarts the feature probing logic on the
cf84484a
LP
1714 next resolver look-up for them at the highest feature level
1715 again.
1716
1717 * The status dump systemd-resolved sends to the logs upon receiving
1718 SIGUSR1 now also includes information about all DNS servers it is
1719 configured to use, and the features levels it probed for them.
d55b0463 1720
fccf5419
LP
1721 Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
1722 Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
76451c1d
LP
1723 Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
1724 Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
1725 Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
1726 Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
1727 ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
1728 Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
1729 Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
1730 John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
1731 Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
1732 Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
1733 Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
1734 Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
1735 Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
1736 Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
1737 Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
1738 Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
1739 Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
1740 Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
fccf5419 1741
c1719d8b 1742 — Berlin, 2017-10-06
fccf5419 1743
4b4da299
LP
1744CHANGES WITH 234:
1745
1746 * Meson is now supported as build system in addition to Automake. It is
1747 our plan to remove Automake in one of our next releases, so that
1748 Meson becomes our exclusive build system. Hence, please start using
1749 the Meson build system in your downstream packaging. There's plenty
1750 of documentation around how to use Meson, the extremely brief
1751 summary:
1752
1753 ./autogen.sh && ./configure && make && sudo make install
1754
1755 becomes:
1756
1757 meson build && ninja -C build && sudo ninja -C build install
1758
1759 * Unit files gained support for a new JobRunningTimeoutUSec= setting,
1760 which permits configuring a timeout on the time a job is
1761 running. This is particularly useful for setting timeouts on jobs for
1762 .device units.
1763
1764 * Unit files gained two new options ConditionUser= and ConditionGroup=
1765 for conditionalizing units based on the identity of the user/group
1766 running a systemd user instance.
1767
1768 * systemd-networkd now understands a new FlowLabel= setting in the
1769 [VXLAN] section of .network files, as well as a Priority= in
1770 [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
1771 and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
1772 gained support for configuration of GENEVE links, and IPv6 address
1773 labels. The [Network] section gained the new IPv6ProxyNDP= setting.
1774
9f09a95a 1775 * .link files now understand a new Port= setting.
4b4da299
LP
1776
1777 * systemd-networkd's DHCP support gained support for DHCP option 119
1778 (domain search list).
1779
1780 * systemd-networkd gained support for serving IPv6 address ranges using
bc99dac5 1781 the Router Advertisement protocol. The new .network configuration
4b4da299
LP
1782 section [IPv6Prefix] may be used to configure the ranges to
1783 serve. This is implemented based on a new, minimal, native server
1784 implementation of RA.
1785
1786 * journalctl's --output= switch gained support for a new parameter
1787 "short-iso-precise" for a mode where timestamps are shown as precise
1788 ISO date values.
1789
1790 * systemd-udevd's "net_id" builtin may now generate stable network
1791 interface names from IBM PowerVM VIO devices as well as ACPI platform
1792 devices.
1793
1794 * MulticastDNS support in systemd-resolved may now be explicitly
1795 enabled/disabled using the new MulticastDNS= configuration file
1796 option.
1797
1798 * systemd-resolved may now optionally use libidn2 instead of the libidn
7f7ab228
ZJS
1799 for processing internationalized domain names. Support for libidn2
1800 should be considered experimental and should not be enabled by
1801 default yet.
4b4da299
LP
1802
1803 * "machinectl pull-tar" and related call may now do verification of
1804 downloaded images using SUSE-style .sha256 checksum files in addition
1805 to the already existing support for validating using Ubuntu-style
1806 SHA256SUMS files.
1807
1808 * sd-bus gained support for a new sd_bus_message_appendv() call which
1809 is va_list equivalent of sd_bus_message_append().
1810
1811 * sd-boot gained support for validating images using SHIM/MOK.
1812
1813 * The SMACK code learnt support for "onlycap".
1814
1815 * systemd-mount --umount is now much smarter in figuring out how to
1816 properly unmount a device given its mount or device path.
5486a31d
ZJS
1817
1818 * The code to call libnss_dns as a fallback from libnss_resolve when
1819 the communication with systemd-resolved fails was removed. This
1820 fallback was redundant and interfered with the [!UNAVAIL=return]
1821 suffix. See nss-resolve(8) for the recommended configuration.
1822
9f09a95a
ZJS
1823 * systemd-logind may now be restarted without losing state. It stores
1824 the file descriptors for devices it manages in the system manager
38d93385 1825 using the FDSTORE= mechanism. Please note that further changes in
9f09a95a
ZJS
1826 other components may be required to make use of this (for example
1827 Xorg has code to listen for stops of systemd-logind and terminate
1828 itself when logind is stopped or restarted, in order to avoid using
1829 stale file descriptors for graphical devices, which is now
1830 counterproductive and must be reverted in order for restarts of
1831 systemd-logind to be safe. See
1832 https://cgit.freedesktop.org/xorg/xserver/commit/?id=dc48bd653c7e101.)
1833
9d8813b3
YW
1834 * All kernel install plugins are called with the environment variable
1835 KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
1836 /etc/machine-id. If the file is missing or empty, the variable is
1837 empty and BOOT_DIR_ABS is the path of a temporary directory which is
38d93385 1838 removed after all the plugins exit. So, if KERNEL_INSTALL_MACHINE_ID
9d8813b3
YW
1839 is empty, all plugins should not put anything in BOOT_DIR_ABS.
1840
184d2c15 1841 Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
ac172e52
LP
1842 Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
1843 Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
1844 Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
1845 Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
184d2c15
LP
1846 Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
1847 Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
1848 Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
1849 Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
ac172e52
LP
1850 Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
1851 hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
1852 Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
1853 Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
1854 Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
1855 Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
1856 Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
1857 Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
1858 Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
1859 Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
1860 Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
1861 Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
1862 Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
1863 Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
1864 Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
1865 Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
184d2c15
LP
1866 H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
1867 Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
ac172e52
LP
1868 userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
1869 Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
1870 Георгиевски
4b4da299 1871
ac172e52 1872 — Berlin, 2017-07-12
4b4da299 1873
a2b53448 1874CHANGES WITH 233:
d08ee7cb 1875
23eb30b3
ZJS
1876 * The "hybrid" control group mode has been modified to improve
1877 compatibility with "legacy" cgroups-v1 setups. Specifically, the
1878 "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
1879 "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
1880 cgroups-v1 hierarchy), the only externally visible change being that
1881 the cgroups-v2 hierarchy is also mounted, to
1882 /sys/fs/cgroup/unified. This should provide a large degree of
1883 compatibility with "legacy" cgroups-v1, while taking benefit of the
1884 better management capabilities of cgroups-v2.
1885
1886 * The default control group setup mode may be selected both a boot-time
1887 via a set of kernel command line parameters (specifically:
1888 systemd.unified_cgroup_hierarchy= and
1889 systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
1890 default selected on the configure command line
1891 (--with-default-hierarchy=). The upstream default is "hybrid"
1892 (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
1893 this will change in a future systemd version to be "unified" (pure
1894 cgroups-v2 mode). The third option for the compile time option is
1895 "legacy", to enter pure cgroups-v1 mode. We recommend downstream
1896 distributions to default to "hybrid" mode for release distributions,
1897 starting with v233. We recommend "unified" for development
1898 distributions (specifically: distributions such as Fedora's rawhide)
1899 as that's where things are headed in the long run. Use "legacy" for
1900 greatest stability and compatibility only.
1901
1902 * Note one current limitation of "unified" and "hybrid" control group
1903 setup modes: the kernel currently does not permit the systemd --user
1904 instance (i.e. unprivileged code) to migrate processes between two
1905 disconnected cgroup subtrees, even if both are managed and owned by
1906 the user. This effectively means "systemd-run --user --scope" doesn't
1907 work when invoked from outside of any "systemd --user" service or
1908 scope. Specifically, it is not supported from session scopes. We are
1909 working on fixing this in a future systemd version. (See #3388 for
1910 further details about this.)
1911
fb7c4eff
MG
1912 * DBus policy files are now installed into /usr rather than /etc. Make
1913 sure your system has dbus >= 1.9.18 running before upgrading to this
1914 version, or override the install path with --with-dbuspolicydir= .
1915
23eb30b3
ZJS
1916 * All python scripts shipped with systemd (specifically: the various
1917 tests written in Python) now require Python 3.
1918
d60c5270 1919 * systemd unit tests can now run standalone (without the source or
4dfe64f8
ZJS
1920 build directories), and can be installed into /usr/lib/systemd/tests/
1921 with 'make install-tests'.
1922
23eb30b3
ZJS
1923 * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
1924 CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
1925 kernel.
1926
1927 * Support for the %c, %r, %R specifiers in unit files has been
1928 removed. Specifiers are not supposed to be dependent on configuration
1929 in the unit file itself (so that they resolve the same regardless
1930 where used in the unit files), but these specifiers were influenced
1931 by the Slice= option.
1932
5cfc0a84
LP
1933 * The shell invoked by debug-shell.service now defaults to /bin/sh in
1934 all cases. If distributions want to use a different shell for this
1935 purpose (for example Fedora's /sbin/sushell) they need to specify
1936 this explicitly at configure time using --with-debug-shell=.
1937
2bcc3309
FB
1938 * The confirmation spawn prompt has been reworked to offer the
1939 following choices:
1940
b0eb2944 1941 (c)ontinue, proceed without asking anymore
dd6f9ac0 1942 (D)ump, show the state of the unit
2bcc3309 1943 (f)ail, don't execute the command and pretend it failed
d172b175 1944 (h)elp
eedf223a 1945 (i)nfo, show a short summary of the unit
56fde33a 1946 (j)obs, show jobs that are in progress
2bcc3309
FB
1947 (s)kip, don't execute the command and pretend it succeeded
1948 (y)es, execute the command
1949
1950 The 'n' choice for the confirmation spawn prompt has been removed,
1951 because its meaning was confusing.
1952
d08ee7cb
LP
1953 The prompt may now also be redirected to an alternative console by
1954 specifying the console as parameter to systemd.confirm_spawn=.
1955
8e458bfe
JW
1956 * Services of Type=notify require a READY=1 notification to be sent
1957 during startup. If no such message is sent, the service now fails,
1958 even if the main process exited with a successful exit code.
1959
85266f9b
LP
1960 * Services that fail to start up correctly now always have their
1961 ExecStopPost= commands executed. Previously, they'd enter "failed"
1962 state directly, without executing these commands.
1963
baf32786
MP
1964 * The option MulticastDNS= of network configuration files has acquired
1965 an actual implementation. With MulticastDNS=yes a host can resolve
23eb30b3 1966 names of remote hosts and reply to mDNS A and AAAA requests.
fa8b4499 1967
631b676b
LP
1968 * When units are about to be started an additional check is now done to
1969 ensure that all dependencies of type BindsTo= (when used in
1970 combination with After=) have been started.
1971
d08ee7cb
LP
1972 * systemd-analyze gained a new verb "syscall-filter" which shows which
1973 system call groups are defined for the SystemCallFilter= unit file
23eb30b3 1974 setting, and which system calls they contain.
d08ee7cb
LP
1975
1976 * A new system call filter group "@filesystem" has been added,
23eb30b3 1977 consisting of various file system related system calls. Group
d08ee7cb 1978 "@reboot" has been added, covering reboot, kexec and shutdown related
23eb30b3 1979 calls. Finally, group "@swap" has been added covering swap
d08ee7cb
LP
1980 configuration related calls.
1981
1982 * A new unit file option RestrictNamespaces= has been added that may be
1983 used to restrict access to the various process namespace types the
1984 Linux kernel provides. Specifically, it may be used to take away the
23eb30b3
ZJS
1985 right for a service unit to create additional file system, network,
1986 user, and other namespaces. This sandboxing option is particularly
1987 relevant due to the high amount of recently discovered namespacing
1988 related vulnerabilities in the kernel.
d08ee7cb 1989
23eb30b3
ZJS
1990 * systemd-udev's .link files gained support for a new AutoNegotiation=
1991 setting for configuring Ethernet auto-negotiation.
d08ee7cb
LP
1992
1993 * systemd-networkd's .network files gained support for a new
1994 ListenPort= setting in the [DHCP] section to explicitly configure the
1995 UDP client port the DHCP client shall listen on.
1996
23eb30b3
ZJS
1997 * .network files gained a new Unmanaged= boolean setting for explicitly
1998 excluding one or more interfaces from management by systemd-networkd.
1999
2000 * The systemd-networkd ProxyARP= option has been renamed to
2001 IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
2002 renamed to ReduceARPProxy=. The old names continue to be available
2003 for compatibility.
2004
2005 * systemd-networkd gained support for configuring IPv6 Proxy NDP
2006 addresses via the new IPv6ProxyNDPAddress= .network file setting.
2007
2008 * systemd-networkd's bonding device support gained support for two new
2009 configuration options ActiveSlave= and PrimarySlave=.
2010
2011 * The various options in the [Match] section of .network files gained
2012 support for negative matching.
2013
d08ee7cb
LP
2014 * New systemd-specific mount options are now understood in /etc/fstab:
2015
2016 x-systemd.mount-timeout= may be used to configure the maximum
2017 permitted runtime of the mount command.
2018
2019 x-systemd.device-bound may be set to bind a mount point to its
2020 backing device unit, in order to automatically remove a mount point
2021 if its backing device is unplugged. This option may also be
2022 configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
2023 on the block device, which is now automatically set for all CDROM
2024 drives, so that mounted CDs are automatically unmounted when they are
2025 removed from the drive.
2026
23eb30b3
ZJS
2027 x-systemd.after= and x-systemd.before= may be used to explicitly
2028 order a mount after or before another unit or mount point.
d08ee7cb
LP
2029
2030 * Enqueued start jobs for device units are now automatically garbage
2031 collected if there are no jobs waiting for them anymore.
2032
23eb30b3
ZJS
2033 * systemctl list-jobs gained two new switches: with --after, for every
2034 queued job the jobs it's waiting for are shown; with --before the
2035 jobs which it's blocking are shown.
d08ee7cb
LP
2036
2037 * systemd-nspawn gained support for ephemeral boots from disk images
2038 (or in other words: --ephemeral and --image= may now be
2039 combined). Moreover, ephemeral boots are now supported for normal
2040 directories, even if the backing file system is not btrfs. Of course,
baf32786
MP
2041 if the file system does not support file system snapshots or
2042 reflinks, the initial copy operation will be relatively expensive, but
2043 this should still be suitable for many use cases.
d08ee7cb
LP
2044
2045 * Calendar time specifications in .timer units now support
2046 specifications relative to the end of a month by using "~" instead of
2047 "-" as separator between month and day. For example, "*-02~03" means
23eb30b3 2048 "the third last day in February". In addition a new syntax for
d08ee7cb
LP
2049 repeated events has been added using the "/" character. For example,
2050 "9..17/2:00" means "every two hours from 9am to 5pm".
2051
2052 * systemd-socket-proxyd gained a new parameter --connections-max= for
2053 configuring the maximum number of concurrent connections.
2054
23eb30b3
ZJS
2055 * sd-id128 gained a new API for generating unique IDs for the host in a
2056 way that does not leak the machine ID. Specifically,
d08ee7cb 2057 sd_id128_get_machine_app_specific() derives an ID based on the
baf32786 2058 machine ID a in well-defined, non-reversible, stable way. This is
d08ee7cb
LP
2059 useful whenever an identifier for the host is needed but where the
2060 identifier shall not be useful to identify the system beyond the
2061 scope of the application itself. (Internally this uses HMAC-SHA256 as
2062 keyed hash function using the machine ID as input.)
2063
2064 * NotifyAccess= gained a new supported value "exec". When set
2065 notifications are accepted from all processes systemd itself invoked,
2066 including all control processes.
2067
2068 * .nspawn files gained support for defining overlay mounts using the
2069 Overlay= and OverlayReadOnly= options. Previously this functionality
2070 was only available on the systemd-nspawn command line.
2071
2072 * systemd-nspawn's --bind= and --overlay= options gained support for
2073 bind/overlay mounts whose source lies within the container tree by
2074 prefixing the source path with "+".
2075
2076 * systemd-nspawn's --bind= and --overlay= options gained support for
2077 automatically allocating a temporary source directory in /var/tmp
2078 that is removed when the container dies. Specifically, if the source
2079 directory is specified as empty string this mechanism is selected. An
2080 example usage is --overlay=+/var::/var, which creates an overlay
2081 mount based on the original /var contained in the image, overlayed
2082 with a temporary directory in the host's /var/tmp. This way changes
2083 to /var are automatically flushed when the container shuts down.
2084
baf32786
MP
2085 * systemd-nspawn --image= option does now permit raw file system block
2086 devices (in addition to images containing partition tables, as
2087 before).
d08ee7cb
LP
2088
2089 * The disk image dissection logic in systemd-nspawn gained support for
2090 automatically setting up LUKS encrypted as well as Verity protected
2091 partitions. When a container is booted from an encrypted image the
2092 passphrase is queried at start-up time. When a container with Verity
2093 data is started, the root hash is search in a ".roothash" file
2094 accompanying the disk image (alternatively, pass the root hash via
2095 the new --root-hash= command line option).
2096
2097 * A new tool /usr/lib/systemd/systemd-dissect has been added that may
2098 be used to dissect disk images the same way as systemd-nspawn does
2099 it, following the Bootable Partition Specification. It may even be
2100 used to mount disk images with complex partition setups (including
2101 LUKS and Verity partitions) to a local host directory, in order to
2102 inspect them. This tool is not considered public API (yet), and is
2103 thus not installed into /usr/bin. Please do not rely on its
3b31c466 2104 existence, since it might go away or be changed in later systemd
d08ee7cb
LP
2105 versions.
2106
2107 * A new generator "systemd-verity-generator" has been added, similar in
baf32786 2108 style to "systemd-cryptsetup-generator", permitting automatic setup of
d08ee7cb
LP
2109 Verity root partitions when systemd boots up. In order to make use of
2110 this your partition setup should follow the Discoverable Partitions
2111 Specification, and the GPT partition ID of the root file system
2112 partition should be identical to the upper 128bit of the Verity root
2113 hash. The GPT partition ID of the Verity partition protecting it
2114 should be the lower 128bit of the Verity root hash. If the partition
2115 image follows this model it is sufficient to specify a single
2116 "roothash=" kernel command line argument to both configure which root
2117 image and verity partition to use as well as the root hash for
2118 it. Note that systemd-nspawn's Verity support follows the same
2119 semantics, meaning that disk images with proper Verity data in place
2120 may be booted in containers with systemd-nspawn as well as on
2121 physical systems via the verity generator. Also note that the "mkosi"
2122 tool available at https://github.com/systemd/mkosi has been updated
2123 to generate Verity protected disk images following this scheme. In
2124 fact, it has been updated to generate disk images that optionally
2125 implement a complete UEFI SecureBoot trust chain, involving a signed
2126 kernel and initrd image that incorporates such a root hash as well as
2127 a Verity-enabled root partition.
2128
d08ee7cb
LP
2129 * The hardware database (hwdb) udev supports has been updated to carry
2130 accelerometer quirks.
2131
2132 * All system services are now run with a fresh kernel keyring set up
2133 for them. The invocation ID is stored by default in it, thus
2134 providing a safe, non-overridable way to determine the invocation
2135 ID of each service.
2136
2137 * Service unit files gained new BindPaths= and BindReadOnlyPaths=
2138 options for bind mounting arbitrary paths in a service-specific
2139 way. When these options are used, arbitrary host or service files and
2140 directories may be mounted to arbitrary locations in the service's
2141 view.
2142
2143 * Documentation has been added that lists all of systemd's low-level
2144 environment variables:
2145
a8a27374 2146 https://github.com/systemd/systemd/blob/master/docs/ENVIRONMENT.md
d08ee7cb
LP
2147
2148 * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
2149 whether a specific socket file descriptor matches a specified socket
2150 address.
2151
2152 * systemd-firstboot has been updated to check for the
2153 systemd.firstboot= kernel command line option. It accepts a boolean
2154 and when set to false the first boot questions are skipped.
2155
d08ee7cb 2156 * systemd-fstab-generator has been updated to check for the
23eb30b3
ZJS
2157 systemd.volatile= kernel command line option, which either takes an
2158 optional boolean parameter or the special value "state". If used the
2159 system may be booted in a "volatile" boot mode. Specifically,
2160 "systemd.volatile" is used, the root directory will be mounted as
d08ee7cb 2161 tmpfs, and only /usr is mounted from the actual root file system. If
23eb30b3
ZJS
2162 "systemd.volatile=state" is used, the root directory will be mounted
2163 as usual, but /var is mounted as tmpfs. This concept provides similar
d08ee7cb
LP
2164 functionality as systemd-nspawn's --volatile= option, but provides it
2165 on physical boots. Use this option for implementing stateless
2166 systems, or testing systems with all state and/or configuration reset
2167 to the defaults. (Note though that many distributions are not
23eb30b3 2168 prepared to boot up without a populated /etc or /var, though.)
d08ee7cb
LP
2169
2170 * systemd-gpt-auto-generator gained support for LUKS encrypted root
2171 partitions. Previously it only supported LUKS encrypted partitions
2172 for all other uses, except for the root partition itself.
2173
2174 * Socket units gained support for listening on AF_VSOCK sockets for
2175 communication in virtualized QEMU environments.
2176
2177 * The "configure" script gained a new option --with-fallback-hostname=
2178 for specifying the fallback hostname to use if none is configured in
2179 /etc/hostname. For example, by specifying
2180 --with-fallback-hostname=fedora it is possible to default to a
23eb30b3 2181 hostname of "fedora" on pristine installations.
d08ee7cb
LP
2182
2183 * systemd-cgls gained support for a new --unit= switch for listing only
2184 the control groups of a specific unit. Similar --user-unit= has been
2185 added for listing only the control groups of a specific user unit.
2186
2187 * systemd-mount gained a new --umount switch for unmounting a mount or
2188 automount point (and all mount/automount points below it).
2189
2190 * systemd will now refuse full configuration reloads (via systemctl
2191 daemon-reload and related calls) unless at least 16MiB of free space
2192 are available in /run. This is a safety precaution in order to ensure
2193 that generators can safely operate after the reload completed.
2194
2195 * A new unit file option RootImage= has been added, which has a similar
2196 effect as RootDirectory= but mounts the service's root directory from
2197 a disk image instead of plain directory. This logic reuses the same
2198 image dissection and mount logic that systemd-nspawn already uses,
2199 and hence supports any disk images systemd-nspawn supports, including
2200 those following the Discoverable Partition Specification, as well as
2201 Verity enabled images. This option enables systemd to run system
2202 services directly off disk images acting as resource bundles,
2203 possibly even including full integrity data.
2204
2205 * A new MountAPIVFS= unit file option has been added, taking a boolean
baf32786 2206 argument. If enabled /proc, /sys and /dev (collectively called the
d08ee7cb
LP
2207 "API VFS") will be mounted for the service. This is only relevant if
2208 RootDirectory= or RootImage= is used for the service, as these mounts
2209 are of course in place in the host mount namespace anyway.
2210
2211 * systemd-nspawn gained support for a new --pivot-root= switch. If
2212 specified the root directory within the container image is pivoted to
2213 the specified mount point, while the original root disk is moved to a
2214 different place. This option enables booting of ostree images
2215 directly with systemd-nspawn.
2216
d08ee7cb 2217 * The systemd build scripts will no longer complain if the NTP server
23eb30b3 2218 addresses are not changed from the defaults. Google now supports
d08ee7cb
LP
2219 these NTP servers officially. We still recommend downstreams to
2220 properly register an NTP pool with the NTP pool project though.
2221
c1ec34d1 2222 * coredumpctl gained a new "--reverse" option for printing the list
d08ee7cb
LP
2223 of coredumps in reverse order.
2224
23eb30b3
ZJS
2225 * coredumpctl will now show additional information about truncated and
2226 inaccessible coredumps, as well as coredumps that are still being
2227 processed. It also gained a new --quiet switch for suppressing
2228 additional informational message in its output.
2229
2230 * coredumpctl gained support for only showing coredumps newer and/or
2231 older than specific timestamps, using the new --since= and --until=
2232 options, reminiscent of journalctl's options by the same name.
2233
d08ee7cb 2234 * The systemd-coredump logic has been improved so that it may be reused
23eb30b3 2235 to collect backtraces in non-compiled languages, for example in
d08ee7cb
LP
2236 scripting languages such as Python.
2237
2238 * machinectl will now show the UID shift of local containers, if user
2239 namespacing is enabled for them.
2240
baf32786 2241 * systemd will now optionally run "environment generator" binaries at
d08ee7cb
LP
2242 configuration load time. They may be used to add environment
2243 variables to the environment block passed to services invoked. One
baf32786 2244 user environment generator is shipped by default that sets up
23eb30b3
ZJS
2245 environment variables based on files dropped into /etc/environment.d
2246 and ~/.config/environment.d/.
d08ee7cb 2247
a2b53448
LP
2248 * systemd-resolved now includes the new, recently published 2017 DNSSEC
2249 root key (KSK).
2250
a2b53448
LP
2251 * hostnamed has been updated to report a new chassis type of
2252 "convertible" to cover "foldable" laptops that can both act as a
2253 tablet and as a laptop, such as various Lenovo Yoga devices.
2254
d08ee7cb
LP
2255 Contributions from: Adrián López, Alexander Galanin, Alexander
2256 Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
2257 Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
2258 Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
2259 Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
2260 David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
2261 Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
2262 Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
2263 Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
a2b53448
LP
2264 Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
2265 Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
2266 Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
2267 Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
2268 Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
2269 Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
2270 Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
2271 Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
2272 Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
2273 Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
2274 Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
2275 Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
2276 Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
2277 Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
2278 Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
2279 Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
2280 Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
2281 Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
2282 YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
2283 Тихонов
2284
2285 — Berlin, 2017-03-01
d08ee7cb 2286
54b24597 2287CHANGES WITH 232:
76153ad4 2288
05f426d2
LP
2289 * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
2290 RestrictAddressFamilies= enabled. These sandboxing options should
2291 generally be compatible with the various external udev call-out
2292 binaries we are aware of, however there may be exceptions, in
2293 particular when exotic languages for these call-outs are used. In
2294 this case, consider turning off these settings locally.
2295
4ffe2479
ZJS
2296 * The new RemoveIPC= option can be used to remove IPC objects owned by
2297 the user or group of a service when that service exits.
2298
6fa44114 2299 * The new ProtectKernelModules= option can be used to disable explicit
4c37970d
LP
2300 load and unload operations of kernel modules by a service. In
2301 addition access to /usr/lib/modules is removed if this option is set.
6fa44114 2302
4a77c53d
ZJS
2303 * ProtectSystem= option gained a new value "strict", which causes the
2304 whole file system tree with the exception of /dev, /proc, and /sys,
2305 to be remounted read-only for a service.
2306
e49e2c25 2307 * The new ProtectKernelTunables= option can be used to disable
4a77c53d
ZJS
2308 modification of configuration files in /sys and /proc by a service.
2309 Various directories and files are remounted read-only, so access is
2310 restricted even if the file permissions would allow it.
2311
6fa44114 2312 * The new ProtectControlGroups= option can be used to disable write
4a77c53d
ZJS
2313 access by a service to /sys/fs/cgroup.
2314
2315 * Various systemd services have been hardened with
2316 ProtectKernelTunables=yes, ProtectControlGroups=yes,
2317 RestrictAddressFamilies=.
2318
4ffe2479
ZJS
2319 * Support for dynamically creating users for the lifetime of a service
2320 has been added. If DynamicUser=yes is specified, user and group IDs
2321 will be allocated from the range 61184..65519 for the lifetime of the
2322 service. They can be resolved using the new nss-systemd.so NSS
2323 module. The module must be enabled in /etc/nsswitch.conf. Services
2324 started in this way have PrivateTmp= and RemoveIPC= enabled, so that
2325 any resources allocated by the service will be cleaned up when the
4a77c53d
ZJS
2326 service exits. They also have ProtectHome=read-only and
2327 ProtectSystem=strict enabled, so they are not able to make any
2328 permanent modifications to the system.
4ffe2479 2329
171ae2cd 2330 * The nss-systemd module also always resolves root and nobody, making
4ffe2479 2331 it possible to have no /etc/passwd or /etc/group files in minimal
171ae2cd 2332 container or chroot environments.
4ffe2479
ZJS
2333
2334 * Services may be started with their own user namespace using the new
171ae2cd
LP
2335 boolean PrivateUsers= option. Only root, nobody, and the uid/gid
2336 under which the service is running are mapped. All other users are
2337 mapped to nobody.
4ffe2479
ZJS
2338
2339 * Support for the cgroup namespace has been added to systemd-nspawn. If
2340 supported by kernel, the container system started by systemd-nspawn
2341 will have its own view of the cgroup hierarchy. This new behaviour
2342 can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
2343
2344 * The new MemorySwapMax= option can be used to limit the maximum swap
2345 usage under the unified cgroup hierarchy.
2346
2347 * Support for the CPU controller in the unified cgroup hierarchy has
2348 been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
2349 options. This controller requires out-of-tree patches for the kernel
2350 and the support is provisional.
2351
171ae2cd
LP
2352 * Mount and automount units may now be created transiently
2353 (i.e. dynamically at runtime via the bus API, instead of requiring
2354 unit files in the file system).
2355
2356 * systemd-mount is a new tool which may mount file systems – much like
2357 mount(8), optionally pulling in additional dependencies through
2358 transient .mount and .automount units. For example, this tool
2359 automatically runs fsck on a backing block device before mounting,
2360 and allows the automount logic to be used dynamically from the
2361 command line for establishing mount points. This tool is particularly
2362 useful when dealing with removable media, as it will ensure fsck is
2363 run – if necessary – before the first access and that the file system
2364 is quickly unmounted after each access by utilizing the automount
2365 logic. This maximizes the chance that the file system on the
2366 removable media stays in a clean state, and if it isn't in a clean
2367 state is fixed automatically.
4ffe2479
ZJS
2368
2369 * LazyUnmount=yes option for mount units has been added to expose the
2370 umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
2371 option.
2372
2373 * /efi will be used as the mount point of the EFI boot partition, if
2374 the directory is present, and the mount point was not configured
2375 through other means (e.g. fstab). If /efi directory does not exist,
2376 /boot will be used as before. This makes it easier to automatically
2377 mount the EFI partition on systems where /boot is used for something
2378 else.
2379
171ae2cd
LP
2380 * When operating on GPT disk images for containers, systemd-nspawn will
2381 now mount the ESP to /boot or /efi according to the same rules as PID
2382 1 running on a host. This allows tools like "bootctl" to operate
2383 correctly within such containers, in order to make container images
2384 bootable on physical systems.
2385
4a77c53d 2386 * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
4ffe2479
ZJS
2387
2388 * Two new user session targets have been added to support running
2389 graphical sessions under the systemd --user instance:
2390 graphical-session.target and graphical-session-pre.target. See
2391 systemd.special(7) for a description of how those targets should be
2392 used.
2393
2394 * The vconsole initialization code has been significantly reworked to
d4c08299 2395 use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
4ffe2479
ZJS
2396 support unicode keymaps. Font and keymap configuration will now be
2397 copied to all allocated virtual consoles.
2398
05ecf467 2399 * FreeBSD's bhyve virtualization is now detected.
4ffe2479 2400
d4c08299 2401 * Information recorded in the journal for core dumps now includes the
4ffe2479
ZJS
2402 contents of /proc/mountinfo and the command line of the process at
2403 the top of the process hierarchy (which is usually the init process
2404 of the container).
2405
171ae2cd 2406 * systemd-journal-gatewayd learned the --directory= option to serve
4ffe2479
ZJS
2407 files from the specified location.
2408
2409 * journalctl --root=… can be used to peruse the journal in the
2410 /var/log/ directories inside of a container tree. This is similar to
2411 the existing --machine= option, but does not require the container to
2412 be active.
2413
2414 * The hardware database has been extended to support
2415 ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
2416 trackball devices.
2417
2418 MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
2419 specify the click rate for mice which include a horizontal wheel with
2420 a click rate that is different than the one for the vertical wheel.
2421
2422 * systemd-run gained a new --wait option that makes service execution
171ae2cd
LP
2423 synchronous. (Specifically, the command will not return until the
2424 specified service binary exited.)
4ffe2479 2425
171ae2cd 2426 * systemctl gained a new --wait option that causes the start command to
4a77c53d
ZJS
2427 wait until the units being started have terminated again.
2428
171ae2cd 2429 * A new journal output mode "short-full" has been added which displays
4ffe2479 2430 timestamps with abbreviated English day names and adds a timezone
171ae2cd
LP
2431 suffix. Those timestamps include more information than the default
2432 "short" output mode, and can be passed directly to journalctl's
2433 --since= and --until= options.
4ffe2479
ZJS
2434
2435 * /etc/resolv.conf will be bind-mounted into containers started by
2436 systemd-nspawn, if possible, so any changes to resolv.conf contents
2437 are automatically propagated to the container.
2438
2439 * The number of instances for socket-activated services originating
171ae2cd
LP
2440 from a single IP address can be limited with
2441 MaxConnectionsPerSource=, extending the existing setting of
2442 MaxConnections=.
4ffe2479 2443
4a77c53d
ZJS
2444 * systemd-networkd gained support for vcan ("Virtual CAN") interface
2445 configuration.
2446
2447 * .netdev and .network configuration can now be extended through
2448 drop-ins.
2449
4ffe2479
ZJS
2450 * UDP Segmentation Offload, TCP Segmentation Offload, Generic
2451 Segmentation Offload, Generic Receive Offload, Large Receive Offload
2452 can be enabled and disabled using the new UDPSegmentationOffload=,
2453 TCPSegmentationOffload=, GenericSegmentationOffload=,
2454 GenericReceiveOffload=, LargeReceiveOffload= options in the
2455 [Link] section of .link files.
2456
171ae2cd
LP
2457 * The Spanning Tree Protocol, Priority, Aging Time, and the Default
2458 Port VLAN ID can be configured for bridge devices using the new STP=,
2459 Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
2460 section of .netdev files.
4ffe2479 2461
171ae2cd 2462 * The route table to which routes received over DHCP or RA should be
4a77c53d
ZJS
2463 added can be configured with the new RouteTable= option in the [DHCP]
2464 and [IPv6AcceptRA] sections of .network files.
2465
171ae2cd 2466 * The Address Resolution Protocol can be disabled on links managed by
4ffe2479
ZJS
2467 systemd-networkd using the ARP=no setting in the [Link] section of
2468 .network files.
2469
171ae2cd
LP
2470 * New environment variables $SERVICE_RESULT, $EXIT_CODE and
2471 $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
2472 encode information about the result and exit codes of the current
2473 service runtime cycle.
4ffe2479 2474
4a77c53d 2475 * systemd-sysctl will now configure kernel parameters in the order
1f4f4cf7 2476 they occur in the configuration files. This matches what sysctl
4a77c53d
ZJS
2477 has been traditionally doing.
2478
2479 * kernel-install "plugins" that are executed to perform various
2480 tasks after a new kernel is added and before an old one is removed
2481 can now return a special value to terminate the procedure and
2482 prevent any later plugins from running.
2483
76153ad4 2484 * Journald's SplitMode=login setting has been deprecated. It has been
d4c08299 2485 removed from documentation, and its use is discouraged. In a future
76153ad4
ZJS
2486 release it will be completely removed, and made equivalent to current
2487 default of SplitMode=uid.
2488
4a77c53d
ZJS
2489 * Storage=both option setting in /etc/systemd/coredump.conf has been
2490 removed. With fast LZ4 compression storing the core dump twice is not
2491 useful.
2492
4ffe2479
ZJS
2493 * The --share-system systemd-nspawn option has been replaced with an
2494 (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
2495 this functionality is discouraged. In addition the variables
2496 $SYSTEMD_NSPAWN_SHARE_NS_IPC, $SYSTEMD_NSPAWN_SHARE_NS_PID,
2497 $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
2498 individual namespaces.
2499
171ae2cd
LP
2500 * "machinectl list" now shows the IP address of running containers in
2501 the output, as well as OS release information.
2502
2503 * "loginctl list" now shows the TTY of each session in the output.
2504
2505 * sd-bus gained new API calls sd_bus_track_set_recursive(),
2506 sd_bus_track_get_recursive(), sd_bus_track_count_name(),
2507 sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
2508 tracking objects in a "recursive" mode, where a single client can be
2509 counted multiple times, if it takes multiple references.
2510
2511 * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
bc99dac5 2512 sd_bus_get_exit_on_disconnect(). They may be used to make a
171ae2cd
LP
2513 process using sd-bus automatically exit if the bus connection is
2514 severed.
2515
2516 * Bus clients of the service manager may now "pin" loaded units into
2517 memory, by taking an explicit reference on them. This is useful to
2518 ensure the client can retrieve runtime data about the service even
2519 after the service completed execution. Taking such a reference is
2520 available only for privileged clients and should be helpful to watch
2521 running services in a race-free manner, and in particular collect
2522 information about exit statuses and results.
2523
4c37970d
LP
2524 * The nss-resolve module has been changed to strictly return UNAVAIL
2525 when communication via D-Bus with resolved failed, and NOTFOUND when
2526 a lookup completed but was negative. This means it is now possible to
2527 neatly configure fallbacks using nsswitch.conf result checking
2528 expressions. Taking benefit of this, the new recommended
2529 configuration line for the "hosts" entry in /etc/nsswitch.conf is:
2530
2531 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
2532
2533 * A new setting CtrlAltDelBurstAction= has been added to
2534 /etc/systemd/system.conf which may be used to configure the precise
2535 behaviour if the user on the console presses Ctrl-Alt-Del more often
2536 than 7 times in 2s. Previously this would unconditionally result in
2537 an expedited, immediate reboot. With this new setting the precise
2538 operation may be configured in more detail, and also turned off
2539 entirely.
2540
2541 * In .netdev files two new settings RemoteChecksumTx= and
2542 RemoteChecksumRx= are now understood that permit configuring the
2543 remote checksumming logic for VXLAN networks.
2544
2545 * The service manager learnt a new "invocation ID" concept for invoked
2546 services. Each runtime cycle of a service will get a new invocation
2547 ID (a 128bit random UUID) assigned that identifies the current
2548 run of the service uniquely and globally. A new invocation ID
2549 is generated each time a service starts up. The journal will store
2550 the invocation ID of a service along with any logged messages, thus
2551 making the invocation ID useful for matching the online runtime of a
2552 service with the offline log data it generated in a safe way without
2553 relying on synchronized timestamps. In many ways this new service
2554 invocation ID concept is similar to the kernel's boot ID concept that
2555 uniquely and globally identifies the runtime of each boot. The
2556 invocation ID of a service is passed to the service itself via an
2557 environment variable ($INVOCATION_ID). A new bus call
2558 GetUnitByInvocationID() has been added that is similar to GetUnit()
2559 but instead of retrieving the bus path for a unit by its name
2560 retrieves it by its invocation ID. The returned path is valid only as
2561 long as the passed invocation ID is current.
2562
2563 * systemd-resolved gained a new "DNSStubListener" setting in
2564 resolved.conf. It either takes a boolean value or the special values
2565 "udp" and "tcp", and configures whether to enable the stub DNS
2566 listener on 127.0.0.53:53.
2567
2568 * IP addresses configured via networkd may now carry additional
2569 configuration settings supported by the kernel. New options include:
2570 HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
2571 PrefixRoute=, AutoJoin=.
2572
2573 * The PAM configuration fragment file for "user@.service" shipped with
2574 systemd (i.e. the --user instance of systemd) has been stripped to
2575 the minimum necessary to make the system boot. Previously, it
2576 contained Fedora-specific stanzas that did not apply to other
2577 distributions. It is expected that downstream distributions add
2578 additional configuration lines, matching their needs to this file,
2579 using it only as rough template of what systemd itself needs. Note
2580 that this reduced fragment does not even include an invocation of
2581 pam_limits which most distributions probably want to add, even though
2582 systemd itself does not need it. (There's also the new build time
2583 option --with-pamconfdir=no to disable installation of the PAM
2584 fragment entirely.)
2585
2586 * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
2587 capability is now also dropped from its set (in addition to
2588 CAP_SYS_MKNOD as before).
2589
2590 * In service unit files it is now possible to connect a specific named
2591 file descriptor with stdin/stdout/stdout of an executed service. The
2592 name may be specified in matching .socket units using the
2593 FileDescriptorName= setting.
2594
2595 * A number of journal settings may now be configured on the kernel
2596 command line. Specifically, the following options are now understood:
2597 systemd.journald.max_level_console=,
2598 systemd.journald.max_level_store=,
2599 systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
2600 systemd.journald.max_level_wall=.
2601
2602 * "systemctl is-enabled --full" will now show by which symlinks a unit
2603 file is enabled in the unit dependency tree.
2604
b4eed568
LP
2605 * Support for VeraCrypt encrypted partitions has been added to the
2606 "cryptsetup" logic and /etc/crypttab.
2607
2608 * systemd-detect-virt gained support for a new --private-users switch
2609 that checks whether the invoking processes are running inside a user
2610 namespace. Similar, a new special value "private-users" for the
2611 existing ConditionVirtualization= setting has been added, permitting
2612 skipping of specific units in user namespace environments.
2613
07393b6e
LP
2614 Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
2615 Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
2616 Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
2617 Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
2618 Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
2619 Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
2620 Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
2621 Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
2622 Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
2623 Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
2624 Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
2625 Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
2626 Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
2627 Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
2628 Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
2629 Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
2630 Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
2631 Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
2632 Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
2633 Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
2634 Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
2635 Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
2636 Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
2637 Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
2638 E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
2639 Jędrzejewski-Szmek, Zeal Jagannatha
2640
54b24597 2641 — Santa Fe, 2016-11-03
07393b6e 2642
5cd118ba
MP
2643CHANGES WITH 231:
2644
fcd30826
LP
2645 * In service units the various ExecXYZ= settings have been extended
2646 with an additional special character as first argument of the
43eb109a 2647 assigned value: if the character '+' is used the specified command
fcd30826
LP
2648 line it will be run with full privileges, regardless of User=,
2649 Group=, CapabilityBoundingSet= and similar options. The effect is
2650 similar to the existing PermissionsStartOnly= option, but allows
2651 configuration of this concept for each executed command line
2652 independently.
2653
2654 * Services may now alter the service watchdog timeout at runtime by
2655 sending a WATCHDOG_USEC= message via sd_notify().
2656
2657 * MemoryLimit= and related unit settings now optionally take percentage
2658 specifications. The percentage is taken relative to the amount of
2659 physical memory in the system (or in case of containers, the assigned
2660 amount of memory). This allows scaling service resources neatly with
771de3f5 2661 the amount of RAM available on the system. Similarly, systemd-logind's
fcd30826
LP
2662 RuntimeDirectorySize= option now also optionally takes percentage
2663 values.
2664
2665 * In similar fashion TasksMax= takes percentage values now, too. The
2666 value is taken relative to the configured maximum number of processes
2667 on the system. The per-service task maximum has been changed to 15%
2668 using this functionality. (Effectively this is an increase of 512 →
2669 4915 for service units, given the kernel's default pid_max setting.)
2670
2671 * Calendar time specifications in .timer units now understand a ".."
2672 syntax for time ranges. Example: "4..7:10" may now be used for
2673 defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
2674 7:10am every day.
2675
2676 * The InaccessableDirectories=, ReadOnlyDirectories= and
2677 ReadWriteDirectories= unit file settings have been renamed to
2678 InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
2679 applied to all kinds of file nodes, and not just directories, with
2680 the exception of symlinks. Specifically these settings may now be
2681 used on block and character device nodes, UNIX sockets and FIFOS as
2682 well as regular files. The old names of these settings remain
2683 available for compatibility.
2684
2685 * systemd will now log about all service processes it kills forcibly
2686 (using SIGKILL) because they remained after the clean shutdown phase
2687 of the service completed. This should help identifying services that
2688 shut down uncleanly. Moreover if KillUserProcesses= is enabled in
2689 systemd-logind's configuration a similar log message is generated for
2690 processes killed at the end of each session due to this setting.
2691
2692 * systemd will now set the $JOURNAL_STREAM environment variable for all
2693 services whose stdout/stderr are connected to the Journal (which
2694 effectively means by default: all services). The variable contains
2695 the device and inode number of the file descriptor used for
2696 stdout/stderr. This may be used by invoked programs to detect whether
2697 their stdout/stderr is connected to the Journal, in which case they
2698 can switch over to direct Journal communication, thus being able to
2699 pass extended, structured metadata along with their log messages. As
2700 one example, this is now used by glib's logging primitives.
2701
2702 * When using systemd's default tmp.mount unit for /tmp, the mount point
2703 will now be established with the "nosuid" and "nodev" options. This
2704 avoids privilege escalation attacks that put traps and exploits into
2705 /tmp. However, this might cause problems if you e. g. put container
5cd118ba
MP
2706 images or overlays into /tmp; if you need this, override tmp.mount's
2707 "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
2708 desired options.
2709
fcd30826 2710 * systemd now supports the "memory" cgroup controller also on
4e1dfa45 2711 cgroup v2.
fcd30826
LP
2712
2713 * The systemd-cgtop tool now optionally takes a control group path as
2714 command line argument. If specified, the control group list shown is
2715 limited to subgroups of that group.
2716
2717 * The SystemCallFilter= unit file setting gained support for
2718 pre-defined, named system call filter sets. For example
2719 SystemCallFilter=@clock is now an effective way to make all clock
771de3f5 2720 changing-related system calls unavailable to a service. A number of
fcd30826
LP
2721 similar pre-defined groups are defined. Writing system call filters
2722 for system services is simplified substantially with this new
2723 concept. Accordingly, all of systemd's own, long-running services now
2724 enable system call filtering based on this, by default.
2725
2726 * A new service setting MemoryDenyWriteExecute= has been added, taking
2727 a boolean value. If turned on, a service may no longer create memory
2728 mappings that are writable and executable at the same time. This
2729 enhances security for services where this is enabled as it becomes
2730 harder to dynamically write and then execute memory in exploited
2731 service processes. This option has been enabled for all of systemd's
2732 own long-running services.
2733
2734 * A new RestrictRealtime= service setting has been added, taking a
2735 boolean argument. If set the service's processes may no longer
2736 acquire realtime scheduling. This improves security as realtime
2737 scheduling may otherwise be used to easily freeze the system.
2738
2739 * systemd-nspawn gained a new switch --notify-ready= taking a boolean
2740 value. This may be used for requesting that the system manager inside
2741 of the container reports start-up completion to nspawn which then
2742 propagates this notification further to the service manager
2743 supervising nspawn itself. A related option NotifyReady= in .nspawn
2744 files has been added too. This functionality allows ordering of the
2745 start-up of multiple containers using the usual systemd ordering
2746 primitives.
2747
2748 * machinectl gained a new command "stop" that is an alias for
2749 "terminate".
2750
2751 * systemd-resolved gained support for contacting DNS servers on
2752 link-local IPv6 addresses.
2753
2754 * If systemd-resolved receives the SIGUSR2 signal it will now flush all
2755 its caches. A method call for requesting the same operation has been
2756 added to the bus API too, and is made available via "systemd-resolve
2757 --flush-caches".
2758
771de3f5 2759 * systemd-resolve gained a new --status switch. If passed a brief
fcd30826
LP
2760 summary of the used DNS configuration with per-interface information
2761 is shown.
2762
2763 * resolved.conf gained a new Cache= boolean option, defaulting to
2764 on. If turned off local DNS caching is disabled. This comes with a
2765 performance penalty in particular when DNSSEC is enabled. Note that
771de3f5 2766 resolved disables its internal caching implicitly anyway, when the
fcd30826
LP
2767 configured DNS server is on a host-local IP address such as ::1 or
2768 127.0.0.1, thus automatically avoiding double local caching.
2769
2770 * systemd-resolved now listens on the local IP address 127.0.0.53:53
2771 for DNS requests. This improves compatibility with local programs
2772 that do not use the libc NSS or systemd-resolved's bus APIs for name
2773 resolution. This minimal DNS service is only available to local
2774 programs and does not implement the full DNS protocol, but enough to
2775 cover local DNS clients. A new, static resolv.conf file, listing just
2776 this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
2777 now recommended to make /etc/resolv.conf a symlink to this file in
2778 order to route all DNS lookups to systemd-resolved, regardless if
2779 done via NSS, the bus API or raw DNS packets. Note that this local
2780 DNS service is not as fully featured as the libc NSS or
2781 systemd-resolved's bus APIs. For example, as unicast DNS cannot be
2782 used to deliver link-local address information (as this implies
2783 sending a local interface index along), LLMNR/mDNS support via this
2784 interface is severely restricted. It is thus strongly recommended for
2785 all applications to use the libc NSS API or native systemd-resolved
2786 bus API instead.
2787
2788 * systemd-networkd's bridge support learned a new setting
2789 VLANFiltering= for controlling VLAN filtering. Moreover a new section
2790 in .network files has been added for configuring VLAN bridging in
2791 more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
2792
2793 * systemd-networkd's IPv6 Router Advertisement code now makes use of
2794 the DNSSL and RDNSS options. This means IPv6 DNS configuration may
2795 now be acquired without relying on DHCPv6. Two new options
2796 UseDomains= and UseDNS= have been added to configure this behaviour.
2797
2798 * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
2799 renamed IPv6AcceptRA=, without altering its behaviour. The old
2800 setting name remains available for compatibility reasons.
2801
2802 * The systemd-networkd VTI/VTI6 tunneling support gained new options
2803 Key=, InputKey= and OutputKey=.
2804
2805 * systemd-networkd gained support for VRF ("Virtual Routing Function")
2806 interface configuration.
2807
2808 * "systemctl edit" may now be used to create new unit files by
2809 specifying the --force switch.
2810
2811 * sd-event gained a new function sd_event_get_iteration() for
2812 requesting the current iteration counter of the event loop. It starts
2813 at zero and is increased by one with each event loop iteration.
2814
43a569a1
ZJS
2815 * A new rpm macro %systemd_ordering is provided by the macros.systemd
2816 file. It can be used in lieu of %systemd_requires in packages which
2817 don't use any systemd functionality and are intended to be installed
2818 in minimal containers without systemd present. This macro provides
ce830873 2819 ordering dependencies to ensure that if the package is installed in
43a569a1
ZJS
2820 the same rpm transaction as systemd, systemd will be installed before
2821 the scriptlets for the package are executed, allowing unit presets
2822 to be handled.
2823
2824 New macros %_systemdgeneratordir and %_systemdusergeneratordir have
2825 been added to simplify packaging of generators.
2826
2827 * The os-release file gained VERSION_CODENAME field for the
2828 distribution nickname (e.g. VERSION_CODENAME=woody).
2829
2830 * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1
2831 can be set to disable parsing of metadata and the creation
2832 of persistent symlinks for that device.
2833
0f1da52b
LP
2834 * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
2835 to make them available to logged-in users has been reverted.
2836
2837 * Much of the common code of the various systemd components is now
2838 built into an internal shared library libsystemd-shared-231.so
2839 (incorporating the systemd version number in the name, to be updated
2840 with future releases) that the components link to. This should
2841 decrease systemd footprint both in memory during runtime and on
2842 disk. Note that the shared library is not for public use, and is
ead6bd25 2843 neither API nor ABI stable, but is likely to change with every new
1ecbf32f
ZJS
2844 released update. Packagers need to make sure that binaries
2845 linking to libsystemd-shared.so are updated in step with the
2846 library.
43a569a1 2847
fcd30826
LP
2848 * Configuration for "mkosi" is now part of the systemd
2849 repository. mkosi is a tool to easily build legacy-free OS images,
2850 and is available on github: https://github.com/systemd/mkosi. If
2851 "mkosi" is invoked in the build tree a new raw OS image is generated
2852 incorporating the systemd sources currently being worked on and a
2853 clean, fresh distribution installation. The generated OS image may be
ce830873 2854 booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
fcd30826
LP
2855 UEFI PC. This functionality is particularly useful to easily test
2856 local changes made to systemd in a pristine, defined environment. See
f09eb768 2857 doc/HACKING for details.
ceeddf79 2858
4ffe2479
ZJS
2859 * configure learned the --with-support-url= option to specify the
2860 distribution's bugtracker.
2861
38b383d9
LP
2862 Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
2863 Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
2864 Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
2865 Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
2866 Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
2867 Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
2868 Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
2869 Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
2870 Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
2871 Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
2872 Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
2873 Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
2874 Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
2875 Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
2876 Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
2877 Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
771de3f5
ZJS
2878 Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
2879 Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
38b383d9 2880 WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
5cd118ba 2881
38b383d9 2882 — Berlin, 2016-07-25
5cd118ba 2883
46e40fab 2884CHANGES WITH 230:
7f6e8043 2885
61ecb465
LP
2886 * DNSSEC is now turned on by default in systemd-resolved (in
2887 "allow-downgrade" mode), but may be turned off during compile time by
2888 passing "--with-default-dnssec=no" to "configure" (and of course,
2889 during runtime with DNSSEC= in resolved.conf). We recommend
2890 downstreams to leave this on at least during development cycles and
2891 report any issues with the DNSSEC logic upstream. We are very
2892 interested in collecting feedback about the DNSSEC validator and its
2893 limitations in the wild. Note however, that DNSSEC support is
2894 probably nothing downstreams should turn on in stable distros just
96d49011 2895 yet, as it might create incompatibilities with a few DNS servers and
61ecb465
LP
2896 networks. We tried hard to make sure we downgrade to non-DNSSEC mode
2897 automatically whenever we detect such incompatible setups, but there
2898 might be systems we do not cover yet. Hence: please help us testing
2899 the DNSSEC code, leave this on where you can, report back, but then
2900 again don't consider turning this on in your stable, LTS or
e40a326c
LP
2901 production release just yet. (Note that you have to enable
2902 nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
2903 and its DNSSEC mode for host name resolution from local
2904 applications.)
61ecb465 2905
96515dbf 2906 * systemd-resolve conveniently resolves DANE records with the --tlsa
e40a326c 2907 option and OPENPGPKEY records with the --openpgp option. It also
e75690c3 2908 supports dumping raw DNS record data via the new --raw= switch.
96515dbf 2909
97e5530c
ZJS
2910 * systemd-logind will now by default terminate user processes that are
2911 part of the user session scope unit (session-XX.scope) when the user
977f2bea 2912 logs out. This behavior is controlled by the KillUserProcesses=
e40a326c
LP
2913 setting in logind.conf, and the previous default of "no" is now
2914 changed to "yes". This means that user sessions will be properly
2915 cleaned up after, but additional steps are necessary to allow
2916 intentionally long-running processes to survive logout.
97e5530c
ZJS
2917
2918 While the user is logged in at least once, user@.service is running,
2919 and any service that should survive the end of any individual login
2920 session can be started at a user service or scope using systemd-run.
e40a326c 2921 systemd-run(1) man page has been extended with an example which shows
8951eaec 2922 how to run screen in a scope unit underneath user@.service. The same
e40a326c 2923 command works for tmux.
97e5530c
ZJS
2924
2925 After the user logs out of all sessions, user@.service will be
2926 terminated too, by default, unless the user has "lingering" enabled.
2927 To effectively allow users to run long-term tasks even if they are
152199f2
ZJS
2928 logged out, lingering must be enabled for them. See loginctl(1) for
2929 details. The default polkit policy was modified to allow users to
2930 set lingering for themselves without authentication.
7f6e8043 2931
95365a57 2932 Previous defaults can be restored at compile time by the
e40a326c 2933 --without-kill-user-processes option to "configure".
7f6e8043 2934
e75690c3
ZJS
2935 * systemd-logind gained new configuration settings SessionsMax= and
2936 InhibitorsMax=, both with a default of 8192. It will not register new
188d3082 2937 user sessions or inhibitors above this limit.
e75690c3
ZJS
2938
2939 * systemd-logind will now reload configuration on SIGHUP.
2940
96515dbf 2941 * The unified cgroup hierarchy added in Linux 4.5 is now supported.
e40a326c 2942 Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
25b0e6cb
LP
2943 enable. Also, support for the "io" cgroup controller in the unified
2944 hierarchy has been added, so that the "memory", "pids" and "io" are
2945 now the controllers that are supported on the unified hierarchy.
e40a326c 2946
96515dbf
ZJS
2947 WARNING: it is not possible to use previous systemd versions with
2948 systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
2949 is necessary to also update systemd in the initramfs if using the
e40a326c 2950 unified hierarchy. An updated SELinux policy is also required.
96515dbf 2951
e40a326c
LP
2952 * LLDP support has been extended, and both passive (receive-only) and
2953 active (sender) modes are supported. Passive mode ("routers-only") is
8951eaec
ZJS
2954 enabled by default in systemd-networkd. Active LLDP mode is enabled
2955 by default for containers on the internal network. The "networkctl
e40a326c
LP
2956 lldp" command may be used to list information gathered. "networkctl
2957 status" will also show basic LLDP information on connected peers now.
96515dbf 2958
e40a326c
LP
2959 * The IAID and DUID unique identifier sent in DHCP requests may now be
2960 configured for the system and each .network file managed by
e75690c3
ZJS
2961 systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
2962
2963 * systemd-networkd gained support for configuring proxy ARP support for
2964 each interface, via the ProxyArp= setting in .network files. It also
2965 gained support for configuring the multicast querier feature of
2966 bridge devices, via the new MulticastQuerier= setting in .netdev
2967 files. Similarly, snooping on the IGMP traffic can be controlled
2968 via the new setting MulticastSnooping=.
2969
2970 A new setting PreferredLifetime= has been added for addresses
2971 configured in .network file to configure the lifetime intended for an
2972 address.
2973
2974 The systemd-networkd DHCP server gained the option EmitRouter=, which
2975 defaults to yes, to configure whether the DHCP Option 3 (Router)
2976 should be emitted.
96515dbf 2977
e40a326c 2978 * The testing tool /usr/lib/systemd/systemd-activate is renamed to
97e5530c
ZJS
2979 systemd-socket-activate and installed into /usr/bin. It is now fully
2980 supported.
2981
e40a326c
LP
2982 * systemd-journald now uses separate threads to flush changes to disk
2983 when closing journal files, thus reducing impact of slow disk I/O on
2984 logging performance.
96515dbf 2985
e75690c3
ZJS
2986 * The sd-journal API gained two new calls
2987 sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
2988 can be used to open journal files using file descriptors instead of
2989 file or directory paths. sd_journal_open_container() has been
2990 deprecated, sd_journal_open_directory_fd() should be used instead
2991 with the flag SD_JOURNAL_OS_ROOT.
2992
2993 * journalctl learned a new output mode "-o short-unix" that outputs log
2994 lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
2995 UTC). It also gained support for a new --no-hostname setting to
2996 suppress the hostname column in the family of "short" output modes.
2997
e40a326c
LP
2998 * systemd-ask-password now optionally skips printing of the password to
2999 stdout with --no-output which can be useful in scripts.
96515dbf
ZJS
3000
3001 * Framebuffer devices (/dev/fb*) and 3D printers and scanners
3002 (devices tagged with ID_MAKER_TOOL) are now tagged with
3003 "uaccess" and are available to logged in users.
3004
e75690c3 3005 * The DeviceAllow= unit setting now supports specifiers (with "%").
e40a326c
LP
3006
3007 * "systemctl show" gained a new --value switch, which allows print a
3008 only the contents of a specific unit property, without also printing
8951eaec
ZJS
3009 the property's name. Similar support was added to "show*" verbs
3010 of loginctl and machinectl that output "key=value" lists.
e40a326c 3011
e75690c3
ZJS
3012 * A new unit type "generated" was added for files dynamically generated
3013 by generator tools. Similarly, a new unit type "transient" is used
3014 for unit files created using the runtime API. "systemctl enable" will
3015 refuse to operate on such files.
3016
e40a326c
LP
3017 * A new command "systemctl revert" has been added that may be used to
3018 revert to the vendor version of a unit file, in case local changes
3019 have been made by adding drop-ins or overriding the unit file.
3020
3021 * "machinectl clean" gained a new verb to automatically remove all or
3022 just hidden container images.
3023
e40a326c
LP
3024 * systemd-tmpfiles gained support for a new line type "e" for emptying
3025 directories, if they exist, without creating them if they don't.
3026
e40a326c
LP
3027 * systemd-nspawn gained support for automatically patching the UID/GIDs
3028 of the owners and the ACLs of all files and directories in a
3029 container tree to match the UID/GID user namespacing range selected
3030 for the container invocation. This mode is enabled via the new
24597ee0
ZJS
3031 --private-users-chown switch. It also gained support for
3032 automatically choosing a free, previously unused UID/GID range when
3033 starting a container, via the new --private-users=pick setting (which
3034 implies --private-users-chown). Together, these options for the first
3035 time make user namespacing for nspawn containers fully automatic and
3036 thus deployable. The systemd-nspawn@.service template unit file has
3037 been changed to use this functionality by default.
e40a326c 3038
25b0e6cb
LP
3039 * systemd-nspawn gained a new --network-zone= switch, that allows
3040 creating ad-hoc virtual Ethernet links between multiple containers,
3041 that only exist as long as at least one container referencing them is
3042 running. This allows easy connecting of multiple containers with a
3043 common link that implements an Ethernet broadcast domain. Each of
3044 these network "zones" may be named relatively freely by the user, and
3045 may be referenced by any number of containers, but each container may
3046 only reference one of these "zones". On the lower level, this is
3047 implemented by an automatically managed bridge network interface for
3048 each zone, that is created when the first container referencing its
3049 zone is created and removed when the last one referencing its zone
3050 terminates.
3051
e40a326c 3052 * The default start timeout may now be configured on the kernel command
8951eaec
ZJS
3053 line via systemd.default_timeout_start_sec=. It was already
3054 configurable via the DefaultTimeoutStartSec= option in
3055 /etc/systemd/system.conf.
e40a326c 3056
030bd839 3057 * Socket units gained a new TriggerLimitIntervalSec= and
e40a326c
LP
3058 TriggerLimitBurst= setting to configure a limit on the activation
3059 rate of the socket unit.
3060
3061 * The LimitNICE= setting now optionally takes normal UNIX nice values
3062 in addition to the raw integer limit value. If the specified
3063 parameter is prefixed with "+" or "-" and is in the range -20..19 the
3064 value is understood as UNIX nice value. If not prefixed like this it
3065 is understood as raw RLIMIT_NICE limit.
3066
999a43f8
LP
3067 * Note that the effect of the PrivateDevices= unit file setting changed
3068 slightly with this release: the per-device /dev file system will be
3069 mounted read-only from this version on, and will have "noexec"
188d3082 3070 set. This (minor) change of behavior might cause some (exceptional)
999a43f8
LP
3071 legacy software to break, when PrivateDevices=yes is set for its
3072 service. Please leave PrivateDevices= off if you run into problems
3073 with this.
3074
e75690c3
ZJS
3075 * systemd-bootchart has been split out to a separate repository:
3076 https://github.com/systemd/systemd-bootchart
3077
3078 * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
3079 merged into the kernel in its current form.
3080
3081 * The compatibility libraries libsystemd-daemon.so,
3082 libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
3083 which have been deprecated since systemd-209 have been removed along
3084 with the corresponding pkg-config files. All symbols provided by
3085 those libraries are provided by libsystemd.so.
3086
3087 * The Capabilities= unit file setting has been removed (it is ignored
3088 for backwards compatibility). AmbientCapabilities= and
3089 CapabilityBoundingSet= should be used instead.
3090
4f9020fa
DR
3091 * A new special target has been added, initrd-root-device.target,
3092 which creates a synchronization point for dependencies of the root
3093 device in early userspace. Initramfs builders must ensure that this
3094 target is now included in early userspace.
3095
e75690c3
ZJS
3096 Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
3097 Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
3098 Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
3099 Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
3100 Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
3101 R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
3102 Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
3103 Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
77ff6022
CG
3104 Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
3105 Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
e75690c3
ZJS
3106 John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
3107 Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
3108 Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
77ff6022
CG
3109 Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
3110 Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
3111 mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
e75690c3
ZJS
3112 Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
3113 Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
3114 Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
3115 Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
3116 Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
3117 Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
46e40fab
ZJS
3118 Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
3119 Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
3120 Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
3121 Jędrzejewski-Szmek
e40a326c 3122
46e40fab 3123 — Fairfax, 2016-05-21
96515dbf 3124
61f32bff
MP
3125CHANGES WITH 229:
3126
d5f8b295
LP
3127 * The systemd-resolved DNS resolver service has gained a substantial
3128 set of new features, most prominently it may now act as a DNSSEC
3129 validating stub resolver. DNSSEC mode is currently turned off by
ed5f8840
ZJS
3130 default, but is expected to be turned on by default in one of the
3131 next releases. For now, we invite everybody to test the DNSSEC logic
3132 by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
3133 service also gained a full set of D-Bus interfaces, including calls
3134 to configure DNS and DNSSEC settings per link (for use by external
3135 network management software). systemd-resolved and systemd-networkd
3136 now distinguish between "search" and "routing" domains. The former
3137 are used to qualify single-label names, the latter are used purely
3138 for routing lookups within certain domains to specific links.
3139 resolved now also synthesizes RRs for all entries from /etc/hosts.
d5f8b295
LP
3140
3141 * The systemd-resolve tool (which is a client utility for
ed5f8840
ZJS
3142 systemd-resolved) has been improved considerably and is now fully
3143 supported and documented. Hence it has moved from /usr/lib/systemd to
3144 /usr/bin.
d5f8b295
LP
3145
3146 * /dev/disk/by-path/ symlink support has been (re-)added for virtio
3147 devices.
3148
a7c723c0
LP
3149 * The coredump collection logic has been reworked: when a coredump is
3150 collected it is now written to disk, compressed and processed
3151 (including stacktrace extraction) from a new instantiated service
3152 systemd-coredump@.service, instead of directly from the
3153 /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
3154 processing large coredumps can take up a substantial amount of
3155 resources and time, and this previously happened entirely outside of
3156 systemd's service supervision. With the new logic the core_pattern
3157 hook only does minimal metadata collection before passing off control
3158 to the new instantiated service, which is configured with a time
3159 limit, a nice level and other settings to minimize negative impact on
3160 the rest of the system. Also note that the new logic will honour the
3161 RLIMIT_CORE setting of the crashed process, which now allows users
3162 and processes to turn off coredumping for their processes by setting
3163 this limit.
3164
3165 * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
3166 and all forked processes by default. Previously, PID 1 would leave
3167 the setting at "0" for all processes, as set by the kernel. Note that
3168 the resource limit traditionally has no effect on the generated
3169 coredumps on the system if the /proc/sys/kernel/core_pattern hook
3170 logic is used. Since the limit is now honoured (see above) its
3171 default has been changed so that the coredumping logic is enabled by
3172 default for all processes, while allowing specific opt-out.
3173
3174 * When the stacktrace is extracted from processes of system users, this
3175 is now done as "systemd-coredump" user, in order to sandbox this
3176 potentially security sensitive parsing operation. (Note that when
3177 processing coredumps of normal users this is done under the user ID
3178 of process that crashed, as before.) Packagers should take notice
3179 that it is now necessary to create the "systemd-coredump" system user
3180 and group at package installation time.
3181
d5f8b295
LP
3182 * The systemd-activate socket activation testing tool gained support
3183 for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
3184 and --seqpacket switches. It also has been extended to support both
3185 new-style and inetd-style file descriptor passing. Use the new
3186 --inetd switch to request inetd-style file descriptor passing.
3187
8968aea0
MP
3188 * Most systemd tools now honor a new $SYSTEMD_COLORS environment
3189 variable, which takes a boolean value. If set to false, ANSI color
d5f8b295
LP
3190 output is disabled in the tools even when run on a terminal that
3191 supports it.
3192
3193 * The VXLAN support in networkd now supports two new settings
3194 DestinationPort= and PortRange=.
3195
3196 * A new systemd.machine_id= kernel command line switch has been added,
3197 that may be used to set the machine ID in /etc/machine-id if it is
3198 not initialized yet. This command line option has no effect if the
3199 file is already initialized.
3200
3201 * systemd-nspawn gained a new --as-pid2 switch that invokes any
3202 specified command line as PID 2 rather than PID 1 in the
ed5f8840
ZJS
3203 container. In this mode PID 1 is a minimal stub init process that
3204 implements the special POSIX and Linux semantics of PID 1 regarding
3205 signal and child process management. Note that this stub init process
3206 is implemented in nspawn itself and requires no support from the
3207 container image. This new logic is useful to support running
3208 arbitrary commands in the container, as normal processes are
d5f8b295
LP
3209 generally not prepared to run as PID 1.
3210
3211 * systemd-nspawn gained a new --chdir= switch for setting the current
3212 working directory for the process started in the container.
3213
ed5f8840
ZJS
3214 * "journalctl /dev/sda" will now output all kernel log messages for
3215 specified device from the current boot, in addition to all devices
3216 that are parents of it. This should make log output about devices
3217 pretty useful, as long as kernel drivers attach enough metadata to
3218 the log messages. (The usual SATA drivers do.)
d5f8b295
LP
3219
3220 * The sd-journal API gained two new calls
3221 sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
3222 that report whether log data from /run or /var has been found.
3223
3224 * journalctl gained a new switch "--fields" that prints all journal
3225 record field names currently in use in the journal. This is backed
3226 by two new sd-journal API calls sd_journal_enumerate_fields() and
3227 sd_journal_restart_fields().
3228
3229 * Most configurable timeouts in systemd now expect an argument of
8968aea0
MP
3230 "infinity" to turn them off, instead of "0" as before. The semantics
3231 from now on is that a timeout of "0" means "now", and "infinity"
3232 means "never". To maintain backwards compatibility, "0" continues to
3233 turn off previously existing timeout settings.
d5f8b295
LP
3234
3235 * "systemctl reload-or-try-restart" has been renamed to "systemctl
8968aea0
MP
3236 try-reload-or-restart" to clarify what it actually does: the "try"
3237 logic applies to both reloading and restarting, not just restarting.
3238 The old name continues to be accepted for compatibility.
3239
3240 * On boot-up, when PID 1 detects that the system clock is behind the
3241 release date of the systemd version in use, the clock is now set
3242 to the latter. Previously, this was already done in timesyncd, in order
d5f8b295
LP
3243 to avoid running with clocks set to the various clock epochs such as
3244 1902, 1938 or 1970. With this change the logic is now done in PID 1
3245 in addition to timesyncd during early boot-up, so that it is enforced
3246 before the first process is spawned by systemd. Note that the logic
3247 in timesyncd remains, as it is more comprehensive and ensures
ed5f8840 3248 clock monotonicity by maintaining a persistent timestamp file in
d5f8b295
LP
3249 /var. Since /var is generally not available in earliest boot or the
3250 initrd, this part of the logic remains in timesyncd, and is not done
3251 by PID 1.
3252
50f48ad3
DM
3253 * Support for tweaking details in net_cls.class_id through the
3254 NetClass= configuration directive has been removed, as the kernel
3255 people have decided to deprecate that controller in cgroup v2.
3256 Userspace tools such as nftables are moving over to setting rules
3257 that are specific to the full cgroup path of a task, which obsoletes
3258 these controllers anyway. The NetClass= directive is kept around for
3259 legacy compatibility reasons. For a more in-depth description of the
3260 kernel change, please refer to the respective upstream commit:
3261
3262 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
3263
d5f8b295 3264 * A new service setting RuntimeMaxSec= has been added that may be used
8968aea0 3265 to specify a maximum runtime for a service. If the timeout is hit, the
d5f8b295
LP
3266 service is terminated and put into a failure state.
3267
8968aea0
MP
3268 * A new service setting AmbientCapabilities= has been added. It allows
3269 configuration of additional Linux process capabilities that are
3270 passed to the activated processes. This is only available on very
d5f8b295
LP
3271 recent kernels.
3272
3273 * The process resource limit settings in service units may now be used
3274 to configure hard and soft limits individually.
3275
8968aea0 3276 * The various libsystemd APIs such as sd-bus or sd-event now publicly
ed5f8840
ZJS
3277 expose support for gcc's __attribute__((cleanup())) C extension.
3278 Specifically, for many object destructor functions alternative
3279 versions have been added that have names suffixed with "p" and take a
3280 pointer to a pointer to the object to destroy, instead of just a
3281 pointer to the object itself. This is useful because these destructor
3282 functions may be used directly as parameters to the cleanup
3283 construct. Internally, systemd has been a heavy user of this GCC
3284 extension for a long time, and with this change similar support is
3285 now available to consumers of the library outside of systemd. Note
8968aea0 3286 that by using this extension in your sources compatibility with old
ed5f8840
ZJS
3287 and strictly ANSI compatible C compilers is lost. However, all gcc or
3288 LLVM versions of recent years support this extension.
d5f8b295
LP
3289
3290 * Timer units gained support for a new setting RandomizedDelaySec= that
8968aea0
MP
3291 allows configuring some additional randomized delay to the configured
3292 time. This is useful to spread out timer events to avoid load peaks in
3293 clusters or larger setups.
d5f8b295
LP
3294
3295 * Calendar time specifications now support sub-second accuracy.
3296
3297 * Socket units now support listening on SCTP and UDP-lite protocol
3298 sockets.
3299
3300 * The sd-event API now comes with a full set of man pages.
3301
3302 * Older versions of systemd contained experimental support for
3303 compressing journal files and coredumps with the LZ4 compressor that
3304 was not compatible with the lz4 binary (due to API limitations of the
3305 lz4 library). This support has been removed; only support for files
3306 compatible with the lz4 binary remains. This LZ4 logic is now
3307 officially supported and no longer considered experimental.
3308
3309 * The dkr image import logic has been removed again from importd. dkr's
3310 micro-services focus doesn't fit into the machine image focus of
3311 importd, and quickly got out of date with the upstream dkr API.
3312
3313 * Creation of the /run/lock/lockdev/ directory was dropped from
3314 tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
61f32bff
MP
3315 been available for many years. If you still need this, you need to
3316 create your own tmpfiles.d config file with:
d5f8b295
LP
3317
3318 d /run/lock/lockdev 0775 root lock -
61f32bff 3319
dd95b381
LP
3320 * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction=
3321 and RebootArgument= have been moved from the [Service] section of
3322 unit files to [Unit], and they are now supported on all unit types,
3323 not just service units. Of course, systemd will continue to
3324 understand these settings also at the old location, in order to
3325 maintain compatibility.
3326
3545ab35
LP
3327 Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
3328 Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
3329 Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
3330 Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
3331 Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
3332 David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
3333 Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
3334 Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
3335 Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
3336 Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
3337 Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
3338 lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
3339 Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
3340 Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
3341 Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
3342 Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
3343 Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
3344 Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
3345 Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
3346
ccddd104 3347 — Berlin, 2016-02-11
61f32bff 3348
a11c7ea5
LP
3349CHANGES WITH 228:
3350
a11c7ea5
LP
3351 * A number of properties previously only settable in unit
3352 files are now also available as properties to set when
3353 creating transient units programmatically via the bus, as it
3354 is exposed with systemd-run's --property=
3355 setting. Specifically, these are: SyslogIdentifier=,
3356 SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
3357 EnvironmentFile=, ReadWriteDirectories=,
3358 ReadOnlyDirectories=, InaccessibleDirectories=,
3359 ProtectSystem=, ProtectHome=, RuntimeDirectory=.
3360
28c85daf
LP
3361 * When creating transient services via the bus API it is now
3362 possible to pass in a set of file descriptors to use as
3363 STDIN/STDOUT/STDERR for the invoked process.
a11c7ea5 3364
f1f8a5a5
LP
3365 * Slice units may now be created transiently via the bus APIs,
3366 similar to the way service and scope units may already be
3367 created transiently.
3368
a11c7ea5
LP
3369 * Wherever systemd expects a calendar timestamp specification
3370 (like in journalctl's --since= and --until= switches) UTC
3371 timestamps are now supported. Timestamps suffixed with "UTC"
3372 are now considered to be in Universal Time Coordinated
3373 instead of the local timezone. Also, timestamps may now
815bb5bd 3374 optionally be specified with sub-second accuracy. Both of
a11c7ea5
LP
3375 these additions also apply to recurring calendar event
3376 specification, such as OnCalendar= in timer units.
3377
28c85daf
LP
3378 * journalctl gained a new "--sync" switch that asks the
3379 journal daemon to write all so far unwritten log messages to
3380 disk and sync the files, before returning.
3381
a11c7ea5
LP
3382 * systemd-tmpfiles learned two new line types "q" and "Q" that
3383 operate like "v", but also set up a basic btrfs quota
3384 hierarchy when used on a btrfs file system with quota
3385 enabled.
3386
f1f8a5a5
LP
3387 * tmpfiles' "v", "q" and "Q" will now create a plain directory
3388 instead of a subvolume (even on a btrfs file system) if the
3389 root directory is a plain directory, and not a
3390 subvolume. This should simplify things with certain chroot()
3391 environments which are not aware of the concept of btrfs
3392 subvolumes.
3393
a11c7ea5
LP
3394 * systemd-detect-virt gained a new --chroot switch to detect
3395 whether execution takes place in a chroot() environment.
3396
28c85daf 3397 * CPUAffinity= now takes CPU index ranges in addition to
a11c7ea5
LP
3398 individual indexes.
3399
28c85daf
LP
3400 * The various memory-related resource limit settings (such as
3401 LimitAS=) now understand the usual K, M, G, ... suffixes to
3402 the base of 1024 (IEC). Similar, the time-related resource
3403 limit settings understand the usual min, h, day, ...
3404 suffixes now.
3405
f1f8a5a5
LP
3406 * There's a new system.conf setting DefaultTasksMax= to
3407 control the default TasksMax= setting for services and
3408 scopes running on the system. (TasksMax= is the primary
3409 setting that exposes the "pids" cgroup controller on systemd
3410 and was introduced in the previous systemd release.) The
3411 setting now defaults to 512, which means services that are
3412 not explicitly configured otherwise will only be able to
3413 create 512 processes or threads at maximum, from this
3414 version on. Note that this means that thread- or
3415 process-heavy services might need to be reconfigured to set
3416 TasksMax= to a higher value. It is sufficient to set
3417 TasksMax= in these specific unit files to a higher value, or
3418 even "infinity". Similar, there's now a logind.conf setting
3419 UserTasksMax= that defaults to 4096 and limits the total
3420 number of processes or tasks each user may own
3421 concurrently. nspawn containers also have the TasksMax=
3422 value set by default now, to 8192. Note that all of this
3423 only has an effect if the "pids" cgroup controller is
3424 enabled in the kernel. The general benefit of these changes
3425 should be a more robust and safer system, that provides a
3426 certain amount of per-service fork() bomb protection.
3427
28c85daf
LP
3428 * systemd-nspawn gained the new --network-veth-extra= switch
3429 to define additional and arbitrarily-named virtual Ethernet
3430 links between the host and the container.
3431
3432 * A new service execution setting PassEnvironment= has been
3433 added that allows importing select environment variables
3434 from PID1's environment block into the environment block of
3435 the service.
3436
ddb4b0d3 3437 * Timer units gained support for a new RemainAfterElapse=
595bfe7d 3438 setting which takes a boolean argument. It defaults to on,
ddb4b0d3
LP
3439 exposing behaviour unchanged to previous releases. If set to
3440 off, timer units are unloaded after they elapsed if they
3441 cannot elapse again. This is particularly useful for
3442 transient timer units, which shall not stay around longer
3443 than until they first elapse.
3444
a11c7ea5 3445 * systemd will now bump the net.unix.max_dgram_qlen to 512 by
28c85daf
LP
3446 default now (the kernel default is 16). This is beneficial
3447 for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
a11c7ea5
LP
3448 allows substantially larger numbers of queued
3449 datagrams. This should increase the capability of systemd to
3450 parallelize boot-up, as logging and sd_notify() are unlikely
3451 to stall execution anymore. If you need to change the value
3452 from the new defaults, use the usual sysctl.d/ snippets.
3453
28c85daf
LP
3454 * The compression framing format used by the journal or
3455 coredump processing has changed to be in line with what the
3456 official LZ4 tools generate. LZ4 compression support in
3457 systemd was considered unsupported previously, as the format
3458 was not compatible with the normal tools. With this release
3459 this has changed now, and it is hence safe for downstream
3460 distributions to turn it on. While not compressing as well
815bb5bd 3461 as the XZ, LZ4 is substantially faster, which makes
28c85daf
LP
3462 it a good default choice for the compression logic in the
3463 journal and in coredump handling.
a11c7ea5 3464
28c85daf
LP
3465 * Any reference to /etc/mtab has been dropped from
3466 systemd. The file has been obsolete since a while, but
3467 systemd refused to work on systems where it was incorrectly
815bb5bd 3468 set up (it should be a symlink or non-existent). Please make
28c85daf
LP
3469 sure to update to util-linux 2.27.1 or newer in conjunction
3470 with this systemd release, which also drops any reference to
3471 /etc/mtab. If you maintain a distribution make sure that no
3472 software you package still references it, as this is a
3473 likely source of bugs. There's also a glibc bug pending,
3474 asking for removal of any reference to this obsolete file:
3475
3476 https://sourceware.org/bugzilla/show_bug.cgi?id=19108
a11c7ea5 3477
d5bd92bb
LP
3478 Note that only util-linux versions built with
3479 --enable-libmount-force-mountinfo are supported.
3480
a11c7ea5
LP
3481 * Support for the ".snapshot" unit type has been removed. This
3482 feature turned out to be little useful and little used, and
3483 has now been removed from the core and from systemctl.
3484
b9e2f7eb
LP
3485 * The dependency types RequiresOverridable= and
3486 RequisiteOverridable= have been removed from systemd. They
3487 have been used only very sparingly to our knowledge and
3488 other options that provide a similar effect (such as
3489 systemctl --mode=ignore-dependencies) are much more useful
3490 and commonly used. Moreover, they were only half-way
3491 implemented as the option to control behaviour regarding
3492 these dependencies was never added to systemctl. By removing
3493 these dependency types the execution engine becomes a bit
3494 simpler. Unit files that use these dependencies should be
3495 changed to use the non-Overridable dependency types
3496 instead. In fact, when parsing unit files with these
3497 options, that's what systemd will automatically convert them
3498 too, but it will also warn, asking users to fix the unit
3499 files accordingly. Removal of these dependency types should
3500 only affect a negligible number of unit files in the wild.
3501
3502 * Behaviour of networkd's IPForward= option changed
3503 (again). It will no longer maintain a per-interface setting,
3504 but propagate one way from interfaces where this is enabled
3505 to the global kernel setting. The global setting will be
3506 enabled when requested by a network that is set up, but
3507 never be disabled again. This change was made to make sure
3508 IPv4 and IPv6 behaviour regarding packet forwarding is
3509 similar (as the Linux IPv6 stack does not support
3510 per-interface control of this setting) and to minimize
3511 surprises.
3512
28c85daf
LP
3513 * In unit files the behaviour of %u, %U, %h, %s has
3514 changed. These specifiers will now unconditionally resolve
3515 to the various user database fields of the user that the
3516 systemd instance is running as, instead of the user
3517 configured in the specific unit via User=. Note that this
3518 effectively doesn't change much, as resolving of these
3519 specifiers was already turned off in the --system instance
3520 of systemd, as we cannot do NSS lookups from PID 1. In the
3521 --user instance of systemd these specifiers where correctly
3522 resolved, but hardly made any sense, since the user instance
3523 lacks privileges to do user switches anyway, and User= is
ce830873 3524 hence useless. Moreover, even in the --user instance of
28c85daf
LP
3525 systemd behaviour was awkward as it would only take settings
3526 from User= assignment placed before the specifier into
3527 account. In order to unify and simplify the logic around
3528 this the specifiers will now always resolve to the
3529 credentials of the user invoking the manager (which in case
3530 of PID 1 is the root user).
3531
3532 Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
3533 Yang, Daniel Machon, Daniel Mack, David Herrmann, David
3534 Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
f1f8a5a5
LP
3535 Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
3536 Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
3537 Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
3538 Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
3539 Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
3540 Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
3541 Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
3542 Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
3543 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
3544 Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
3545 Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
3546 Jędrzejewski-Szmek
28c85daf 3547
ccddd104 3548 — Berlin, 2015-11-18
a11c7ea5 3549
c97e586d
DM
3550CHANGES WITH 227:
3551
3552 * systemd now depends on util-linux v2.27. More specifically,
3553 the newly added mount monitor feature in libmount now
3554 replaces systemd's former own implementation.
3555
3556 * libmount mandates /etc/mtab not to be regular file, and
3557 systemd now enforces this condition at early boot.
3558 /etc/mtab has been deprecated and warned about for a very
3559 long time, so systems running systemd should already have
3560 stopped having this file around as anything else than a
3561 symlink to /proc/self/mounts.
3562
d046fb93
LP
3563 * Support for the "pids" cgroup controller has been added. It
3564 allows accounting the number of tasks in a cgroup and
c97e586d
DM
3565 enforcing limits on it. This adds two new setting
3566 TasksAccounting= and TasksMax= to each unit, as well as a
6fd5517b 3567 global option DefaultTasksAccounting=.
c97e586d
DM
3568
3569 * Support for the "net_cls" cgroup controller has been added.
fe08a30b
LP
3570 It allows assigning a net class ID to each task in the
3571 cgroup, which can then be used in firewall rules and traffic
3572 shaping configurations. Note that the kernel netfilter net
3573 class code does not currently work reliably for ingress
3574 packets on unestablished sockets.
c97e586d
DM
3575
3576 This adds a new config directive called NetClass= to CGroup
6fd5517b 3577 enabled units. Allowed values are positive numbers for fixed
c97e586d
DM
3578 assignments and "auto" for picking a free value
3579 automatically.
3580
21d86c61
DM
3581 * 'systemctl is-system-running' now returns 'offline' if the
3582 system is not booted with systemd. This command can now be
3583 used as a substitute for 'systemd-notify --booted'.
3584
3585 * Watchdog timeouts have been increased to 3 minutes for all
3586 in-tree service files. Apparently, disk IO issues are more
3587 frequent than we hoped, and user reported >1 minute waiting
3588 for disk IO.
3589
3590 * 'machine-id-commit' functionality has been merged into
3591 'machine-id-setup --commit'. The separate binary has been
3592 removed.
3593
d046fb93
LP
3594 * The WorkingDirectory= directive in unit files may now be set
3595 to the special value '~'. In this case, the working
3596 directory is set to the home directory of the user
3597 configured in User=.
21d86c61 3598
fe08a30b
LP
3599 * "machinectl shell" will now open the shell in the home
3600 directory of the selected user by default.
3601
21d86c61 3602 * The CrashChVT= configuration file setting is renamed to
d046fb93
LP
3603 CrashChangeVT=, following our usual logic of not
3604 abbreviating unnecessarily. The old directive is still
3605 supported for compat reasons. Also, this directive now takes
3606 an integer value between 1 and 63, or a boolean value. The
3607 formerly supported '-1' value for disabling stays around for
3608 compat reasons.
21d86c61 3609
fe08a30b 3610 * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
8b5f9d15 3611 NoNewPrivileges=, TTYPath=, WorkingDirectory= and
fe08a30b
LP
3612 RootDirectory= properties can now be set for transient
3613 units.
3614
3615 * The systemd-analyze tool gained a new "set-log-target" verb
3616 to change the logging target the system manager logs to
3617 dynamically during runtime. This is similar to how
3618 "systemd-analyze set-log-level" already changes the log
3619 level.
3620
3621 * In nspawn /sys is now mounted as tmpfs, with only a selected
3622 set of subdirectories mounted in from the real sysfs. This
3623 enhances security slightly, and is useful for ensuring user
3624 namespaces work correctly.
3625
3626 * Support for USB FunctionFS activation has been added. This
3627 allows implementation of USB gadget services that are
3628 activated as soon as they are requested, so that they don't
595bfe7d 3629 have to run continuously, similar to classic socket
fe08a30b
LP
3630 activation.
3631
3632 * The "systemctl exit" command now optionally takes an
3633 additional parameter that sets the exit code to return from
3634 the systemd manager when exiting. This is only relevant when
3635 running the systemd user instance, or when running the
3636 system instance in a container.
3637
3638 * sd-bus gained the new API calls sd_bus_path_encode_many()
3639 and sd_bus_path_decode_many() that allow easy encoding and
3640 decoding of multiple identifier strings inside a D-Bus
3641 object path. Another new call sd_bus_default_flush_close()
3642 has been added to flush and close per-thread default
3643 connections.
3644
3645 * systemd-cgtop gained support for a -M/--machine= switch to
3646 show the control groups within a certain container only.
3647
3648 * "systemctl kill" gained support for an optional --fail
3649 switch. If specified the requested operation will fail of no
3650 processes have been killed, because the unit had no
3651 processes attached, or similar.
3652
bdba9227
DM
3653 * A new systemd.crash_reboot=1 kernel command line option has
3654 been added that triggers a reboot after crashing. This can
3655 also be set through CrashReboot= in systemd.conf.
3656
3657 * The RuntimeDirectory= setting now understands unit
3658 specifiers like %i or %f.
3659
ce830873 3660 * A new (still internal) library API sd-ipv4acd has been added,
fe08a30b
LP
3661 that implements address conflict detection for IPv4. It's
3662 based on code from sd-ipv4ll, and will be useful for
3663 detecting DHCP address conflicts.
3664
bdba9227
DM
3665 * File descriptors passed during socket activation may now be
3666 named. A new API sd_listen_fds_with_names() is added to
a8eaaee7 3667 access the names. The default names may be overridden,
bdba9227
DM
3668 either in the .socket file using the FileDescriptorName=
3669 parameter, or by passing FDNAME= when storing the file
3670 descriptors using sd_notify().
fe08a30b 3671
d046fb93
LP
3672 * systemd-networkd gained support for:
3673
0053598f 3674 - Setting the IPv6 Router Advertisement settings via
edf4126f 3675 IPv6AcceptRouterAdvertisements= in .network files.
d046fb93
LP
3676
3677 - Configuring the HelloTimeSec=, MaxAgeSec= and
3678 ForwardDelaySec= bridge parameters in .netdev files.
3679
3680 - Configuring PreferredSource= for static routes in
edf4126f 3681 .network files.
fe08a30b 3682
bdba9227
DM
3683 * The "ask-password" framework used to query for LUKS harddisk
3684 passwords or SSL passwords during boot gained support for
3685 caching passwords in the kernel keyring, if it is
3686 available. This makes sure that the user only has to type in
3687 a passphrase once if there are multiple objects to unlock
3688 with the same one. Previously, such password caching was
3689 available only when Plymouth was used; this moves the
3690 caching logic into the systemd codebase itself. The
3691 "systemd-ask-password" utility gained a new --keyname=
3692 switch to control which kernel keyring key to use for
3693 caching a password in. This functionality is also useful for
3694 enabling display managers such as gdm to automatically
3695 unlock the user's GNOME keyring if its passphrase, the
3696 user's password and the harddisk password are the same, if
3697 gdm-autologin is used.
fe08a30b
LP
3698
3699 * When downloading tar or raw images using "machinectl
3700 pull-tar" or "machinectl pull-raw", a matching ".nspawn"
3701 file is now also downloaded, if it is available and stored
3702 next to the image file.
c97e586d 3703
91d0d699
LP
3704 * Units of type ".socket" gained a new boolean setting
3705 Writable= which is only useful in conjunction with
3706 ListenSpecial=. If true, enables opening the specified
3707 special file in O_RDWR mode rather than O_RDONLY mode.
3708
3709 * systemd-rfkill has been reworked to become a singleton
3710 service that is activated through /dev/rfkill on each rfkill
3711 state change and saves the settings to disk. This way,
3712 systemd-rfkill is now compatible with devices that exist
3713 only intermittendly, and even restores state if the previous
3714 system shutdown was abrupt rather than clean.
3715
d046fb93
LP
3716 * The journal daemon gained support for vacuuming old journal
3717 files controlled by the number of files that shall remain,
3718 in addition to the already existing control by size and by
3719 date. This is useful as journal interleaving performance
6dd6a9c4 3720 degrades with too many separate journal files, and allows
d046fb93
LP
3721 putting an effective limit on them. The new setting defaults
3722 to 100, but this may be changed by setting SystemMaxFiles=
3723 and RuntimeMaxFiles= in journald.conf. Also, the
3724 "journalctl" tool gained the new --vacuum-files= switch to
3725 manually vacuum journal files to leave only the specified
3726 number of files in place.
c48eb61f 3727
bdba9227
DM
3728 * udev will now create /dev/disk/by-path links for ATA devices
3729 on kernels where that is supported.
c30f086f 3730
efce0ffe 3731 * Galician, Serbian, Turkish and Korean translations were added.
c97e586d 3732
61e6771c
LP
3733 Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
3734 Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
3735 (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
3736 Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
3737 Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
3738 de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
3739 Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
3740 Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
3741 Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
3742 Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
3743 Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
3744 Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
3745 Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
3746 Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
3747 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
3748 Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
3749 Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
3750 Zbigniew Jędrzejewski-Szmek, Марко М. Костић
3751
ccddd104 3752 — Berlin, 2015-10-07
c97e586d 3753
c9912c5e
DH
3754CHANGES WITH 226:
3755
5e8d4254
LP
3756 * The DHCP implementation of systemd-networkd gained a set of
3757 new features:
3758
3759 - The DHCP server now supports emitting DNS and NTP
3760 information. It may be enabled and configured via
3761 EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
3762 and NTP information is enabled, but no servers are
3763 configured, the corresponding uplink information (if there
3764 is any) is propagated.
3765
3766 - Server and client now support transmission and reception
3767 of timezone information. It can be configured via the
3768 newly introduced network options UseTimezone=,
3769 EmitTimezone=, and Timezone=. Transmission of timezone
3770 information is enabled between host and containers by
3771 default now: the container will change its local timezone
3772 to what the host has set.
3773
3774 - Lease timeouts can now be configured via
3775 MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
3776
3777 - The DHCP server improved on the stability of
3778 leases. Clients are more likely to get the same lease
3779 information back, even if the server loses state.
3780
3781 - The DHCP server supports two new configuration options to
3782 control the lease address pool metrics, PoolOffset= and
3783 PoolSize=.
3784
3785 * The encapsulation limit of tunnels in systemd-networkd may
3786 now be configured via 'EncapsulationLimit='. It allows
3787 modifying the maximum additional levels of encapsulation
3788 that are permitted to be prepended to a packet.
3789
3790 * systemd now supports the concept of user buses replacing
3791 session buses, if used with dbus-1.10 (and enabled via dbus
3792 --enable-user-session). It previously only supported this on
3793 kdbus-enabled systems, and this release expands this to
3794 'dbus-daemon' systems.
3795
3796 * systemd-networkd now supports predictable interface names
3797 for virtio devices.
3798
3799 * systemd now optionally supports the new Linux kernel
3800 "unified" control group hierarchy. If enabled via the kernel
3801 command-line option 'systemd.unified_cgroup_hierarchy=1',
3802 systemd will try to mount the unified cgroup hierarchy
3803 directly on /sys/fs/cgroup. If not enabled, or not
3804 available, systemd will fall back to the legacy cgroup
3805 hierarchy setup, as before. Host system and containers can
3806 mix and match legacy and unified hierarchies as they
856ca72b 3807 wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
5e8d4254
LP
3808 environment variable to individually select the hierarchy to
3809 use for executed containers. By default, nspawn will use the
3810 unified hierarchy for the containers if the host uses the
3811 unified hierarchy, and the legacy hierarchy otherwise.
3812 Please note that at this point the unified hierarchy is an
3813 experimental kernel feature and is likely to change in one
3814 of the next kernel releases. Therefore, it should not be
3815 enabled by default in downstream distributions yet. The
3816 minimum required kernel version for the unified hierarchy to
3817 work is 4.2. Note that when the unified hierarchy is used
3818 for the first time delegated access to controllers is
3819 safe. Because of this systemd-nspawn containers will get
3820 access to controllers now, as will systemd user
3821 sessions. This means containers and user sessions may now
3822 manage their own resources, partitioning up what the system
3823 grants them.
3824
3825 * A new special scope unit "init.scope" has been introduced
3826 that encapsulates PID 1 of the system. It may be used to
3827 determine resource usage and enforce resource limits on PID
3828 1 itself. PID 1 hence moved out of the root of the control
3829 group tree.
3830
3831 * The cgtop tool gained support for filtering out kernel
3832 threads when counting tasks in a control group. Also, the
3833 count of processes is now recursively summed up by
3834 default. Two options -k and --recursive= have been added to
3835 revert to old behaviour. The tool has also been updated to
3836 work correctly in containers now.
3837
3838 * systemd-nspawn's --bind= and --bind-ro= options have been
3839 extended to allow creation of non-recursive bind mounts.
3840
c626bf1d
DM
3841 * libsystemd gained two new calls sd_pid_get_cgroup() and
3842 sd_peer_get_cgroup() which return the control group path of
5e8d4254
LP
3843 a process or peer of a connected AF_UNIX socket. This
3844 function call is particularly useful when implementing
3845 delegated subtrees support in the control group hierarchy.
3846
3847 * The "sd-event" event loop API of libsystemd now supports
3848 correct dequeuing of real-time signals, without losing
3849 signal events.
3850
d35f51ea
ZJS
3851 * When systemd requests a polkit decision when managing units it
3852 will now add additional fields to the request, including unit
3853 name and desired operation. This enables more powerful polkit
3854 policies, that make decisions depending on these parameters.
c9912c5e 3855
47f5a38c
LP
3856 * nspawn learnt support for .nspawn settings files, that may
3857 accompany the image files or directories of containers, and
3858 may contain additional settings for the container. This is
3859 an alternative to configuring container parameters via the
3860 nspawn command line.
3861
2f77decc
LP
3862 Contributions from: Cristian Rodríguez, Daniel Mack, David
3863 Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
3864 Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
3865 Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
3866 Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
3867 Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
3868 Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
23d08d1b 3869 Andersen, Tom Gundersen, Torstein Husebø
c9912c5e 3870
ccddd104 3871 — Berlin, 2015-09-08
c9912c5e 3872
ec5249a2
DM
3873CHANGES WITH 225:
3874
5e8d4254
LP
3875 * machinectl gained a new verb 'shell' which opens a fresh
3876 shell on the target container or the host. It is similar to
3877 the existing 'login' command of machinectl, but spawns the
3878 shell directly without prompting for username or
3879 password. The pseudo machine '.host' now refers to the local
3880 host and is used by default. Hence, 'machinectl shell' can
3881 be used as replacement for 'su -' which spawns a session as
3882 a fresh systemd unit in a way that is fully isolated from
3883 the originating session.
3884
3885 * systemd-networkd learned to cope with private-zone DHCP
3886 options and allows other programs to query the values.
3887
3888 * SELinux access control when enabling/disabling units is no
d35f51ea
ZJS
3889 longer enforced with this release. The previous implementation
3890 was incorrect, and a new corrected implementation is not yet
3891 available. As unit file operations are still protected via
3892 polkit and D-Bus policy this is not a security problem. Yet,
3893 distributions which care about optimal SELinux support should
3894 probably not stabilize on this release.
5e8d4254
LP
3895
3896 * sd-bus gained support for matches of type "arg0has=", that
3897 test for membership of strings in string arrays sent in bus
3898 messages.
3899
3900 * systemd-resolved now dumps the contents of its DNS and LLMNR
3901 caches to the logs on reception of the SIGUSR1 signal. This
3902 is useful to debug DNS behaviour.
3903
3904 * The coredumpctl tool gained a new --directory= option to
3905 operate on journal files in a specific directory.
3906
3907 * "systemctl reboot" and related commands gained a new
3908 "--message=" option which may be used to set a free-text
3909 wall message when shutting down or rebooting the
3910 system. This message is also logged, which is useful for
3911 figuring out the reason for a reboot or shutdown a
3912 posteriori.
3913
3914 * The "systemd-resolve-host" tool's -i switch now takes
3915 network interface numbers as alternative to interface names.
3916
3917 * A new unit file setting for services has been introduced:
3918 UtmpMode= allows configuration of how precisely systemd
3919 handles utmp and wtmp entries for the service if this is
3920 enabled. This allows writing services that appear similar to
3921 user sessions in the output of the "w", "who", "last" and
3922 "lastlog" tools.
3923
3924 * systemd-resolved will now locally synthesize DNS resource
3925 records for the "localhost" and "gateway" domains as well as
3926 the local hostname. This should ensure that clients querying
3927 RRs via resolved will get similar results as those going via
3928 NSS, if nss-myhostname is enabled.
3929
3930 Contributions from: Alastair Hughes, Alex Crawford, Daniel
3931 Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
3932 Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
3933 Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
3934 Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
3935 Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
3936 Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
3937 Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
3938 reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
3939 Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
3940 Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
3941 WaLyong Cho, Zbigniew Jędrzejewski-Szmek
e1439a14 3942
ccddd104 3943 — Berlin, 2015-08-27
ec5249a2 3944
11811e85
DH
3945CHANGES WITH 224:
3946
10fa421c
DH
3947 * The systemd-efi-boot-generator functionality was merged into
3948 systemd-gpt-auto-generator.
3949
5e8d4254
LP
3950 * systemd-networkd now supports Group Policy for vxlan
3951 devices. It can be enabled via the new boolean configuration
3952 option called 'GroupPolicyExtension='.
10fa421c 3953
11811e85
DH
3954 Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
3955 Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
3956 Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
3957
ccddd104 3958 — Berlin, 2015-07-31
11811e85 3959
e57eaef8
DH
3960CHANGES WITH 223:
3961
3962 * The python-systemd code has been removed from the systemd repository.
3963 A new repository has been created which accommodates the code from
3964 now on, and we kindly ask distributions to create a separate package
3965 for this: https://github.com/systemd/python-systemd
3966
01608bc8 3967 * The systemd daemon will now reload its main configuration
e57eaef8
DH
3968 (/etc/systemd/system.conf) on daemon-reload.
3969
3970 * sd-dhcp now exposes vendor specific extensions via
3971 sd_dhcp_lease_get_vendor_specific().
3972
931618d0
DM
3973 * systemd-networkd gained a number of new configuration options.
3974
3975 - A new boolean configuration option for TAP devices called
37d54b93 3976 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
931618d0
DM
3977 device, thus allowing to send and receive GSO packets.
3978
3979 - A new tunnel configuration option called 'CopyDSCP='.
3980 If enabled, the DSCP field of ip6 tunnels is copied into the
3981 decapsulated packet.
3982
3983 - A set of boolean bridge configuration options were added.
3984 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
3985 and 'UnicastFlood=' are now parsed by networkd and applied to the
3986 respective bridge link device via the respective IFLA_BRPORT_*
3987 netlink attribute.
3988
3989 - A new string configuration option to override the hostname sent
3990 to a DHCP server, called 'Hostname='. If set and 'SendHostname='
3991 is true, networkd will use the configured hostname instead of the
3992 system hostname when sending DHCP requests.
3993
3994 - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
3995 networkd will configure the IPv6 flow-label of the tunnel device
3996 according to RFC2460.
e57eaef8 3997
f5f113f6
DH
3998 - The 'macvtap' virtual network devices are now supported, similar to
3999 the already supported 'macvlan' devices.
4000
e57eaef8 4001 * systemd-resolved now implements RFC5452 to improve resilience against
01608bc8 4002 cache poisoning. Additionally, source port randomization is enabled
e57eaef8
DH
4003 by default to further protect against DNS spoofing attacks.
4004
4005 * nss-mymachines now supports translating UIDs and GIDs of running
4006 containers with user-namespaces enabled. If a container 'foo'
4007 translates a host uid 'UID' to the container uid 'TUID', then
4008 nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
4009 (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
4010 mapped as 'vg-foo-TGID'.
4011
4012 Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
e4e66993
DH
4013 Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
4014 HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
4015 Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
4016 Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
4017 Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
4018 Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
4019 Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
4020 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
4021 Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
4022
ccddd104 4023 — Berlin, 2015-07-29
e57eaef8 4024
0db83ad7 4025CHANGES WITH 222:
5541c889 4026
861b02eb
KS
4027 * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
4028 There are no known issues with current sysfs, and udev does not need
4029 or should be used to work around such bugs.
4030
4031 * udev does no longer enable USB HID power management. Several reports
4032 indicate, that some devices cannot handle that setting.
0db83ad7
DH
4033
4034 * The udev accelerometer helper was removed. The functionality
4035 is now fully included in iio-sensor-proxy. But this means,
4036 older iio-sensor-proxy versions will no longer provide
4037 accelerometer/orientation data with this systemd version.
4038 Please upgrade iio-sensor-proxy to version 1.0.
4039
5541c889
DH
4040 * networkd gained a new configuration option IPv6PrivacyExtensions=
4041 which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
4042 for Stateless Address") on selected networks.
4043
9b361114
DM
4044 * For the sake of fewer build-time dependencies and less code in the
4045 main repository, the python bindings are about to be removed in the
4046 next release. A new repository has been created which accommodates
4047 the code from now on, and we kindly ask distributions to create a
4048 separate package for this. The removal will take place in v223.
4049
4050 https://github.com/systemd/python-systemd
4051
0db83ad7
DH
4052 Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
4053 Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
4054 daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
4055 Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
5541c889
DH
4056 Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
4057 (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
0db83ad7
DH
4058 Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
4059 Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
2d1ca112
DH
4060 Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
4061 Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
0db83ad7 4062
ccddd104 4063 — Berlin, 2015-07-07
0db83ad7 4064
0f0467e6
MP
4065CHANGES WITH 221:
4066
470e72d4 4067 * The sd-bus.h and sd-event.h APIs have now been declared
5f92d24f 4068 stable and have been added to the official interface of
470e72d4
LP
4069 libsystemd.so. sd-bus implements an alternative D-Bus client
4070 library, that is relatively easy to use, very efficient and
4071 supports both classic D-Bus as well as kdbus as transport
4072 backend. sd-event is a generic event loop abstraction that
4073 is built around Linux epoll, but adds features such as event
0aee49d5 4074 prioritization or efficient timer handling. Both APIs are good
470e72d4
LP
4075 choices for C programs looking for a bus and/or event loop
4076 implementation that is minimal and does not have to be
5f92d24f 4077 portable to other kernels.
0f0467e6 4078
470e72d4
LP
4079 * kdbus support is no longer compile-time optional. It is now
4080 always built-in. However, it can still be disabled at
4081 runtime using the kdbus=0 kernel command line setting, and
c6551464 4082 that setting may be changed to default to off, by specifying
470e72d4
LP
4083 --disable-kdbus at build-time. Note though that the kernel
4084 command line setting has no effect if the kdbus.ko kernel
4085 module is not installed, in which case kdbus is (obviously)
4086 also disabled. We encourage all downstream distributions to
0aee49d5 4087 begin testing kdbus by adding it to the kernel images in the
470e72d4
LP
4088 development distributions, and leaving kdbus support in
4089 systemd enabled.
0f0467e6 4090
470e72d4
LP
4091 * The minimal required util-linux version has been bumped to
4092 2.26.
4093
4094 * Support for chkconfig (--enable-chkconfig) was removed in
0aee49d5 4095 favor of calling an abstraction tool
470e72d4
LP
4096 /lib/systemd/systemd-sysv-install. This needs to be
4097 implemented for your distribution. See "SYSV INIT.D SCRIPTS"
4098 in README for details.
4099
4100 * If there's a systemd unit and a SysV init script for the
4101 same service name, and the user executes "systemctl enable"
4102 for it (or a related call), then this will now enable both
4103 (or execute the related operation on both), not just the
4104 unit.
4105
4106 * The libudev API documentation has been converted from gtkdoc
4107 into man pages.
4108
4109 * gudev has been removed from the systemd tree, it is now an
4110 external project.
4111
4112 * The systemd-cgtop tool learnt a new --raw switch to generate
0aee49d5 4113 "raw" (machine parsable) output.
470e72d4
LP
4114
4115 * networkd's IPForwarding= .network file setting learnt the
4116 new setting "kernel", which ensures that networkd does not
4117 change the IP forwarding sysctl from the default kernel
4118 state.
4119
4120 * The systemd-logind bus API now exposes a new boolean
4121 property "Docked" that reports whether logind considers the
4122 system "docked", i.e. connected to a docking station or not.
4123
4124 Contributions from: Alex Crawford, Andreas Pokorny, Andrei
4125 Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
4126 Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
4127 David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
4128 Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
4129 Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
4130 Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
4131 Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
4132 Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
4133 Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
4134 Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
b912e251
LP
4135 Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
4136 Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
4137 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
4138 Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
4139 Fink, Zbigniew Jędrzejewski-Szmek
470e72d4 4140
ccddd104 4141 — Berlin, 2015-06-19
0f0467e6 4142
481a0aa2
LP
4143CHANGES WITH 220:
4144
f7a73a25
DH
4145 * The gudev library has been extracted into a separate repository
4146 available at: https://git.gnome.org/browse/libgudev/
4147 It is now managed as part of the Gnome project. Distributions
4148 are recommended to pass --disable-gudev to systemd and use
4149 gudev from the Gnome project instead. gudev is still included
4150 in systemd, for now. It will be removed soon, though. Please
4151 also see the announcement-thread on systemd-devel:
56cadcb6 4152 https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
f7a73a25 4153
481a0aa2
LP
4154 * systemd now exposes a CPUUsageNSec= property for each
4155 service unit on the bus, that contains the overall consumed
4156 CPU time of a service (the sum of what each process of the
4157 service consumed). This value is only available if
4158 CPUAccounting= is turned on for a service, and is then shown
4159 in the "systemctl status" output.
4160
4161 * Support for configuring alternative mappings of the old SysV
4162 runlevels to systemd targets has been removed. They are now
29d1fcb4 4163 hardcoded in a way that runlevels 2, 3, 4 all map to
481a0aa2
LP
4164 multi-user.target and 5 to graphical.target (which
4165 previously was already the default behaviour).
4166
4167 * The auto-mounter logic gained support for mount point
4168 expiry, using a new TimeoutIdleSec= setting in .automount
4169 units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
4170
4171 * The EFI System Partition (ESP) as mounted to /boot by
4172 systemd-efi-boot-generator will now be unmounted
29d1fcb4 4173 automatically after 2 minutes of not being used. This should
481a0aa2
LP
4174 minimize the risk of ESP corruptions.
4175
4176 * New /etc/fstab options x-systemd.requires= and
4177 x-systemd.requires-mounts-for= are now supported to express
4178 additional dependencies for mounts. This is useful for
4179 journalling file systems that support external journal
4180 devices or overlay file systems that require underlying file
4181 systems to be mounted.
4182
4183 * systemd does not support direct live-upgrades (via systemctl
4184 daemon-reexec) from versions older than v44 anymore. As no
4185 distribution we are aware of shipped such old versions in a
4186 stable release this should not be problematic.
4187
4188 * When systemd forks off a new per-connection service instance
4189 it will now set the $REMOTE_ADDR environment variable to the
4190 remote IP address, and $REMOTE_PORT environment variable to
4191 the remote IP port. This behaviour is similar to the
4192 corresponding environment variables defined by CGI.
4193
4194 * systemd-networkd gained support for uplink failure
4195 detection. The BindCarrier= option allows binding interface
4196 configuration dynamically to the link sense of other
4197 interfaces. This is useful to achieve behaviour like in
4198 network switches.
4199
4200 * systemd-networkd gained support for configuring the DHCP
4201 client identifier to use when requesting leases.
4202
4203 * systemd-networkd now has a per-network UseNTP= option to
4204 configure whether NTP server information acquired via DHCP
4205 is passed on to services like systemd-timesyncd.
4206
4207 * systemd-networkd gained support for vti6 tunnels.
4208
1579dd2c
LP
4209 * Note that systemd-networkd manages the sysctl variable
4210 /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
4211 it is configured for since v219. The variable controls IP
4212 forwarding, and is a per-interface alternative to the global
4213 /proc/sys/net/ipv[46]/ip_forward. This setting is
4214 configurable in the IPForward= option, which defaults to
4215 "no". This means if networkd is used for an interface it is
4216 no longer sufficient to set the global sysctl option to turn
4217 on IP forwarding! Instead, the .network file option
4218 IPForward= needs to be turned on! Note that the
4219 implementation of this behaviour was broken in v219 and has
4220 been fixed in v220.
4221
481a0aa2
LP
4222 * Many bonding and vxlan options are now configurable in
4223 systemd-networkd.
4224
4225 * systemd-nspawn gained a new --property= setting to set unit
4226 properties for the container scope. This is useful for
ce830873 4227 setting resource parameters (e.g. "CPUShares=500") on
481a0aa2
LP
4228 containers started from the command line.
4229
4230 * systemd-nspawn gained a new --private-users= switch to make
4231 use of user namespacing available on recent Linux kernels.
4232
4233 * systemd-nspawn may now be called as part of a shell pipeline
4234 in which case the pipes used for stdin and stdout are passed
4235 directly to the process invoked in the container, without
4236 indirection via a pseudo tty.
4237
4238 * systemd-nspawn gained a new switch to control the UNIX
4239 signal to use when killing the init process of the container
4240 when shutting down.
4241
4242 * systemd-nspawn gained a new --overlay= switch for mounting
4243 overlay file systems into the container using the new kernel
4244 overlayfs support.
4245
4246 * When a container image is imported via systemd-importd and
4247 the host file system is not btrfs, a loopback block device
4248 file is created in /var/lib/machines.raw with a btrfs file
4249 system inside. It is then mounted to /var/lib/machines to
4250 enable btrfs features for container management. The loopback
4251 file and btrfs file system is grown as needed when container
4252 images are imported via systemd-importd.
4253
4254 * systemd-machined/systemd-importd gained support for btrfs
4255 quota, to enforce container disk space limits on disk. This
4256 is exposed in "machinectl set-limit".
4257
4258 * systemd-importd now can import containers from local .tar,
4259 .raw and .qcow2 images, and export them to .tar and .raw. It
4260 can also import dkr v2 images now from the network (on top
4261 of v1 as before).
4262
4263 * systemd-importd gained support for verifying downloaded
4264 images with gpg2 (previously only gpg1 was supported).
4265
d35f51ea
ZJS
4266 * systemd-machined, systemd-logind, systemd: most bus calls are
4267 now accessible to unprivileged processes via polkit. Also,
4268 systemd-logind will now allow users to kill their own sessions
4269 without further privileges or authorization.
481a0aa2
LP
4270
4271 * systemd-shutdownd has been removed. This service was
4272 previously responsible for implementing scheduled shutdowns
4273 as exposed in /usr/bin/shutdown's time parameter. This
4274 functionality has now been moved into systemd-logind and is
4275 accessible via a bus interface.
4276
4277 * "systemctl reboot" gained a new switch --firmware-setup that
4278 can be used to reboot into the EFI firmware setup, if that
4279 is available. systemd-logind now exposes an API on the bus
4280 to trigger such reboots, in case graphical desktop UIs want
4281 to cover this functionality.
4282
4283 * "systemctl enable", "systemctl disable" and "systemctl mask"
1579dd2c 4284 now support a new "--now" switch. If specified the units
481a0aa2
LP
4285 that are enabled will also be started, and the ones
4286 disabled/masked also stopped.
4287
4288 * The Gummiboot EFI boot loader tool has been merged into
1a2d5fbe
DH
4289 systemd, and renamed to "systemd-boot". The bootctl tool has been
4290 updated to support systemd-boot.
481a0aa2
LP
4291
4292 * An EFI kernel stub has been added that may be used to create
4293 kernel EFI binaries that contain not only the actual kernel,
4294 but also an initrd, boot splash, command line and OS release
4295 information. This combined binary can then be signed as a
4296 single image, so that the firmware can verify it all in one
1a2d5fbe 4297 step. systemd-boot has special support for EFI binaries created
481a0aa2
LP
4298 like this and can extract OS release information from them
4299 and show them in the boot menu. This functionality is useful
4300 to implement cryptographically verified boot schemes.
4301
4302 * Optional support has been added to systemd-fsck to pass
4303 fsck's progress report to an AF_UNIX socket in the file
4304 system.
4305
4306 * udev will no longer create device symlinks for all block
4307 devices by default. A blacklist for excluding special block
4308 devices from this logic has been turned into a whitelist
4309 that requires picking block devices explicitly that require
4310 device symlinks.
4311
4312 * A new (currently still internal) API sd-device.h has been
4313 added to libsystemd. This modernized API is supposed to
4314 replace libudev eventually. In fact, already much of libudev
4315 is now just a wrapper around sd-device.h.
4316
4317 * A new hwdb database for storing metadata about pointing
4318 stick devices has been added.
4319
4320 * systemd-tmpfiles gained support for setting file attributes
4321 similar to the "chattr" tool with new 'h' and 'H' lines.
4322
4323 * systemd-journald will no longer unconditionally set the
4324 btrfs NOCOW flag on new journal files. This is instead done
4325 with tmpfiles snippet using the new 'h' line type. This
4326 allows easy disabling of this logic, by masking the
4327 journal-nocow.conf tmpfiles file.
4328
4329 * systemd-journald will now translate audit message types to
4330 human readable identifiers when writing them to the
4331 journal. This should improve readability of audit messages.
4332
4333 * The LUKS logic gained support for the offset= and skip=
4334 options in /etc/crypttab, as previously implemented by
4335 Debian.
4336
4337 * /usr/lib/os-release gained a new optional field VARIANT= for
4338 distributions that support multiple variants (such as a
4339 desktop edition, a server edition, ...)
4340
4341 Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
4342 Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
4343 Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
4344 Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
4345 Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
4346 Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
4347 Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
4348 Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
4349 Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
4350 Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
4351 Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
4352 Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
4353 Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
4354 Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
4355 De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
4356 Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
4357 Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
4358 Michael Biebl, Michael Marineau, Michael Olbrich, Michal
4359 Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
4360 Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
4361 Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
4362 Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
4363 Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
4364 Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
4365 Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
4366 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
4367 Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
4368
ccddd104 4369 — Berlin, 2015-05-22
481a0aa2 4370
615aaf41
LP
4371CHANGES WITH 219:
4372
615aaf41
LP
4373 * Introduce a new API "sd-hwdb.h" for querying the hardware
4374 metadata database. With this minimal interface one can query
4375 and enumerate the udev hwdb, decoupled from the old libudev
4376 library. libudev's interface for this is now only a wrapper
4377 around sd-hwdb. A new tool systemd-hwdb has been added to
4378 interface with and update the database.
4379
4380 * When any of systemd's tools copies files (for example due to
4381 tmpfiles' C lines) a btrfs reflink will attempted first,
4382 before bytewise copying is done.
4383
4384 * systemd-nspawn gained a new --ephemeral switch. When
4385 specified a btrfs snapshot is taken of the container's root
4386 directory, and immediately removed when the container
4387 terminates again. Thus, a container can be started whose
4388 changes never alter the container's root directory, and are
4389 lost on container termination. This switch can also be used
4390 for starting a container off the root file system of the
4391 host without affecting the host OS. This switch is only
4392 available on btrfs file systems.
4393
4394 * systemd-nspawn gained a new --template= switch. It takes the
4395 path to a container tree to use as template for the tree
7edecf21 4396 specified via --directory=, should that directory be
615aaf41
LP
4397 missing. This allows instantiating containers dynamically,
4398 on first run. This switch is only available on btrfs file
4399 systems.
4400
4401 * When a .mount unit refers to a mount point on which multiple
4402 mounts are stacked, and the .mount unit is stopped all of
4403 the stacked mount points will now be unmounted until no
4404 mount point remains.
4405
4406 * systemd now has an explicit notion of supported and
4407 unsupported unit types. Jobs enqueued for unsupported unit
4408 types will now fail with an "unsupported" error code. More
4409 specifically .swap, .automount and .device units are not
4410 supported in containers, .busname units are not supported on
4411 non-kdbus systems. .swap and .automount are also not
4412 supported if their respective kernel compile time options
4413 are disabled.
4414
4415 * machinectl gained support for two new "copy-from" and
4416 "copy-to" commands for copying files from a running
4417 container to the host or vice versa.
4418
4419 * machinectl gained support for a new "bind" command to bind
4420 mount host directories into local containers. This is
4421 currently only supported for nspawn containers.
4422
4423 * networkd gained support for configuring bridge forwarding
4424 database entries (fdb) from .network files.
4425
4426 * A new tiny daemon "systemd-importd" has been added that can
4427 download container images in tar, raw, qcow2 or dkr formats,
4428 and make them available locally in /var/lib/machines, so
4429 that they can run as nspawn containers. The daemon can GPG
4430 verify the downloads (not supported for dkr, since it has no
4431 provisions for verifying downloads). It will transparently
4432 decompress bz2, xz, gzip compressed downloads if necessary,
4433 and restore sparse files on disk. The daemon uses privilege
4434 separation to ensure the actual download logic runs with
94e5ba37 4435 fewer privileges than the daemon itself. machinectl has
615aaf41
LP
4436 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
4437 make the functionality of importd available to the
4438 user. With this in place the Fedora and Ubuntu "Cloud"
4439 images can be downloaded and booted as containers unmodified
4440 (the Fedora images lack the appropriate GPG signature files
4441 currently, so they cannot be verified, but this will change
4442 soon, hopefully). Note that downloading images is currently
4443 only fully supported on btrfs.
4444
4445 * machinectl is now able to list container images found in
4446 /var/lib/machines, along with some metadata about sizes of
4447 disk and similar. If the directory is located on btrfs and
4448 quota is enabled, this includes quota display. A new command
4449 "image-status" has been added that shows additional
4450 information about images.
4451
4452 * machinectl is now able to clone container images
4453 efficiently, if the underlying file system (btrfs) supports
f59dba26 4454 it, with the new "machinectl clone" command. It also
615aaf41
LP
4455 gained commands for renaming and removing images, as well as
4456 marking them read-only or read-write (supported also on
4457 legacy file systems).
4458
4459 * networkd gained support for collecting LLDP network
4460 announcements, from hardware that supports this. This is
4461 shown in networkctl output.
4462
4463 * systemd-run gained support for a new -t (--pty) switch for
4464 invoking a binary on a pty whose input and output is
4465 connected to the invoking terminal. This allows executing
4466 processes as system services while interactively
4467 communicating with them via the terminal. Most interestingly
4468 this is supported across container boundaries. Invoking
4469 "systemd-run -t /bin/bash" is an alternative to running a
4470 full login session, the difference being that the former
4471 will not register a session, nor go through the PAM session
4472 setup.
4473
4474 * tmpfiles gained support for a new "v" line type for creating
4475 btrfs subvolumes. If the underlying file system is a legacy
4476 file system, this automatically degrades to creating a
4477 normal directory. Among others /var/lib/machines is now
4478 created like this at boot, should it be missing.
4479
4480 * The directory /var/lib/containers/ has been deprecated and
4481 been replaced by /var/lib/machines. The term "machines" has
4482 been used in the systemd context as generic term for both
4483 VMs and containers, and hence appears more appropriate for
4484 this, as the directory can also contain raw images bootable
4485 via qemu/kvm.
4486
4487 * systemd-nspawn when invoked with -M but without --directory=
4488 or --image= is now capable of searching for the container
4489 root directory, subvolume or disk image automatically, in
4490 /var/lib/machines. systemd-nspawn@.service has been updated
4491 to make use of this, thus allowing it to be used for raw
4492 disk images, too.
4493
4494 * A new machines.target unit has been introduced that is
4495 supposed to group all containers/VMs invoked as services on
4496 the system. systemd-nspawn@.service has been updated to
4497 integrate with that.
4498
4499 * machinectl gained a new "start" command, for invoking a
4500 container as a service. "machinectl start foo" is mostly
4501 equivalent to "systemctl start systemd-nspawn@foo.service",
4502 but handles escaping in a nicer way.
4503
4504 * systemd-nspawn will now mount most of the cgroupfs tree
4505 read-only into each container, with the exception of the
4506 container's own subtree in the name=systemd hierarchy.
4507
4508 * journald now sets the special FS_NOCOW file flag for its
4509 journal files. This should improve performance on btrfs, by
4510 avoiding heavy fragmentation when journald's write-pattern
4511 is used on COW file systems. It degrades btrfs' data
4512 integrity guarantees for the files to the same levels as for
4513 ext3/ext4 however. This should be OK though as journald does
4514 its own data integrity checks and all its objects are
4515 checksummed on disk. Also, journald should handle btrfs disk
4516 full events a lot more gracefully now, by processing SIGBUS
4517 errors, and not relying on fallocate() anymore.
4518
4519 * When journald detects that journal files it is writing to
4520 have been deleted it will immediately start new journal
4521 files.
4522
4523 * systemd now provides a way to store file descriptors
4c37970d 4524 per-service in PID 1. This is useful for daemons to ensure
615aaf41 4525 that fds they require are not lost during a daemon
94e5ba37 4526 restart. The fds are passed to the daemon on the next
615aaf41
LP
4527 invocation in the same way socket activation fds are
4528 passed. This is now used by journald to ensure that the
4529 various sockets connected to all the system's stdout/stderr
4530 are not lost when journald is restarted. File descriptors
4531 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
4532 an extension to sd_notify(). Note that a limit is enforced
4533 on the number of fds a service can store in PID 1, and it
4534 defaults to 0, so that no fds may be stored, unless this is
4535 explicitly turned on.
4536
4537 * The default TERM variable to use for units connected to a
4538 terminal, when no other value is explicitly is set is now
4539 vt220 rather than vt102. This should be fairly safe still,
4540 but allows PgUp/PgDn work.
4541
4542 * The /etc/crypttab option header= as known from Debian is now
4543 supported.
4544
4545 * "loginctl user-status" and "loginctl session-status" will
4546 now show the last 10 lines of log messages of the
4547 user/session following the status output. Similar,
4548 "machinectl status" will show the last 10 log lines
4549 associated with a virtual machine or container
4550 service. (Note that this is usually not the log messages
4551 done in the VM/container itself, but simply what the
4552 container manager logs. For nspawn this includes all console
4553 output however.)
4554
4555 * "loginctl session-status" without further argument will now
4556 show the status of the session of the caller. Similar,
4557 "lock-session", "unlock-session", "activate",
4558 "enable-linger", "disable-linger" may now be called without
4559 session/user parameter in which case they apply to the
4560 caller's session/user.
4561
4562 * An X11 session scriptlet is now shipped that uploads
4563 $DISPLAY and $XAUTHORITY into the environment of the systemd
4564 --user daemon if a session begins. This should improve
4565 compatibility with X11 enabled applications run as systemd
4566 user services.
4567
4568 * Generators are now subject to masking via /etc and /run, the
4569 same way as unit files.
4570
4571 * networkd .network files gained support for configuring
4572 per-link IPv4/IPv6 packet forwarding as well as IPv4
4573 masquerading. This is by default turned on for veth links to
4574 containers, as registered by systemd-nspawn. This means that
4575 nspawn containers run with --network-veth will now get
4576 automatic routed access to the host's networks without any
4577 further configuration or setup, as long as networkd runs on
4578 the host.
4579
4580 * systemd-nspawn gained the --port= (-p) switch to expose TCP
4581 or UDP posts of a container on the host. With this in place
4582 it is possible to run containers with private veth links
4583 (--network-veth), and have their functionality exposed on
4584 the host as if their services were running directly on the
4585 host.
4586
dd2fd155 4587 * systemd-nspawn's --network-veth switch now gained a short
615aaf41
LP
4588 version "-n", since with the changes above it is now truly
4589 useful out-of-the-box. The systemd-nspawn@.service has been
4590 updated to make use of it too by default.
4591
4592 * systemd-nspawn will now maintain a per-image R/W lock, to
4593 ensure that the same image is not started more than once
4594 writable. (It's OK to run an image multiple times
4595 simultaneously in read-only mode.)
4596
4597 * systemd-nspawn's --image= option is now capable of
4598 dissecting and booting MBR and GPT disk images that contain
4599 only a single active Linux partition. Previously it
4600 supported only GPT disk images with proper GPT type
4601 IDs. This allows running cloud images from major
4602 distributions directly with systemd-nspawn, without
4603 modification.
4604
4605 * In addition to collecting mouse dpi data in the udev
4606 hardware database, there's now support for collecting angle
4607 information for mouse scroll wheels. The database is
7edecf21 4608 supposed to guarantee similar scrolling behavior on mice
615aaf41
LP
4609 that it knows about. There's also support for collecting
4610 information about Touchpad types.
4611
4612 * udev's input_id built-in will now also collect touch screen
4613 dimension data and attach it to probed devices.
4614
4615 * /etc/os-release gained support for a Distribution Privacy
4616 Policy link field.
4617
4618 * networkd gained support for creating "ipvlan", "gretap",
4619 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
4620
4621 * systemd-tmpfiles gained support for "a" lines for setting
4622 ACLs on files.
4623
4624 * systemd-nspawn will now mount /tmp in the container to
4625 tmpfs, automatically.
4626
4627 * systemd now exposes the memory.usage_in_bytes cgroup
4628 attribute and shows it for each service in the "systemctl
4629 status" output, if available.
4630
4631 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
4632 immediate reboot is triggered. This useful if shutdown is
4633 hung and is unable to complete, to expedite the
4634 operation. Note that this kind of reboot will still unmount
4635 all file systems, and hence should not result in fsck being
4636 run on next reboot.
4637
4638 * A .device unit for an optical block device will now be
4639 considered active only when a medium is in the drive. Also,
4640 mount units are now bound to their backing devices thus
4641 triggering automatic unmounting when devices become
4642 unavailable. With this in place systemd will now
4643 automatically unmount left-over mounts when a CD-ROM is
4644 ejected or an USB stick is yanked from the system.
4645
4646 * networkd-wait-online now has support for waiting for
4647 specific interfaces only (with globbing), and for giving up
4648 after a configurable timeout.
4649
4650 * networkd now exits when idle. It will be automatically
4651 restarted as soon as interfaces show up, are removed or
4652 change state. networkd will stay around as long as there is
4653 at least one DHCP state machine or similar around, that keep
4654 it non-idle.
4655
4656 * networkd may now configure IPv6 link-local addressing in
4657 addition to IPv4 link-local addressing.
4658
4659 * The IPv6 "token" for use in SLAAC may now be configured for
4660 each .network interface in networkd.
4661
4662 * Routes configured with networkd may now be assigned a scope
4663 in .network files.
4664
4665 * networkd's [Match] sections now support globbing and lists
4666 of multiple space-separated matches per item.
4667
11ea2781 4668 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
d2c643c6
LP
4669 Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
4670 Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
4671 Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
4672 Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
4673 Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
4674 Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
4675 Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
4676 Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
4677 Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
4678 Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
4679 Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
4680 Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
4681 Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
4682 Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
11ea2781
LP
4683 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
4684 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
4685 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
4686 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
4687 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
4688 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
4689 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
d2c643c6
LP
4690 Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
4691 Hoffmann, Zbigniew Jędrzejewski-Szmek
11ea2781 4692
ccddd104 4693 — Berlin, 2015-02-16
11ea2781 4694
d4f5a1f4
DH
4695CHANGES WITH 218:
4696
f9e00a9f
LP
4697 * When querying unit file enablement status (for example via
4698 "systemctl is-enabled"), a new state "indirect" is now known
4699 which indicates that a unit might not be enabled itself, but
c7683ffb 4700 another unit listed in its Also= setting might be.
f9e00a9f
LP
4701
4702 * Similar to the various existing ConditionXYZ= settings for
b938cb90 4703 units, there are now matching AssertXYZ= settings. While
f9e00a9f
LP
4704 failing conditions cause a unit to be skipped, but its job
4705 to succeed, failing assertions declared like this will cause
4706 a unit start operation and its job to fail.
4707
4708 * hostnamed now knows a new chassis type "embedded".
4709
4710 * systemctl gained a new "edit" command. When used on a unit
b938cb90 4711 file, this allows extending unit files with .d/ drop-in
f9e00a9f
LP
4712 configuration snippets or editing the full file (after
4713 copying it from /usr/lib to /etc). This will invoke the
4714 user's editor (as configured with $EDITOR), and reload the
4715 modified configuration after editing.
4716
4717 * "systemctl status" now shows the suggested enablement state
4718 for a unit, as declared in the (usually vendor-supplied)
4719 system preset files.
4720
4721 * nss-myhostname will now resolve the single-label host name
4722 "gateway" to the locally configured default IP routing
4723 gateways, ordered by their metrics. This assigns a stable
4724 name to the used gateways, regardless which ones are
4725 currently configured. Note that the name will only be
4726 resolved after all other name sources (if nss-myhostname is
4727 configured properly) and should hence not negatively impact
4728 systems that use the single-label host name "gateway" in
4729 other contexts.
4730
4731 * systemd-inhibit now allows filtering by mode when listing
4732 inhibitors.
4733
122676c9 4734 * Scope and service units gained a new "Delegate" boolean
b938cb90 4735 property, which, when set, allows processes running inside the
122676c9
LP
4736 unit to further partition resources. This is primarily
4737 useful for systemd user instances as well as container
4738 managers.
f9e00a9f
LP
4739
4740 * journald will now pick up audit messages directly from
4741 the kernel, and log them like any other log message. The
4742 audit fields are split up and fully indexed. This means that
4743 journalctl in many ways is now a (nicer!) alternative to
4744 ausearch, the traditional audit client. Note that this
b938cb90 4745 implements only a minimal audit client. If you want the
f9e00a9f
LP
4746 special audit modes like reboot-on-log-overflow, please use
4747 the traditional auditd instead, which can be used in
4748 parallel to journald.
4749
4750 * The ConditionSecurity= unit file option now understands the
4751 special string "audit" to check whether auditing is
4752 available.
4753
4754 * journalctl gained two new commands --vacuum-size= and
4755 --vacuum-time= to delete old journal files until the
a8eaaee7 4756 remaining ones take up no more than the specified size on disk,
f9e00a9f
LP
4757 or are not older than the specified time.
4758
4759 * A new, native PPPoE library has been added to sd-network,
4760 systemd's library of light-weight networking protocols. This
4761 library will be used in a future version of networkd to
4762 enable PPPoE communication without an external pppd daemon.
4763
4764 * The busctl tool now understands a new "capture" verb that
4765 works similar to "monitor", but writes a packet capture
4766 trace to STDOUT that can be redirected to a file which is
4767 compatible with libcap's capture file format. This can then
4768 be loaded in Wireshark and similar tools to inspect bus
4769 communication.
4770
4771 * The busctl tool now understands a new "tree" verb that shows
4772 the object trees of a specific service on the bus, or of all
4773 services.
4774
4775 * The busctl tool now understands a new "introspect" verb that
4776 shows all interfaces and members of objects on the bus,
4777 including their signature and values. This is particularly
4778 useful to get more information about bus objects shown by
4779 the new "busctl tree" command.
4780
4781 * The busctl tool now understands new verbs "call",
4782 "set-property" and "get-property" for invoking bus method
4783 calls, setting and getting bus object properties in a
4784 friendly way.
4785
4786 * busctl gained a new --augment-creds= argument that controls
4787 whether the tool shall augment credential information it
4788 gets from the bus with data from /proc, in a possibly
4789 race-ful way.
4790
4791 * nspawn's --link-journal= switch gained two new values
4792 "try-guest" and "try-host" that work like "guest" and
17c29493 4793 "host", but do not fail if the host has no persistent
f9e00a9f
LP
4794 journalling enabled. -j is now equivalent to
4795 --link-journal=try-guest.
4796
4797 * macvlan network devices created by nspawn will now have
4798 stable MAC addresses.
4799
4800 * A new SmackProcessLabel= unit setting has been added, which
4801 controls the SMACK security label processes forked off by
4802 the respective unit shall use.
4803
d4f5a1f4
DH
4804 * If compiled with --enable-xkbcommon, systemd-localed will
4805 verify x11 keymap settings by compiling the given keymap. It
4806 will spew out warnings if the compilation fails. This
4807 requires libxkbcommon to be installed.
4808
b938cb90 4809 * When a coredump is collected, a larger number of metadata
f9e00a9f 4810 fields is now collected and included in the journal records
b938cb90 4811 created for it. More specifically, control group membership,
f9e00a9f
LP
4812 environment variables, memory maps, working directory,
4813 chroot directory, /proc/$PID/status, and a list of open file
4814 descriptors is now stored in the log entry.
4815
17c29493 4816 * The udev hwdb now contains DPI information for mice. For
f9e00a9f
LP
4817 details see:
4818
4819 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
4820
4821 * All systemd programs that read standalone configuration
4822 files in /etc now also support a corresponding series of
997b2b43
JT
4823 .conf.d configuration directories in /etc/, /run/,
4824 /usr/local/lib/, /usr/lib/, and (if configured with
4825 --enable-split-usr) /lib/. In particular, the following
4826 configuration files now have corresponding configuration
4827 directories: system.conf user.conf, logind.conf,
4828 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
4829 resolved.conf, timesyncd.conf, journal-remote.conf, and
4830 journal-upload.conf. Note that distributions should use the
4831 configuration directories in /usr/lib/; the directories in
4832 /etc/ are reserved for the system administrator.
4833
f9e00a9f
LP
4834 * systemd-rfkill will no longer take the rfkill device name
4835 into account when storing rfkill state on disk, as the name
4836 might be dynamically assigned and not stable. Instead, the
4837 ID_PATH udev variable combined with the rfkill type (wlan,
4838 bluetooth, ...) is used.
4839
4840 * A new service systemd-machine-id-commit.service has been
4841 added. When used on systems where /etc is read-only during
4842 boot, and /etc/machine-id is not initialized (but an empty
4843 file), this service will copy the temporary machine ID
4844 created as replacement into /etc after the system is fully
4845 booted up. This is useful for systems that are freshly
4846 installed with a non-initialized machine ID, but should get
4847 a fixed machine ID for subsequent boots.
4848
4849 * networkd's .netdev files now provide a large set of
a8eaaee7 4850 configuration parameters for VXLAN devices. Similarly, the
f9e00a9f
LP
4851 bridge port cost parameter is now configurable in .network
4852 files. There's also new support for configuring IP source
4853 routing. networkd .link files gained support for a new
4854 OriginalName= match that is useful to match against the
4855 original interface name the kernel assigned. .network files
4856 may include MTU= and MACAddress= fields for altering the MTU
4857 and MAC address while being connected to a specific network
4858 interface.
4859
4860 * The LUKS logic gained supported for configuring
4861 UUID-specific key files. There's also new support for naming
4862 LUKS device from the kernel command line, using the new
4863 luks.name= argument.
4864
4865 * Timer units may now be transiently created via the bus API
4866 (this was previously already available for scope and service
4867 units). In addition it is now possible to create multiple
4868 transient units at the same time with a single bus call. The
4869 "systemd-run" tool has been updated to make use of this for
4870 running commands on a specified time, in at(1)-style.
4871
4872 * tmpfiles gained support for "t" lines, for assigning
4873 extended attributes to files. Among other uses this may be
4874 used to assign SMACK labels to files.
4875
13e92f39
LP
4876 Contributions from: Alin Rauta, Alison Chaiken, Andrej
4877 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
4878 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
4879 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
4880 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
4881 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
4882 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
4883 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
4884 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
4885 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
4886 Michael Biebl, Michael Chapman, Michael Marineau, Michal
7da81d33
LP
4887 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
4888 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
4889 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
4890 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
4891 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
4892 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
4893 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
13e92f39 4894
ccddd104 4895 — Berlin, 2014-12-10
f9e00a9f 4896
b62a309a
ZJS
4897CHANGES WITH 217:
4898
78b6b7ce
LP
4899 * journalctl gained the new options -t/--identifier= to match
4900 on the syslog identifier (aka "tag"), as well as --utc to
4901 show log timestamps in the UTC timezone. journalctl now also
4902 accepts -n/--lines=all to disable line capping in a pager.
b62a309a 4903
a65b8245
ZJS
4904 * journalctl gained a new switch, --flush, that synchronously
4905 flushes logs from /run/log/journal to /var/log/journal if
4906 persistent storage is enabled. systemd-journal-flush.service
4907 now waits until the operation is complete.
2a97b03b 4908
b62a309a
ZJS
4909 * Services can notify the manager before they start a reload
4910 (by sending RELOADING=1) or shutdown (by sending
4bdc60cb
LP
4911 STOPPING=1). This allows the manager to track and show the
4912 internal state of daemons and closes a race condition when
78b6b7ce 4913 the process is still running but has closed its D-Bus
4bdc60cb 4914 connection.
b62a309a 4915
78b6b7ce
LP
4916 * Services with Type=oneshot do not have to have any ExecStart
4917 commands anymore.
b62a309a
ZJS
4918
4919 * User units are now loaded also from
4920 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
4921 /run/systemd/user directory that was already previously
4922 supported, but is under the control of the user.
4923
3f9a0a52 4924 * Job timeouts (i.e. timeouts on the time a job that is
4ffd29fd
LP
4925 queued stays in the run queue) can now optionally result in
4926 immediate reboot or power-off actions (JobTimeoutAction= and
4927 JobTimeoutRebootArgument=). This is useful on ".target"
4928 units, to limit the maximum time a target remains
4929 undispatched in the run queue, and to trigger an emergency
4930 operation in such a case. This is now used by default to
4931 turn off the system if boot-up (as defined by everything in
4932 basic.target) hangs and does not complete for at least
4933 15min. Also, if power-off or reboot hang for at least 30min
4934 an immediate power-off/reboot operation is triggered. This
4935 functionality is particularly useful to increase reliability
4936 on embedded devices, but also on laptops which might
4937 accidentally get powered on when carried in a backpack and
4938 whose boot stays stuck in a hard disk encryption passphrase
4939 question.
4940
b62a309a
ZJS
4941 * systemd-logind can be configured to also handle lid switch
4942 events even when the machine is docked or multiple displays
4943 are attached (HandleLidSwitchDocked= option).
4944
4945 * A helper binary and a service have been added which can be
4946 used to resume from hibernation in the initramfs. A
4947 generator will parse the resume= option on the kernel
81c7dd89 4948 command line to trigger resume.
b62a309a 4949
78b6b7ce
LP
4950 * A user console daemon systemd-consoled has been
4951 added. Currently, it is a preview, and will so far open a
4952 single terminal on each session of the user marked as
09077149 4953 Desktop=systemd-console.
b62a309a
ZJS
4954
4955 * Route metrics can be specified for DHCP routes added by
4956 systemd-networkd.
4957
ba8df74b 4958 * The SELinux context of socket-activated services can be set
78b6b7ce 4959 from the information provided by the networking stack
b62a309a
ZJS
4960 (SELinuxContextFromNet= option).
4961
4962 * Userspace firmware loading support has been removed and
4963 the minimum supported kernel version is thus bumped to 3.7.
4964
4965 * Timeout for udev workers has been increased from 1 to 3
4966 minutes, but a warning will be printed after 1 minute to
4967 help diagnose kernel modules that take a long time to load.
4968
78b6b7ce 4969 * Udev rules can now remove tags on devices with TAG-="foobar".
b62a309a 4970
4bdc60cb 4971 * systemd's readahead implementation has been removed. In many
f6d1de85 4972 circumstances it didn't give expected benefits even for
b62a309a 4973 rotational disk drives and was becoming less relevant in the
78b6b7ce
LP
4974 age of SSDs. As none of the developers has been using
4975 rotating media anymore, and nobody stepped up to actively
4976 maintain this component of systemd it has now been removed.
b62a309a 4977
c4ac9900 4978 * Swap units can use Options= to specify discard options.
b62a309a
ZJS
4979 Discard options specified for swaps in /etc/fstab are now
4980 respected.
4981
4982 * Docker containers are now detected as a separate type of
4983 virtualization.
4984
4985 * The Password Agent protocol gained support for queries where
ba8df74b 4986 the user input is shown, useful e.g. for user names.
78b6b7ce
LP
4987 systemd-ask-password gained a new --echo option to turn that
4988 on.
b62a309a 4989
e6c253e3
MS
4990 * The default sysctl.d/ snippets will now set:
4991
4992 net.core.default_qdisc = fq_codel
4993
ba8df74b
KS
4994 This selects Fair Queuing Controlled Delay as the default
4995 queuing discipline for network interfaces. fq_codel helps
e6c253e3
MS
4996 fight the network bufferbloat problem. It is believed to be
4997 a good default with no tuning required for most workloads.
4998 Downstream distributions may override this choice. On 10Gbit
4999 servers that do not do forwarding, "fq" may perform better.
5000 Systems without a good clocksource should use "pfifo_fast".
5001
4bdc60cb
LP
5002 * If kdbus is enabled during build a new option BusPolicy= is
5003 available for service units, that allows locking all service
5004 processes into a stricter bus policy, in order to limit
5005 access to various bus services, or even hide most of them
5006 from the service's view entirely.
5007
5008 * networkctl will now show the .network and .link file
5009 networkd has applied to a specific interface.
5010
5011 * sd-login gained a new API call sd_session_get_desktop() to
5012 query which desktop environment has been selected for a
5013 session.
5014
5015 * UNIX utmp support is now compile-time optional to support
5016 legacy-free systems.
5017
78b6b7ce
LP
5018 * systemctl gained two new commands "add-wants" and
5019 "add-requires" for pulling in units from specific targets
5020 easily.
5021
5022 * If the word "rescue" is specified on the kernel command line
5023 the system will now boot into rescue mode (aka
5024 rescue.target), which was previously available only by
5025 specifying "1" or "systemd.unit=rescue.target" on the kernel
5026 command line. This new kernel command line option nicely
5027 mirrors the already existing "emergency" kernel command line
5028 option.
5029
5030 * New kernel command line options mount.usr=, mount.usrflags=,
d4474c41 5031 mount.usrfstype= have been added that match root=, rootflags=,
78b6b7ce
LP
5032 rootfstype= but allow mounting a specific file system to
5033 /usr.
5034
f6d1de85 5035 * The $NOTIFY_SOCKET is now also passed to control processes of
78b6b7ce
LP
5036 services, not only the main process.
5037
5038 * This version reenables support for fsck's -l switch. This
5039 means at least version v2.25 of util-linux is required for
5040 operation, otherwise dead-locks on device nodes may
5041 occur. Again: you need to update util-linux to at least
5042 v2.25 when updating systemd to v217.
5043
3769415e
TT
5044 * The "multi-seat-x" tool has been removed from systemd, as
5045 its functionality has been integrated into X servers 1.16,
5046 and the tool is hence redundant. It is recommended to update
5047 display managers invoking this tool to simply invoke X
5048 directly from now on, again.
5049
fae9332b 5050 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
d35f51ea
ZJS
5051 message flag has been added for all of systemd's polkit
5052 authenticated method calls has been added. In particular this
5053 now allows optional interactive authorization via polkit for
5054 many of PID1's privileged operations such as unit file
5055 enabling and disabling.
fae9332b 5056
cfa1571b
LP
5057 * "udevadm hwdb --update" learnt a new switch "--usr" for
5058 placing the rebuilt hardware database in /usr instead of
5059 /etc. When used only hardware database entries stored in
5060 /usr will be used, and any user database entries in /etc are
5061 ignored. This functionality is useful for vendors to ship a
5062 pre-built database on systems where local configuration is
5063 unnecessary or unlikely.
5064
7e63dd10
LP
5065 * Calendar time specifications in .timer units now also
5066 understand the strings "semi-annually", "quarterly" and
ba8df74b 5067 "minutely" as shortcuts (in addition to the preexisting
7e63dd10
LP
5068 "anually", "hourly", ...).
5069
d4474c41
TG
5070 * systemd-tmpfiles will now correctly create files in /dev
5071 at boot which are marked for creation only at boot. It is
5072 recommended to always create static device nodes with 'c!'
5073 and 'b!', so that they are created only at boot and not
5074 overwritten at runtime.
5075
3b187c5c
LP
5076 * When the watchdog logic is used for a service (WatchdogSec=)
5077 and the watchdog timeout is hit the service will now be
5078 terminated with SIGABRT (instead of just SIGTERM), in order
5079 to make sure a proper coredump and backtrace is
5080 generated. This ensures that hanging services will result in
5081 similar coredump/backtrace behaviour as services that hit a
5082 segmentation fault.
5083
4b08dd87
LP
5084 Contributions from: Andreas Henriksson, Andrei Borzenkov,
5085 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
5086 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
5087 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
5088 Herrmann, David Sommerseth, David Strauss, Emil Renner
5089 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
5090 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
5091 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
5092 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
5093 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
5094 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
5095 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
5096 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
5097 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
5098 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
5099 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
5100 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
5101 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
5102 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
5103 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
5104 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
13e92f39 5105 Jędrzejewski-Szmek
4b08dd87 5106
ccddd104 5107 — Berlin, 2014-10-28
4b08dd87 5108
b72ddf0f 5109CHANGES WITH 216:
b2ca0d63
LP
5110
5111 * timedated no longer reads NTP implementation unit names from
b72ddf0f 5112 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
b2ca0d63
LP
5113 implementations should add a
5114
b72ddf0f 5115 Conflicts=systemd-timesyncd.service
b2ca0d63
LP
5116
5117 to their unit files to take over and replace systemd's NTP
5118 default functionality.
5119
5120 * systemd-sysusers gained a new line type "r" for configuring
5121 which UID/GID ranges to allocate system users/groups
5122 from. Lines of type "u" may now add an additional column
5123 that specifies the home directory for the system user to be
5124 created. Also, systemd-sysusers may now optionally read user
5125 information from STDIN instead of a file. This is useful for
5126 invoking it from RPM preinst scriptlets that need to create
5127 users before the first RPM file is installed since these
5128 files might need to be owned by them. A new
5129 %sysusers_create_inline RPM macro has been introduced to do
5130 just that. systemd-sysusers now updates the shadow files as
5131 well as the user/group databases, which should enhance
5132 compatibility with certain tools like grpck.
5133
d35f51ea
ZJS
5134 * A number of bus APIs of PID 1 now optionally consult polkit to
5135 permit access for otherwise unprivileged clients under certain
5136 conditions. Note that this currently doesn't support
5137 interactive authentication yet, but this is expected to be
5138 added eventually, too.
b2ca0d63
LP
5139
5140 * /etc/machine-info now has new fields for configuring the
5141 deployment environment of the machine, as well as the
5142 location of the machine. hostnamectl has been updated with
5143 new command to update these fields.
5144
5145 * systemd-timesyncd has been updated to automatically acquire
5146 NTP server information from systemd-networkd, which might
5147 have been discovered via DHCP.
5148
5149 * systemd-resolved now includes a caching DNS stub resolver
5150 and a complete LLMNR name resolution implementation. A new
daa05349
AB
5151 NSS module "nss-resolve" has been added which can be used
5152 instead of glibc's own "nss-dns" to resolve hostnames via
b2ca0d63
LP
5153 systemd-resolved. Hostnames, addresses and arbitrary RRs may
5154 be resolved via systemd-resolved D-Bus APIs. In contrast to
5155 the glibc internal resolver systemd-resolved is aware of
5156 multi-homed system, and keeps DNS server and caches separate
5f02e26c 5157 and per-interface. Queries are sent simultaneously on all
b2ca0d63
LP
5158 interfaces that have DNS servers configured, in order to
5159 properly handle VPNs and local LANs which might resolve
5160 separate sets of domain names. systemd-resolved may acquire
a1a4a25e 5161 DNS server information from systemd-networkd automatically,
b2ca0d63
LP
5162 which in turn might have discovered them via DHCP. A tool
5163 "systemd-resolve-host" has been added that may be used to
5164 query the DNS logic in resolved. systemd-resolved implements
5165 IDNA and automatically uses IDNA or UTF-8 encoding depending
5166 on whether classic DNS or LLMNR is used as transport. In the
5167 next releases we intend to add a DNSSEC and mDNS/DNS-SD
5168 implementation to systemd-resolved.
5169
5170 * A new NSS module nss-mymachines has been added, that
5171 automatically resolves the names of all local registered
5172 containers to their respective IP addresses.
5173
5174 * A new client tool "networkctl" for systemd-networkd has been
5175 added. It currently is entirely passive and will query
5176 networking configuration from udev, rtnetlink and networkd,
5f02e26c 5177 and present it to the user in a very friendly
b2ca0d63
LP
5178 way. Eventually, we hope to extend it to become a full
5179 control utility for networkd.
5180
5181 * .socket units gained a new DeferAcceptSec= setting that
5182 controls the kernels' TCP_DEFER_ACCEPT sockopt for
a8eaaee7 5183 TCP. Similarly, support for controlling TCP keep-alive
b2ca0d63
LP
5184 settings has been added (KeepAliveTimeSec=,
5185 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
5186 turning off Nagle's algorithm on TCP has been added
5187 (NoDelay=).
5188
a1a4a25e 5189 * logind learned a new session type "web", for use in projects
b2ca0d63
LP
5190 like Cockpit which register web clients as PAM sessions.
5191
5192 * timer units with at least one OnCalendar= setting will now
46ae28d8 5193 be started only after time-sync.target has been
b2ca0d63
LP
5194 reached. This way they will not elapse before the system
5195 clock has been corrected by a local NTP client or
5196 similar. This is particular useful on RTC-less embedded
5197 machines, that come up with an invalid system clock.
5198
5199 * systemd-nspawn's --network-veth= switch should now result in
5200 stable MAC addresses for both the outer and the inner side
5201 of the link.
5202
5203 * systemd-nspawn gained a new --volatile= switch for running
5204 container instances with /etc or /var unpopulated.
5205
5206 * The kdbus client code has been updated to use the new Linux
5207 3.17 memfd subsystem instead of the old kdbus-specific one.
5208
5209 * systemd-networkd's DHCP client and server now support
01da80b1
LP
5210 FORCERENEW. There are also new configuration options to
5211 configure the vendor client identifier and broadcast mode
5212 for DHCP.
b2ca0d63
LP
5213
5214 * systemd will no longer inform the kernel about the current
5215 timezone, as this is necessarily incorrect and racy as the
5216 kernel has no understanding of DST and similar
5217 concepts. This hence means FAT timestamps will be always
5218 considered UTC, similar to what Android is already
5219 doing. Also, when the RTC is configured to the local time
5220 (rather than UTC) systemd will never synchronize back to it,
5221 as this might confuse Windows at a later boot.
5222
5223 * systemd-analyze gained a new command "verify" for offline
5224 validation of unit files.
5225
5226 * systemd-networkd gained support for a couple of additional
5227 settings for bonding networking setups. Also, the metric for
5228 statically configured routes may now be configured. For
5229 network interfaces where this is appropriate the peer IP
5230 address may now be configured.
5231
26568403
TG
5232 * systemd-networkd's DHCP client will no longer request
5233 broadcasting by default, as this tripped up some networks.
5234 For hardware where broadcast is required the feature should
5235 be switched back on using RequestBroadcast=yes.
5236
5237 * systemd-networkd will now set up IPv4LL addresses (when
5238 enabled) even if DHCP is configured successfully.
5239
5240 * udev will now default to respect network device names given
5241 by the kernel when the kernel indicates that these are
5242 predictable. This behavior can be tweaked by changing
5243 NamePolicy= in the relevant .link file.
5244
b2ca0d63
LP
5245 * A new library systemd-terminal has been added that
5246 implements full TTY stream parsing and rendering. This
5247 library is supposed to be used later on for implementing a
5248 full userspace VT subsystem, replacing the current kernel
5249 implementation.
5250
5251 * A new tool systemd-journal-upload has been added to push
5252 journal data to a remote system running
5253 systemd-journal-remote.
5254
5255 * journald will no longer forward all local data to another
5256 running syslog daemon. This change has been made because
5257 rsyslog (which appears to be the most commonly used syslog
5258 implementation these days) no longer makes use of this, and
5259 instead pulls the data out of the journal on its own. Since
5f02e26c 5260 forwarding the messages to a non-existent syslog server is
b2ca0d63
LP
5261 more expensive than we assumed we have now turned this
5262 off. If you run a syslog server that is not a recent rsyslog
5263 version, you have to turn this option on again
5264 (ForwardToSyslog= in journald.conf).
5265
5266 * journald now optionally supports the LZ4 compressor for
5267 larger journal fields. This compressor should perform much
5268 better than XZ which was the previous default.
5269
5270 * machinectl now shows the IP addresses of local containers,
5271 if it knows them, plus the interface name of the container.
5272
5273 * A new tool "systemd-escape" has been added that makes it
5274 easy to escape strings to build unit names and similar.
5275
5276 * sd_notify() messages may now include a new ERRNO= field
5277 which is parsed and collected by systemd and shown among the
5278 "systemctl status" output for a service.
5279
5280 * A new component "systemd-firstboot" has been added that
5281 queries the most basic systemd information (timezone,
a1a4a25e 5282 hostname, root password) interactively on first
b2ca0d63
LP
5283 boot. Alternatively it may also be used to provision these
5284 things offline on OS images installed into directories.
5285
01da80b1
LP
5286 * The default sysctl.d/ snippets will now set
5287
5288 net.ipv4.conf.default.promote_secondaries=1
5289
5290 This has the benefit of no flushing secondary IP addresses
5291 when primary addresses are removed.
5292
b2ca0d63
LP
5293 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
5294 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
5295 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
5296 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
5297 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
5298 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
5299 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
5300 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
5301 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
5302 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
5303 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
5304 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
5305 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
5306 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
5307 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
5308
ccddd104 5309 — Berlin, 2014-08-19
b72ddf0f 5310
3dff3e00 5311CHANGES WITH 215:
24a2bf4c
LP
5312
5313 * A new tool systemd-sysusers has been added. This tool
5314 creates system users and groups in /etc/passwd and
5315 /etc/group, based on static declarative system user/group
5316 definitions in /usr/lib/sysusers.d/. This is useful to
5317 enable factory resets and volatile systems that boot up with
5318 an empty /etc directory, and thus need system users and
5319 groups created during early boot. systemd now also ships
5320 with two default sysusers.d/ files for the most basic
5321 users and groups systemd and the core operating system
5322 require.
5323
5324 * A new tmpfiles snippet has been added that rebuilds the
5325 essential files in /etc on boot, should they be missing.
5326
5327 * A directive for ensuring automatic clean-up of
5328 /var/cache/man/ has been removed from the default
5329 configuration. This line should now be shipped by the man
5330 implementation. The necessary change has been made to the
5331 man-db implementation. Note that you need to update your man
5332 implementation to one that ships this line, otherwise no
5333 automatic clean-up of /var/cache/man will take place.
5334
5335 * A new condition ConditionNeedsUpdate= has been added that
5336 may conditionalize services to only run when /etc or /var
5337 are "older" than the vendor operating system resources in
5338 /usr. This is useful for reconstructing or updating /etc
5339 after an offline update of /usr or a factory reset, on the
5340 next reboot. Services that want to run once after such an
5341 update or reset should use this condition and order
5342 themselves before the new systemd-update-done.service, which
5343 will mark the two directories as fully updated. A number of
5344 service files have been added making use of this, to rebuild
5345 the udev hardware database, the journald message catalog and
5346 dynamic loader cache (ldconfig). The systemd-sysusers tool
5347 described above also makes use of this now. With this in
5348 place it is now possible to start up a minimal operating
ce1dde29 5349 system with /etc empty cleanly. For more information on the
24a2bf4c
LP
5350 concepts involved see this recent blog story:
5351
5352 http://0pointer.de/blog/projects/stateless.html
5353
5354 * A new system group "input" has been introduced, and all
5355 input device nodes get this group assigned. This is useful
5356 for system-level software to get access to input devices. It
3dff3e00
KS
5357 complements what is already done for "audio" and "video".
5358
24a2bf4c
LP
5359 * systemd-networkd learnt minimal DHCPv4 server support in
5360 addition to the existing DHCPv4 client support. It also
5361 learnt DHCPv6 client and IPv6 Router Solicitation client
5362 support. The DHCPv4 client gained support for static routes
5363 passed in from the server. Note that the [DHCPv4] section
5364 known in older systemd-networkd versions has been renamed to
5365 [DHCP] and is now also used by the DHCPv6 client. Existing
c7435cc9
LP
5366 .network files using settings of this section should be
5367 updated, though compatibility is maintained. Optionally, the
5368 client hostname may now be sent to the DHCP server.
24a2bf4c 5369
c7435cc9
LP
5370 * networkd gained support for vxlan virtual networks as well
5371 as tun/tap and dummy devices.
24a2bf4c
LP
5372
5373 * networkd gained support for automatic allocation of address
5374 ranges for interfaces from a system-wide pool of
5375 addresses. This is useful for dynamically managing a large
5376 number of interfaces with a single network configuration
5377 file. In particular this is useful to easily assign
5378 appropriate IP addresses to the veth links of a large number
5379 of nspawn instances.
5380
5381 * RPM macros for processing sysusers, sysctl and binfmt
5382 drop-in snippets at package installation time have been
5383 added.
5384
5385 * The /etc/os-release file should now be placed in
5386 /usr/lib/os-release. The old location is automatically
5387 created as symlink. /usr/lib is the more appropriate
5388 location of this file, since it shall actually describe the
5389 vendor operating system shipped in /usr, and not the
5390 configuration stored in /etc.
5391
5392 * .mount units gained a new boolean SloppyOptions= setting
5393 that maps to mount(8)'s -s option which enables permissive
5394 parsing of unknown mount options.
5395
5396 * tmpfiles learnt a new "L+" directive which creates a symlink
5397 but (unlike "L") deletes a pre-existing file first, should
5398 it already exist and not already be the correct
a8eaaee7 5399 symlink. Similarly, "b+", "c+" and "p+" directives have been
24a2bf4c
LP
5400 added as well, which create block and character devices, as
5401 well as fifos in the filesystem, possibly removing any
5402 pre-existing files of different types.
5403
5404 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
5405 'argument' field (which so far specified the source to
ce1dde29 5406 symlink/copy the files from) is now optional. If omitted the
24a2bf4c
LP
5407 same file os copied from /usr/share/factory/ suffixed by the
5408 full destination path. This is useful for populating /etc
5409 with essential files, by copying them from vendor defaults
5410 shipped in /usr/share/factory/etc.
5411
5412 * A new command "systemctl preset-all" has been added that
5413 applies the service preset settings to all installed unit
5414 files. A new switch --preset-mode= has been added that
5415 controls whether only enable or only disable operations
5416 shall be executed.
5417
5418 * A new command "systemctl is-system-running" has been added
5419 that allows checking the overall state of the system, for
ce1dde29 5420 example whether it is fully up and running.
24a2bf4c
LP
5421
5422 * When the system boots up with an empty /etc, the equivalent
5423 to "systemctl preset-all" is executed during early boot, to
5424 make sure all default services are enabled after a factory
5425 reset.
5426
5427 * systemd now contains a minimal preset file that enables the
5428 most basic services systemd ships by default.
5429
5430 * Unit files' [Install] section gained a new DefaultInstance=
5431 field for defining the default instance to create if a
5432 template unit is enabled with no instance specified.
5433
5434 * A new passive target cryptsetup-pre.target has been added
5435 that may be used by services that need to make they run and
5436 finish before the first LUKS cryptographic device is set up.
5437
5438 * The /dev/loop-control and /dev/btrfs-control device nodes
5439 are now owned by the "disk" group by default, opening up
5440 access to this group.
5441
5442 * systemd-coredump will now automatically generate a
5443 stack trace of all core dumps taking place on the system,
5444 based on elfutils' libdw library. This stack trace is logged
5445 to the journal.
5446
5447 * systemd-coredump may now optionally store coredumps directly
5448 on disk (in /var/lib/systemd/coredump, possibly compressed),
5449 instead of storing them unconditionally in the journal. This
5450 mode is the new default. A new configuration file
5451 /etc/systemd/coredump.conf has been added to configure this
5452 and other parameters of systemd-coredump.
5453
5454 * coredumpctl gained a new "info" verb to show details about a
5455 specific coredump. A new switch "-1" has also been added
5456 that makes sure to only show information about the most
5457 recent entry instead of all entries. Also, as the tool is
5458 generally useful now the "systemd-" prefix of the binary
5459 name has been removed. Distributions that want to maintain
5460 compatibility with the old name should add a symlink from
5461 the old name to the new name.
5462
5463 * journald's SplitMode= now defaults to "uid". This makes sure
ce1dde29 5464 that unprivileged users can access their own coredumps with
24a2bf4c
LP
5465 coredumpctl without restrictions.
5466
5467 * New kernel command line options "systemd.wants=" (for
5468 pulling an additional unit during boot), "systemd.mask="
5469 (for masking a specific unit for the boot), and
5470 "systemd.debug-shell" (for enabling the debug shell on tty9)
5471 have been added. This is implemented in the new generator
5472 "systemd-debug-generator".
5473
5474 * systemd-nspawn will now by default filter a couple of
5475 syscalls for containers, among them those required for
5476 kernel module loading, direct x86 IO port access, swap
5477 management, and kexec. Most importantly though
5478 open_by_handle_at() is now prohibited for containers,
5479 closing a hole similar to a recently discussed vulnerability
5480 in docker regarding access to files on file hierarchies the
b938cb90
JE
5481 container should normally not have access to. Note that, for
5482 nspawn, we generally make no security claims anyway (and
24a2bf4c
LP
5483 this is explicitly documented in the man page), so this is
5484 just a fix for one of the most obvious problems.
5485
5486 * A new man page file-hierarchy(7) has been added that
5487 contains a minimized, modernized version of the file system
5488 layout systemd expects, similar in style to the FHS
c7435cc9
LP
5489 specification or hier(5). A new tool systemd-path(1) has
5490 been added to query many of these paths for the local
5491 machine and user.
24a2bf4c
LP
5492
5493 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
5494 longer done. Since the directory now has a per-user size
5495 limit, and is cleaned on logout this appears unnecessary,
5496 in particular since this now brings the lifecycle of this
5497 directory closer in line with how IPC objects are handled.
5498
5499 * systemd.pc now exports a number of additional directories,
5500 including $libdir (which is useful to identify the library
5501 path for the primary architecture of the system), and a
5502 couple of drop-in directories.
5503
3058e017
TLSC
5504 * udev's predictable network interface names now use the dev_port
5505 sysfs attribute, introduced in linux 3.15 instead of dev_id to
5506 distinguish between ports of the same PCI function. dev_id should
5507 only be used for ports using the same HW address, hence the need
5508 for dev_port.
5509
c7435cc9
LP
5510 * machined has been updated to export the OS version of a
5511 container (read from /etc/os-release and
5512 /usr/lib/os-release) on the bus. This is now shown in
5513 "machinectl status" for a machine.
5514
5515 * A new service setting RestartForceExitStatus= has been
5516 added. If configured to a set of exit signals or process
5517 return values, the service will be restarted when the main
5518 daemon process exits with any of them, regardless of the
5519 Restart= setting.
5520
5521 * systemctl's -H switch for connecting to remote systemd
5522 machines has been extended so that it may be used to
5523 directly connect to a specific container on the
5524 host. "systemctl -H root@foobar:waldi" will now connect as
5525 user "root" to host "foobar", and then proceed directly to
5526 the container named "waldi". Note that currently you have to
5527 authenticate as user "root" for this to work, as entering
5528 containers is a privileged operation.
5529
5530 Contributions from: Andreas Henriksson, Benjamin Steinwender,
5531 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
5532 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
5533 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
5534 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
5535 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
5536 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
5537 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
5538 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
5539 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
5540 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
5541 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
5542
ccddd104 5543 — Berlin, 2014-07-03
c7435cc9 5544
4196a3ea
KS
5545CHANGES WITH 214:
5546
5547 * As an experimental feature, udev now tries to lock the
5548 disk device node (flock(LOCK_SH|LOCK_NB)) while it
5549 executes events for the disk or any of its partitions.
5550 Applications like partitioning programs can lock the
5551 disk device node (flock(LOCK_EX)) and claim temporary
5552 device ownership that way; udev will entirely skip all event
5553 handling for this disk and its partitions. If the disk
5554 was opened for writing, the close will trigger a partition
5555 table rescan in udev's "watch" facility, and if needed
71449caf 5556 synthesize "change" events for the disk and all its partitions.
8d0e0ddd 5557 This is now unconditionally enabled, and if it turns out to
4196a3ea 5558 cause major problems, we might turn it on only for specific
45df8656 5559 devices, or might need to disable it entirely. Device Mapper
4196a3ea
KS
5560 devices are excluded from this logic.
5561
04e91da2
LP
5562 * We temporarily dropped the "-l" switch for fsck invocations,
5563 since they collide with the flock() logic above. util-linux
5564 upstream has been changed already to avoid this conflict,
5565 and we will readd "-l" as soon as util-linux with this
5566 change has been released.
5567
5568 * The dependency on libattr has been removed. Since a long
8d0e0ddd 5569 time, the extended attribute calls have moved to glibc, and
04e91da2
LP
5570 libattr is thus unnecessary.
5571
ce830873 5572 * Virtualization detection works without privileges now. This
04e91da2
LP
5573 means the systemd-detect-virt binary no longer requires
5574 CAP_SYS_PTRACE file capabilities, and our daemons can run
71449caf 5575 with fewer privileges.
04e91da2
LP
5576
5577 * systemd-networkd now runs under its own "systemd-network"
5578 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
5579 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
5580 loses the ability to write to files owned by root this way.
5581
a8eaaee7 5582 * Similarly, systemd-resolved now runs under its own
04e91da2
LP
5583 "systemd-resolve" user with no capabilities remaining.
5584
a8eaaee7 5585 * Similarly, systemd-bus-proxyd now runs under its own
04e91da2
LP
5586 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
5587
5588 * systemd-networkd gained support for setting up "veth"
a8eaaee7 5589 virtual Ethernet devices for container connectivity, as well
04e91da2
LP
5590 as GRE and VTI tunnels.
5591
5592 * systemd-networkd will no longer automatically attempt to
5593 manually load kernel modules necessary for certain tunnel
8d0e0ddd 5594 transports. Instead, it is assumed the kernel loads them
04e91da2
LP
5595 automatically when required. This only works correctly on
5596 very new kernels. On older kernels, please consider adding
c54bed5d 5597 the kernel modules to /etc/modules-load.d/ as a work-around.
04e91da2 5598
cd14eda3 5599 * The resolv.conf file systemd-resolved generates has been
8d0e0ddd
JE
5600 moved to /run/systemd/resolve/. If you have a symlink from
5601 /etc/resolv.conf, it might be necessary to correct it.
cd14eda3 5602
ef392da6 5603 * Two new service settings, ProtectHome= and ProtectSystem=,
8d0e0ddd 5604 have been added. When enabled, they will make the user data
04e91da2
LP
5605 (such as /home) inaccessible or read-only and the system
5606 (such as /usr) read-only, for specific services. This allows
5607 very light-weight per-service sandboxing to avoid
5608 modifications of user data or system files from
5609 services. These two new switches have been enabled for all
5610 of systemd's long-running services, where appropriate.
5611
5612 * Socket units gained new SocketUser= and SocketGroup=
5613 settings to set the owner user and group of AF_UNIX sockets
5614 and FIFOs in the file system.
5615
8d0e0ddd 5616 * Socket units gained a new RemoveOnStop= setting. If enabled,
04e91da2
LP
5617 all FIFOS and sockets in the file system will be removed
5618 when the specific socket unit is stopped.
5619
5620 * Socket units gained a new Symlinks= setting. It takes a list
5621 of symlinks to create to file system sockets or FIFOs
45df8656 5622 created by the specific Unix sockets. This is useful to
71449caf 5623 manage symlinks to socket nodes with the same life-cycle as
04e91da2
LP
5624 the socket itself.
5625
5626 * The /dev/log socket and /dev/initctl FIFO have been moved to
5627 /run, and have been replaced by symlinks. This allows
5628 connecting to these facilities even if PrivateDevices=yes is
5629 used for a service (which makes /dev/log itself unavailable,
5630 but /run is left). This also has the benefit of ensuring
5631 that /dev only contains device nodes, directories and
5632 symlinks, and nothing else.
5633
5634 * sd-daemon gained two new calls sd_pid_notify() and
5635 sd_pid_notifyf(). They are similar to sd_notify() and
5636 sd_notifyf(), but allow overriding of the source PID of
5637 notification messages if permissions permit this. This is
5638 useful to send notify messages on behalf of a different
5639 process (for example, the parent process). The
5640 systemd-notify tool has been updated to make use of this
5641 when sending messages (so that notification messages now
5642 originate from the shell script invoking systemd-notify and
5643 not the systemd-notify process itself. This should minimize
5644 a race where systemd fails to associate notification
5645 messages to services when the originating process already
5646 vanished.
5647
5648 * A new "on-abnormal" setting for Restart= has been added. If
8d0e0ddd 5649 set, it will result in automatic restarts on all "abnormal"
04e91da2
LP
5650 reasons for a process to exit, which includes unclean
5651 signals, core dumps, timeouts and watchdog timeouts, but
5652 does not include clean and unclean exit codes or clean
5653 signals. Restart=on-abnormal is an alternative for
5654 Restart=on-failure for services that shall be able to
5655 terminate and avoid restarts on certain errors, by
5656 indicating so with an unclean exit code. Restart=on-failure
5657 or Restart=on-abnormal is now the recommended setting for
5658 all long-running services.
5659
5660 * If the InaccessibleDirectories= service setting points to a
5661 mount point (or if there are any submounts contained within
5662 it), it is now attempted to completely unmount it, to make
5663 the file systems truly unavailable for the respective
5664 service.
5665
5666 * The ReadOnlyDirectories= service setting and
5667 systemd-nspawn's --read-only parameter are now recursively
5668 applied to all submounts, too.
5669
5670 * Mount units may now be created transiently via the bus APIs.
5671
5672 * The support for SysV and LSB init scripts has been removed
5673 from the systemd daemon itself. Instead, it is now
5674 implemented as a generator that creates native systemd units
5675 from these scripts when needed. This enables us to remove a
5676 substantial amount of legacy code from PID 1, following the
5677 fact that many distributions only ship a very small number
5678 of LSB/SysV init scripts nowadays.
5679
cc98b302 5680 * Privileged Xen (dom0) domains are not considered
04e91da2
LP
5681 virtualization anymore by the virtualization detection
5682 logic. After all, they generally have unrestricted access to
71449caf 5683 the hardware and usually are used to manage the unprivileged
04e91da2
LP
5684 (domU) domains.
5685
5686 * systemd-tmpfiles gained a new "C" line type, for copying
5687 files or entire directories.
5688
5689 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
8d0e0ddd
JE
5690 lines. So far, they have been non-globbing versions of the
5691 latter, and have thus been redundant. In future, it is
5692 recommended to only use "z". "m" has hence been removed
04e91da2
LP
5693 from the documentation, even though it stays supported.
5694
5695 * A tmpfiles snippet to recreate the most basic structure in
5696 /var has been added. This is enough to create the /var/run →
5697 /run symlink and create a couple of structural
5698 directories. This allows systems to boot up with an empty or
8d0e0ddd
JE
5699 volatile /var. Of course, while with this change, the core OS
5700 now is capable with dealing with a volatile /var, not all
04e91da2 5701 user services are ready for it. However, we hope that sooner
8d0e0ddd 5702 or later, many service daemons will be changed upstream so
04e91da2
LP
5703 that they are able to automatically create their necessary
5704 directories in /var at boot, should they be missing. This is
5705 the first step to allow state-less systems that only require
5706 the vendor image for /usr to boot.
5707
5708 * systemd-nspawn has gained a new --tmpfs= switch to mount an
5709 empty tmpfs instance to a specific directory. This is
5710 particularly useful for making use of the automatic
5711 reconstruction of /var (see above), by passing --tmpfs=/var.
5712
5713 * Access modes specified in tmpfiles snippets may now be
5714 prefixed with "~", which indicates that they shall be masked
daa05349 5715 by whether the existing file or directory is currently
8d0e0ddd 5716 writable, readable or executable at all. Also, if specified,
04e91da2
LP
5717 the sgid/suid/sticky bits will be masked for all
5718 non-directories.
5719
5720 * A new passive target unit "network-pre.target" has been
5721 added which is useful for services that shall run before any
5722 network is configured, for example firewall scripts.
5723
4c0d13bd
LP
5724 * The "floppy" group that previously owned the /dev/fd*
5725 devices is no longer used. The "disk" group is now used
5726 instead. Distributions should probably deprecate usage of
5727 this group.
5728
dc1d6c02
LP
5729 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
5730 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
5731 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
5732 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
5733 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
5734 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
5735 Jędrzejewski-Szmek
5736
ccddd104 5737 — Berlin, 2014-06-11
dc1d6c02 5738
6936cd89
LP
5739CHANGES WITH 213:
5740
5741 * A new "systemd-timesyncd" daemon has been added for
69beda1f 5742 synchronizing the system clock across the network. It
6936cd89 5743 implements an SNTP client. In contrast to NTP
8d0e0ddd 5744 implementations such as chrony or the NTP reference server,
6936cd89 5745 this only implements a client side, and does not bother with
c9679c65
LP
5746 the full NTP complexity, focusing only on querying time from
5747 one remote server and synchronizing the local clock to
6936cd89 5748 it. Unless you intend to serve NTP to networked clients or
8d0e0ddd 5749 want to connect to local hardware clocks, this simple NTP
6936cd89
LP
5750 client should be more than appropriate for most
5751 installations. The daemon runs with minimal privileges, and
5752 has been hooked up with networkd to only operate when
5753 network connectivity is available. The daemon saves the
5754 current clock to disk every time a new NTP sync has been
5755 acquired, and uses this to possibly correct the system clock
69beda1f 5756 early at bootup, in order to accommodate for systems that
6936cd89 5757 lack an RTC such as the Raspberry Pi and embedded devices,
8d0e0ddd 5758 and to make sure that time monotonically progresses on these
c9679c65 5759 systems, even if it is not always correct. To make use of
8d0e0ddd 5760 this daemon, a new system user and group "systemd-timesync"
c9679c65 5761 needs to be created on installation of systemd.
6936cd89 5762
69beda1f
KS
5763 * The queue "seqnum" interface of libudev has been disabled, as
5764 it was generally incompatible with device namespacing as
6936cd89
LP
5765 sequence numbers of devices go "missing" if the devices are
5766 part of a different namespace.
5767
5768 * "systemctl list-timers" and "systemctl list-sockets" gained
5769 a --recursive switch for showing units of these types also
499b604b
ZJS
5770 for all local containers, similar in style to the already
5771 supported --recursive switch for "systemctl list-units".
6936cd89
LP
5772
5773 * A new RebootArgument= setting has been added for service
5774 units, which may be used to specify a kernel reboot argument
499b604b 5775 to use when triggering reboots with StartLimitAction=.
6936cd89
LP
5776
5777 * A new FailureAction= setting has been added for service
5778 units which may be used to specify an operation to trigger
499b604b 5779 when a service fails. This works similarly to
8d0e0ddd 5780 StartLimitAction=, but unlike it, controls what is done
6936cd89
LP
5781 immediately rather than only after several attempts to
5782 restart the service in question.
5783
5784 * hostnamed got updated to also expose the kernel name,
499b604b
ZJS
5785 release, and version on the bus. This is useful for
5786 executing commands like hostnamectl with the -H switch.
5787 systemd-analyze makes use of this to properly display
5788 details when running non-locally.
6936cd89
LP
5789
5790 * The bootchart tool can now show cgroup information in the
5791 graphs it generates.
5792
5793 * The CFS CPU quota cgroup attribute is now exposed for
5794 services. The new CPUQuota= switch has been added for this
5795 which takes a percentage value. Setting this will have the
5796 result that a service may never get more CPU time than the
5797 specified percentage, even if the machine is otherwise idle.
5798
5799 * systemd-networkd learned IPIP and SIT tunnel support.
5800
5801 * LSB init scripts exposing a dependency on $network will now
5802 get a dependency on network-online.target rather than simply
5803 network.target. This should bring LSB handling closer to
5804 what it was on SysV systems.
5805
5806 * A new fsck.repair= kernel option has been added to control
5807 how fsck shall deal with unclean file systems at boot.
5808
5809 * The (.ini) configuration file parser will now silently
5810 ignore sections whose name begins with "X-". This may be
5811 used to maintain application-specific extension sections in unit
5812 files.
5813
5814 * machined gained a new API to query the IP addresses of
5815 registered containers. "machinectl status" has been updated
5816 to show these addresses in its output.
5817
5818 * A new call sd_uid_get_display() has been added to the
5819 sd-login APIs for querying the "primary" session of a
5820 user. The "primary" session of the user is elected from the
5821 user's sessions and generally a graphical session is
5822 preferred over a text one.
5823
5824 * A minimal systemd-resolved daemon has been added. It
5825 currently simply acts as a companion to systemd-networkd and
5826 manages resolv.conf based on per-interface DNS
5827 configuration, possibly supplied via DHCP. In the long run
5828 we hope to extend this into a local DNSSEC enabled DNS and
5829 mDNS cache.
5830
68dd0956
TG
5831 * The systemd-networkd-wait-online tool is now enabled by
5832 default. It will delay network-online.target until a network
5833 connection has been configured. The tool primarily integrates
5834 with networkd, but will also make a best effort to make sense
5835 of network configuration performed in some other way.
5836
6936cd89 5837 * Two new service options StartupCPUShares= and
499b604b 5838 StartupBlockIOWeight= have been added that work similarly to
6936cd89 5839 CPUShares= and BlockIOWeight= however only apply during
69beda1f 5840 system startup. This is useful to prioritize certain services
6936cd89
LP
5841 differently during bootup than during normal runtime.
5842
8e7acf67
LP
5843 * hostnamed has been changed to prefer the statically
5844 configured hostname in /etc/hostname (unless set to
5845 'localhost' or empty) over any dynamic one supplied by
8d0e0ddd 5846 dhcp. With this change, the rules for picking the hostname
8e7acf67
LP
5847 match more closely the rules of other configuration settings
5848 where the local administrator's configuration in /etc always
5849 overrides any other settings.
5850
5851 Contributions fron: Ali H. Caliskan, Alison Chaiken, Bas van
6936cd89
LP
5852 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
5853 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
5854 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
5855 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
5856 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
5857 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
5858 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
5859 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
8e7acf67
LP
5860 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
5861 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
5862 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
5863 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
5864 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
5865 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
5866 Lindskog, WaLyong Cho, Will Woods, Zbigniew
6936cd89
LP
5867 Jędrzejewski-Szmek
5868
ccddd104 5869 — Beijing, 2014-05-28
6936cd89 5870
51c61cda
LP
5871CHANGES WITH 212:
5872
5873 * When restoring the screen brightness at boot, stay away from
5874 the darkest setting or from the lowest 5% of the available
5875 range, depending on which is the larger value of both. This
5876 should effectively protect the user from rebooting into a
5877 black screen, should the brightness have been set to minimum
5878 by accident.
5879
5880 * sd-login gained a new sd_machine_get_class() call to
5881 determine the class ("vm" or "container") of a machine
5882 registered with machined.
5883
5884 * sd-login gained new calls
5885 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
5886 to query the identity of the peer of a local AF_UNIX
499b604b 5887 connection. They operate similarly to their sd_pid_get_xyz()
51c61cda
LP
5888 counterparts.
5889
5890 * PID 1 will now maintain a system-wide system state engine
5891 with the states "starting", "running", "degraded",
5892 "maintenance", "stopping". These states are bound to system
5893 startup, normal runtime, runtime with at least one failed
5894 service, rescue/emergency mode and system shutdown. This
5895 state is shown in the "systemctl status" output when no unit
5896 name is passed. It is useful to determine system state, in
5897 particularly when doing so for many systems or containers at
5898 once.
5899
5900 * A new command "list-machines" has been added to "systemctl"
5901 that lists all local OS containers and shows their system
5902 state (see above), if systemd runs inside of them.
5903
5904 * systemctl gained a new "-r" switch to recursively enumerate
5905 units on all local containers, when used with the
5906 "list-unit" command (which is the default one that is
5907 executed when no parameters are specified).
5908
5909 * The GPT automatic partition discovery logic will now honour
5910 two GPT partition flags: one may be set on a partition to
5911 cause it to be mounted read-only, and the other may be set
5912 on a partition to ignore it during automatic discovery.
5913
5914 * Two new GPT type UUIDs have been added for automatic root
70a44afe 5915 partition discovery, for 32-bit and 64-bit ARM. This is not
51c61cda
LP
5916 particularly useful for discovering the root directory on
5917 these architectures during bare-metal boots (since UEFI is
5918 not common there), but still very useful to allow booting of
5919 ARM disk images in nspawn with the -i option.
5920
5921 * MAC addresses of interfaces created with nspawn's
5922 --network-interface= switch will now be generated from the
5923 machine name, and thus be stable between multiple invocations
5924 of the container.
5925
5926 * logind will now automatically remove all IPC objects owned
5927 by a user if she or he fully logs out. This makes sure that
5928 users who are logged out cannot continue to consume IPC
5929 resources. This covers SysV memory, semaphores and message
5930 queues as well as POSIX shared memory and message
b8bde116
JE
5931 queues. Traditionally, SysV and POSIX IPC had no life-cycle
5932 limits. With this functionality, that is corrected. This may
5933 be turned off by using the RemoveIPC= switch of logind.conf.
51c61cda
LP
5934
5935 * The systemd-machine-id-setup and tmpfiles tools gained a
5936 --root= switch to operate on a specific root directory,
5937 instead of /.
5938
5939 * journald can now forward logged messages to the TTYs of all
5940 logged in users ("wall"). This is the default for all
5941 emergency messages now.
5942
5943 * A new tool systemd-journal-remote has been added to stream
5944 journal log messages across the network.
5945
5946 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
5947 controller trees are mounted into it. Note that the
5948 directories mounted beneath it are not read-only. This is a
5949 security measure and is particularly useful because glibc
5950 actually includes a search logic to pick any tmpfs it can
5951 find to implement shm_open() if /dev/shm is not available
5952 (which it might very well be in namespaced setups).
5953
5954 * machinectl gained a new "poweroff" command to cleanly power
5955 down a local OS container.
5956
5957 * The PrivateDevices= unit file setting will now also drop the
5958 CAP_MKNOD capability from the capability bound set, and
5959 imply DevicePolicy=closed.
5960
5961 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
5962 comprehensively on all long-running systemd services where
5963 this is appropriate.
5964
5965 * systemd-udevd will now run in a disassociated mount
b8bde116 5966 namespace. To mount directories from udev rules, make sure to
51c61cda
LP
5967 pull in mount units via SYSTEMD_WANTS properties.
5968
5969 * The kdbus support gained support for uploading policy into
5970 the kernel. sd-bus gained support for creating "monitoring"
5971 connections that can eavesdrop into all bus communication
5972 for debugging purposes.
5973
5974 * Timestamps may now be specified in seconds since the UNIX
5975 epoch Jan 1st, 1970 by specifying "@" followed by the value
5976 in seconds.
5977
5978 * Native tcpwrap support in systemd has been removed. tcpwrap
5979 is old code, not really maintained anymore and has serious
5980 shortcomings, and better options such as firewalls
5981 exist. For setups that require tcpwrap usage, please
5982 consider invoking your socket-activated service via tcpd,
5983 like on traditional inetd.
5984
5985 * A new system.conf configuration option
5986 DefaultTimerAccuracySec= has been added that controls the
5987 default AccuracySec= setting of .timer units.
5988
b8bde116 5989 * Timer units gained a new WakeSystem= switch. If enabled,
51c61cda
LP
5990 timers configured this way will cause the system to resume
5991 from system suspend (if the system supports that, which most
5992 do these days).
5993
b8bde116 5994 * Timer units gained a new Persistent= switch. If enabled,
51c61cda
LP
5995 timers configured this way will save to disk when they have
5996 been last triggered. This information is then used on next
5997 reboot to possible execute overdue timer events, that
d28315e4
JE
5998 could not take place because the system was powered off.
5999 This enables simple anacron-like behaviour for timer units.
51c61cda
LP
6000
6001 * systemctl's "list-timers" will now also list the time a
6002 timer unit was last triggered in addition to the next time
6003 it will be triggered.
6004
6005 * systemd-networkd will now assign predictable IPv4LL
6006 addresses to its local interfaces.
6007
6008 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
6009 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
6010 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
6011 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
6012 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
6013 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
6014 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
6015 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
6016 Jędrzejewski-Szmek
6017
ccddd104 6018 — Berlin, 2014-03-25
51c61cda 6019
699b6b34
LP
6020CHANGES WITH 211:
6021
6022 * A new unit file setting RestrictAddressFamilies= has been
6023 added to restrict which socket address families unit
6024 processes gain access to. This takes address family names
6025 like "AF_INET" or "AF_UNIX", and is useful to minimize the
6026 attack surface of services via exotic protocol stacks. This
6027 is built on seccomp system call filters.
6028
6029 * Two new unit file settings RuntimeDirectory= and
6030 RuntimeDirectoryMode= have been added that may be used to
6031 manage a per-daemon runtime directories below /run. This is
6032 an alternative for setting up directory permissions with
6033 tmpfiles snippets, and has the advantage that the runtime
6034 directory's lifetime is bound to the daemon runtime and that
6035 the daemon starts up with an empty directory each time. This
6036 is particularly useful when writing services that drop
f1721625 6037 privileges using the User= or Group= setting.
699b6b34
LP
6038
6039 * The DeviceAllow= unit setting now supports globbing for
6040 matching against device group names.
6041
6042 * The systemd configuration file system.conf gained new
6043 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
6044 DefaultMemoryAccounting= to globally turn on/off accounting
6045 for specific resources (cgroups) for all units. These
22e7062d 6046 settings may still be overridden individually in each unit
699b6b34
LP
6047 though.
6048
6049 * systemd-gpt-auto-generator is now able to discover /srv and
6050 root partitions in addition to /home and swap partitions. It
6051 also supports LUKS-encrypted partitions now. With this in
b8bde116 6052 place, automatic discovery of partitions to mount following
699b6b34 6053 the Discoverable Partitions Specification
56cadcb6 6054 (https://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
699b6b34
LP
6055 is now a lot more complete. This allows booting without
6056 /etc/fstab and without root= on the kernel command line on
b8bde116 6057 systems prepared appropriately.
699b6b34
LP
6058
6059 * systemd-nspawn gained a new --image= switch which allows
6060 booting up disk images and Linux installations on any block
6061 device that follow the Discoverable Partitions Specification
6062 (see above). This means that installations made with
6063 appropriately updated installers may now be started and
6064 deployed using container managers, completely
6065 unmodified. (We hope that libvirt-lxc will add support for
6066 this feature soon, too.)
6067
6068 * systemd-nspawn gained a new --network-macvlan= setting to
6069 set up a private macvlan interface for the
499b604b 6070 container. Similarly, systemd-networkd gained a new
699b6b34
LP
6071 Kind=macvlan setting in .netdev files.
6072
6073 * systemd-networkd now supports configuring local addresses
6074 using IPv4LL.
6075
6076 * A new tool systemd-network-wait-online has been added to
6077 synchronously wait for network connectivity using
6078 systemd-networkd.
6079
6080 * The sd-bus.h bus API gained a new sd_bus_track object for
6081 tracking the life-cycle of bus peers. Note that sd-bus.h is
6082 still not a public API though (unless you specify
6083 --enable-kdbus on the configure command line, which however
6084 voids your warranty and you get no API stability guarantee).
6085
6086 * The $XDG_RUNTIME_DIR runtime directories for each user are
6087 now individual tmpfs instances, which has the benefit of
6088 introducing separate pools for each user, with individual
4ef6e535 6089 size limits, and thus making sure that unprivileged clients
699b6b34
LP
6090 can no longer negatively impact the system or other users by
6091 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
6092 RuntimeDirectorySize= has been introduced that allows
6093 controlling the default size limit for all users. It
6094 defaults to 10% of the available physical memory. This is no
6095 replacement for quotas on tmpfs though (which the kernel
6096 still does not support), as /dev/shm and /tmp are still
4ef6e535 6097 shared resources used by both the system and unprivileged
699b6b34
LP
6098 users.
6099
6100 * logind will now automatically turn off automatic suspending
6101 on laptop lid close when more than one display is
6102 connected. This was previously expected to be implemented
6103 individually in desktop environments (such as GNOME),
6104 however has been added to logind now, in order to fix a
6105 boot-time race where a desktop environment might not have
6106 been started yet and thus not been able to take an inhibitor
6107 lock at the time where logind already suspends the system
6108 due to a closed lid.
6109
6110 * logind will now wait at least 30s after each system
6111 suspend/resume cycle, and 3min after system boot before
6112 suspending the system due to a closed laptop lid. This
6113 should give USB docking stations and similar enough time to
4ef6e535 6114 be probed and configured after system resume and boot in
699b6b34
LP
6115 order to then act as suspend blocker.
6116
6117 * systemd-run gained a new --property= setting which allows
6118 initialization of resource control properties (and others)
6119 for the created scope or service unit. Example: "systemd-run
6120 --property=BlockIOWeight=10 updatedb" may be used to run
6121 updatedb at a low block IO scheduling weight.
6122
6123 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
6124 now also work in --scope mode.
6125
6126 * When systemd is compiled with kdbus support, basic support
6127 for enforced policies is now in place. (Note that enabling
6128 kdbus still voids your warranty and no API compatibility
6129 promises are made.)
6130
6131 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
6132 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
6133 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
6134 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
6135 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
6136 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
6137 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
6138 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
6139 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
6140 Zbigniew Jędrzejewski-Szmek
6141
ccddd104 6142 — Berlin, 2014-03-12
699b6b34 6143
43c71255
LP
6144CHANGES WITH 210:
6145
6146 * systemd will now relabel /dev after loading the SMACK policy
6147 according to SMACK rules.
6148
67dd87c5 6149 * A new unit file option AppArmorProfile= has been added to
43c71255
LP
6150 set the AppArmor profile for the processes of a unit.
6151
6152 * A new condition check ConditionArchitecture= has been added
6153 to conditionalize units based on the system architecture, as
6154 reported by uname()'s "machine" field.
6155
6156 * systemd-networkd now supports matching on the system
6157 virtualization, architecture, kernel command line, host name
6158 and machine ID.
6159
ed28905e 6160 * logind is now a lot more aggressive when suspending the
43c71255 6161 machine due to a closed laptop lid. Instead of acting only
b8bde116 6162 on the lid close action, it will continuously watch the lid
43c71255
LP
6163 status and act on it. This is useful for laptops where the
6164 power button is on the outside of the chassis so that it can
ed28905e 6165 be reached without opening the lid (such as the Lenovo
b8bde116 6166 Yoga). On those machines, logind will now immediately
ed28905e 6167 re-suspend the machine if the power button has been
43c71255
LP
6168 accidentally pressed while the laptop was suspended and in a
6169 backpack or similar.
6170
6171 * logind will now watch SW_DOCK switches and inhibit reaction
6172 to the lid switch if it is pressed. This means that logind
d27893ef 6173 will not suspend the machine anymore if the lid is closed
949138cc 6174 and the system is docked, if the laptop supports SW_DOCK
43c71255
LP
6175 notifications via the input layer. Note that ACPI docking
6176 stations do not generate this currently. Also note that this
6177 logic is usually not fully sufficient and Desktop
6178 Environments should take a lid switch inhibitor lock when an
6179 external display is connected, as systemd will not watch
6180 this on its own.
6181
6182 * nspawn will now make use of the devices cgroup controller by
6183 default, and only permit creation of and access to the usual
6184 API device nodes like /dev/null or /dev/random, as well as
6185 access to (but not creation of) the pty devices.
6186
6187 * We will now ship a default .network file for
6188 systemd-networkd that automatically configures DHCP for
6189 network interfaces created by nspawn's --network-veth or
6190 --network-bridge= switches.
6191
6192 * systemd will now understand the usual M, K, G, T suffixes
6193 according to SI conventions (i.e. to the base 1000) when
6194 referring to throughput and hardware metrics. It will stay
6195 with IEC conventions (i.e. to the base 1024) for software
6196 metrics, according to what is customary according to
6197 Wikipedia. We explicitly document which base applies for
6198 each configuration option.
6199
6200 * The DeviceAllow= setting in unit files now supports a syntax
ed28905e 6201 to whitelist an entire group of devices node majors at once,
43c71255 6202 based on the /proc/devices listing. For example, with the
b8bde116 6203 string "char-pts", it is now possible to whitelist all
43c71255
LP
6204 current and future pseudo-TTYs at once.
6205
6206 * sd-event learned a new "post" event source. Event sources of
6207 this type are triggered by the dispatching of any event
6208 source of a type that is not "post". This is useful for
6209 implementing clean-up and check event sources that are
6210 triggered by other work being done in the program.
6211
6212 * systemd-networkd is no longer statically enabled, but uses
6213 the usual [Install] sections so that it can be
6214 enabled/disabled using systemctl. It still is enabled by
6215 default however.
6216
b8bde116 6217 * When creating a veth interface pair with systemd-nspawn, the
43c71255
LP
6218 host side will now be prefixed with "vb-" if
6219 --network-bridge= is used, and with "ve-" if --network-veth
b8bde116 6220 is used. This way, it is easy to distinguish these cases on
43c71255
LP
6221 the host, for example to apply different configuration to
6222 them with systemd-networkd.
6223
d27893ef
LP
6224 * The compatibility libraries for libsystemd-journal.so,
6225 libsystem-id128.so, libsystemd-login.so and
6226 libsystemd-daemon.so do not make use of IFUNC
b8bde116 6227 anymore. Instead, we now build libsystemd.so multiple times
d27893ef
LP
6228 under these alternative names. This means that the footprint
6229 is drastically increased, but given that these are
b8bde116 6230 transitional compatibility libraries, this should not matter
d27893ef
LP
6231 much. This change has been made necessary to support the ARM
6232 platform for these compatibility libraries, as the ARM
d28315e4 6233 toolchain is not really at the same level as the toolchain
ed28905e 6234 for other architectures like x86 and does not support
d27893ef
LP
6235 IFUNC. Please make sure to use --enable-compat-libs only
6236 during a transitional period!
6237
13b28d82 6238 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
43c71255
LP
6239 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
6240 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
6241 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
6242 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
6243 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
6244 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
6245 Zbigniew Jędrzejewski-Szmek
6246
ccddd104 6247 — Berlin, 2014-02-24
43c71255 6248
e49b5aad
LP
6249CHANGES WITH 209:
6250
6251 * A new component "systemd-networkd" has been added that can
6252 be used to configure local network interfaces statically or
8b7d0494
JSJ
6253 via DHCP. It is capable of bringing up bridges, VLANs, and
6254 bonding. Currently, no hook-ups for interactive network
4670e9d5 6255 configuration are provided. Use this for your initrd,
8b7d0494
JSJ
6256 container, embedded, or server setup if you need a simple,
6257 yet powerful, network configuration solution. This
4670e9d5 6258 configuration subsystem is quite nifty, as it allows wildcard
1e190502 6259 hotplug matching in interfaces. For example, with a single
4670e9d5 6260 configuration snippet, you can configure that all Ethernet
1e190502
ZJS
6261 interfaces showing up are automatically added to a bridge,
6262 or similar. It supports link-sensing and more.
e49b5aad
LP
6263
6264 * A new tool "systemd-socket-proxyd" has been added which can
4c2413bf 6265 act as a bidirectional proxy for TCP sockets. This is
e49b5aad
LP
6266 useful for adding socket activation support to services that
6267 do not actually support socket activation, including virtual
4c2413bf 6268 machines and the like.
e49b5aad
LP
6269
6270 * Add a new tool to save/restore rfkill state on
6271 shutdown/boot.
6272
8b7d0494
JSJ
6273 * Save/restore state of keyboard backlights in addition to
6274 display backlights on shutdown/boot.
e49b5aad
LP
6275
6276 * udev learned a new SECLABEL{} construct to label device
6277 nodes with a specific security label when they appear. For
4c2413bf 6278 now, only SECLABEL{selinux} is supported, but the syntax is
e49b5aad
LP
6279 prepared for additional security frameworks.
6280
6281 * udev gained a new scheme to configure link-level attributes
6282 from files in /etc/systemd/network/*.link. These files can
8b7d0494 6283 match against MAC address, device path, driver name and type,
4c2413bf 6284 and will apply attributes like the naming policy, link speed,
8b7d0494 6285 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
e49b5aad
LP
6286 address assignment policy (randomized, ...).
6287
dfb08b05
ZJS
6288 * The configuration of network interface naming rules for
6289 "permanent interface names" has changed: a new NamePolicy=
6290 setting in the [Link] section of .link files determines the
a8eaaee7 6291 priority of possible naming schemes (onboard, slot, MAC,
dfb08b05
ZJS
6292 path). The default value of this setting is determined by
6293 /usr/lib/net/links/99-default.link. Old
6294 80-net-name-slot.rules udev configuration file has been
6295 removed, so local configuration overriding this file should
ce830873 6296 be adapted to override 99-default.link instead.
dfb08b05 6297
e49b5aad 6298 * When the User= switch is used in a unit file, also
4c2413bf 6299 initialize $SHELL= based on the user database entry.
e49b5aad
LP
6300
6301 * systemd no longer depends on libdbus. All communication is
6302 now done with sd-bus, systemd's low-level bus library
6303 implementation.
6304
6305 * kdbus support has been added to PID 1 itself. When kdbus is
4c2413bf 6306 enabled, this causes PID 1 to set up the system bus and
e49b5aad
LP
6307 enable support for a new ".busname" unit type that
6308 encapsulates bus name activation on kdbus. It works a little
6309 bit like ".socket" units, except for bus names. A new
6310 generator has been added that converts classic dbus1 service
6311 activation files automatically into native systemd .busname
6312 and .service units.
6313
6314 * sd-bus: add a light-weight vtable implementation that allows
6315 defining objects on the bus with a simple static const
6316 vtable array of its methods, signals and properties.
6317
8b7d0494 6318 * systemd will not generate or install static dbus
e49b5aad 6319 introspection data anymore to /usr/share/dbus-1/interfaces,
1e190502 6320 as the precise format of these files is unclear, and
e49b5aad
LP
6321 nothing makes use of it.
6322
6323 * A proxy daemon is now provided to proxy clients connecting
6324 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
6325 compatibility with classic D-Bus.
6326
6327 * A bus driver implementation has been added that supports the
6328 classic D-Bus bus driver calls on kdbus, also for
6329 compatibility purposes.
6330
6331 * A new API "sd-event.h" has been added that implements a
6332 minimal event loop API built around epoll. It provides a
6333 couple of features that direct epoll usage is lacking:
b9761003 6334 prioritization of events, scales to large numbers of timer
e49b5aad
LP
6335 events, per-event timer slack (accuracy), system-wide
6336 coalescing of timer events, exit handlers, watchdog
6337 supervision support using systemd's sd_notify() API, child
6338 process handling.
6339
6340 * A new API "sd-rntl.h" has been added that provides an API
6341 around the route netlink interface of the kernel, similar in
6342 style to "sd-bus.h".
6343
7e95eda5
PF
6344 * A new API "sd-dhcp-client.h" has been added that provides a
6345 small DHCPv4 client-side implementation. This is used by
e49b5aad
LP
6346 "systemd-networkd".
6347
4c2413bf 6348 * There is a new kernel command line option
8b7d0494
JSJ
6349 "systemd.restore_state=0|1". When set to "0", none of the
6350 systemd tools will restore saved runtime state to hardware
6351 devices. More specifically, the rfkill and backlight states
6352 are not restored.
e49b5aad
LP
6353
6354 * The FsckPassNo= compatibility option in mount/service units
6355 has been removed. The fstab generator will now add the
6356 necessary dependencies automatically, and does not require
6357 PID1's support for that anymore.
6358
8b7d0494 6359 * journalctl gained a new switch, --list-boots, that lists
e49b5aad
LP
6360 recent boots with their times and boot IDs.
6361
6362 * The various tools like systemctl, loginctl, timedatectl,
6363 busctl, systemd-run, ... have gained a new switch "-M" to
6364 connect to a specific, local OS container (as direct
6365 connection, without requiring SSH). This works on any
6366 container that is registered with machined, such as those
6367 created by libvirt-lxc or nspawn.
6368
6369 * systemd-run and systemd-analyze also gained support for "-H"
4c2413bf 6370 to connect to remote hosts via SSH. This is particularly
8b7d0494
JSJ
6371 useful for systemd-run because it enables queuing of jobs
6372 onto remote systems.
e49b5aad
LP
6373
6374 * machinectl gained a new command "login" to open a getty
6375 login in any local container. This works with any container
6376 that is registered with machined (such as those created by
8e420494 6377 libvirt-lxc or nspawn), and which runs systemd inside.
e49b5aad
LP
6378
6379 * machinectl gained a new "reboot" command that may be used to
6380 trigger a reboot on a specific container that is registered
6381 with machined. This works on any container that runs an init
6382 system of some kind.
6383
6384 * systemctl gained a new "list-timers" command to print a nice
6385 listing of installed timer units with the times they elapse
6386 next.
6387
6388 * Alternative reboot() parameters may now be specified on the
6389 "systemctl reboot" command line and are passed to the
6390 reboot() system call.
6391
6392 * systemctl gained a new --job-mode= switch to configure the
6393 mode to queue a job with. This is a more generic version of
8b7d0494 6394 --fail, --irreversible, and --ignore-dependencies, which are
e49b5aad
LP
6395 still available but not advertised anymore.
6396
e49b5aad
LP
6397 * /etc/systemd/system.conf gained new settings to configure
6398 various default timeouts of units, as well as the default
b9761003 6399 start limit interval and burst. These may still be overridden
e49b5aad
LP
6400 within each Unit.
6401
270f1624
LP
6402 * PID1 will now export on the bus profile data of the security
6403 policy upload process (such as the SELinux policy upload to
8e420494 6404 the kernel).
e49b5aad 6405
4670e9d5 6406 * journald: when forwarding logs to the console, include
1e190502
ZJS
6407 timestamps (following the setting in
6408 /sys/module/printk/parameters/time).
e49b5aad
LP
6409
6410 * OnCalendar= in timer units now understands the special
6411 strings "yearly" and "annually". (Both are equivalent)
6412
6413 * The accuracy of timer units is now configurable with the new
6414 AccuracySec= setting. It defaults to 1min.
6415
6416 * A new dependency type JoinsNamespaceOf= has been added that
6417 allows running two services within the same /tmp and network
6418 namespace, if PrivateNetwork= or PrivateTmp= are used.
6419
6420 * A new command "cat" has been added to systemctl. It outputs
6421 the original unit file of a unit, and concatenates the
1e190502
ZJS
6422 contents of additional "drop-in" unit file snippets, so that
6423 the full configuration is shown.
e49b5aad
LP
6424
6425 * systemctl now supports globbing on the various "list-xyz"
6426 commands, like "list-units" or "list-sockets", as well as on
1e190502
ZJS
6427 those commands which take multiple unit names.
6428
6429 * journalctl's --unit= switch gained support for globbing.
e49b5aad
LP
6430
6431 * All systemd daemons now make use of the watchdog logic so
6432 that systemd automatically notices when they hang.
6433
4c2413bf 6434 * If the $container_ttys environment variable is set,
e49b5aad
LP
6435 getty-generator will automatically spawn a getty for each
6436 listed tty. This is useful for container managers to request
6437 login gettys to be spawned on as many ttys as needed.
6438
6439 * %h, %s, %U specifier support is not available anymore when
6440 used in unit files for PID 1. This is because NSS calls are
6441 not safe from PID 1. They stay available for --user
6442 instances of systemd, and as special case for the root user.
6443
e49b5aad
LP
6444 * loginctl gained a new "--no-legend" switch to turn off output
6445 of the legend text.
6446
6447 * The "sd-login.h" API gained three new calls:
6448 sd_session_is_remote(), sd_session_get_remote_user(),
6449 sd_session_get_remote_host() to query information about
6450 remote sessions.
6451
8e420494
LP
6452 * The udev hardware database now also carries vendor/product
6453 information of SDIO devices.
e49b5aad
LP
6454
6455 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
6456 determine whether watchdog notifications are requested by
6457 the system manager.
6458
1e190502 6459 * Socket-activated per-connection services now include a
e49b5aad
LP
6460 short description of the connection parameters in the
6461 description.
6462
4c2413bf 6463 * tmpfiles gained a new "--boot" option. When this is not used,
e49b5aad 6464 only lines where the command character is not suffixed with
4670e9d5 6465 "!" are executed. When this option is specified, those
1e190502
ZJS
6466 options are executed too. This partitions tmpfiles
6467 directives into those that can be safely executed at any
6468 time, and those which should be run only at boot (for
6469 example, a line that creates /run/nologin).
e49b5aad 6470
c0c5af00 6471 * A new API "sd-resolve.h" has been added which provides a simple
4c2413bf 6472 asynchronous wrapper around glibc NSS host name resolution
e49b5aad 6473 calls, such as getaddrinfo(). In contrast to glibc's
4c2413bf
JE
6474 getaddrinfo_a(), it does not use signals. In contrast to most
6475 other asynchronous name resolution libraries, this one does
6476 not reimplement DNS, but reuses NSS, so that alternate
e49b5aad 6477 host name resolution systems continue to work, such as mDNS,
8b7d0494 6478 LDAP, etc. This API is based on libasyncns, but it has been
e49b5aad
LP
6479 cleaned up for inclusion in systemd.
6480
6300b3ec
LP
6481 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
6482 "sd-daemon.h" are no longer found in individual libraries
6483 libsystemd-journal.so, libsystemd-login.so,
6484 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
8b7d0494
JSJ
6485 merged them into a single library, libsystemd.so, which
6486 provides all symbols. The reason for this is cyclic
e49b5aad 6487 dependencies, as these libraries tend to use each other's
d28315e4 6488 symbols. So far, we have managed to workaround that by linking
6300b3ec
LP
6489 a copy of a good part of our code into each of these
6490 libraries again and again, which, however, makes certain
6491 things hard to do, like sharing static variables. Also, it
6492 substantially increases footprint. With this change, there
6493 is only one library for the basic APIs systemd
6494 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
6495 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
6496 library as well, however are subject to the --enable-kdbus
6497 switch (see below). Note that "sd-dhcp-client.h" is not part
6498 of this library (this is because it only consumes, never
6499 provides, services of/to other APIs). To make the transition
8b7d0494 6500 easy from the separate libraries to the unified one, we
4c2413bf 6501 provide the --enable-compat-libs compile-time switch which
e49b5aad
LP
6502 will generate stub libraries that are compatible with the
6503 old ones but redirect all calls to the new one.
6504
8b7d0494 6505 * All of the kdbus logic and the new APIs "sd-bus.h",
e49b5aad 6506 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
8b7d0494
JSJ
6507 and "sd-utf8.h" are compile-time optional via the
6508 "--enable-kdbus" switch, and they are not compiled in by
6509 default. To make use of kdbus, you have to explicitly enable
4c2413bf 6510 the switch. Note however, that neither the kernel nor the
e49b5aad
LP
6511 userspace API for all of this is considered stable yet. We
6512 want to maintain the freedom to still change the APIs for
4c2413bf 6513 now. By specifying this build-time switch, you acknowledge
e49b5aad 6514 that you are aware of the instability of the current
ad42cf73
KS
6515 APIs.
6516
6517 * Also, note that while kdbus is pretty much complete,
e49b5aad 6518 it lacks one thing: proper policy support. This means you
8b7d0494 6519 can build a fully working system with all features; however,
4c2413bf
JE
6520 it will be highly insecure. Policy support will be added in
6521 one of the next releases, at the same time that we will
6522 declare the APIs stable.
e49b5aad 6523
81c7dd89 6524 * When the kernel command line argument "kdbus" is specified,
ad42cf73 6525 systemd will automatically load the kdbus.ko kernel module. At
8b7d0494 6526 this stage of development, it is only useful for testing kdbus
ad42cf73 6527 and should not be used in production. Note: if "--enable-kdbus"
8b7d0494 6528 is specified, and the kdbus.ko kernel module is available, and
ad42cf73
KS
6529 "kdbus" is added to the kernel command line, the entire system
6530 runs with kdbus instead of dbus-daemon, with the above mentioned
6531 problem of missing the system policy enforcement. Also a future
6532 version of kdbus.ko or a newer systemd will not be compatible with
6533 each other, and will unlikely be able to boot the machine if only
6534 one of them is updated.
6535
e49b5aad 6536 * systemctl gained a new "import-environment" command which
4c2413bf 6537 uploads the caller's environment (or parts thereof) into the
e49b5aad
LP
6538 service manager so that it is inherited by services started
6539 by the manager. This is useful to upload variables like
6540 $DISPLAY into the user service manager.
6541
6542 * A new PrivateDevices= switch has been added to service units
6543 which allows running a service with a namespaced /dev
6544 directory that does not contain any device nodes for
4c2413bf 6545 physical devices. More specifically, it only includes devices
8b7d0494 6546 such as /dev/null, /dev/urandom, and /dev/zero which are API
e49b5aad
LP
6547 entry points.
6548
6549 * logind has been extended to support behaviour like VT
6550 switching on seats that do not support a VT. This makes
6551 multi-session available on seats that are not the first seat
6552 (seat0), and on systems where kernel support for VTs has
8b7d0494 6553 been disabled at compile-time.
e49b5aad
LP
6554
6555 * If a process holds a delay lock for system sleep or shutdown
1e190502 6556 and fails to release it in time, we will now log its
e49b5aad
LP
6557 identity. This makes it easier to identify processes that
6558 cause slow suspends or power-offs.
6559
1e190502
ZJS
6560 * When parsing /etc/crypttab, support for a new key-slot=
6561 option as supported by Debian is added. It allows indicating
6562 which LUKS slot to use on disk, speeding up key loading.
e49b5aad 6563
000b1ba5 6564 * The sd_journal_sendv() API call has been checked and
1e190502
ZJS
6565 officially declared to be async-signal-safe so that it may
6566 be invoked from signal handlers for logging purposes.
e49b5aad
LP
6567
6568 * Boot-time status output is now enabled automatically after a
6569 short timeout if boot does not progress, in order to give
8e420494 6570 the user an indication what she or he is waiting for.
1e190502
ZJS
6571
6572 * The boot-time output has been improved to show how much time
6573 remains until jobs expire.
e49b5aad
LP
6574
6575 * The KillMode= switch in service units gained a new possible
8b7d0494 6576 value "mixed". If set, and the unit is shut down, then the
e49b5aad 6577 initial SIGTERM signal is sent only to the main daemon
8e420494 6578 process, while the following SIGKILL signal is sent to
e49b5aad
LP
6579 all remaining processes of the service.
6580
4c2413bf
JE
6581 * When a scope unit is registered, a new property "Controller"
6582 may be set. If set to a valid bus name, systemd will send a
e49b5aad
LP
6583 RequestStop() signal to this name when it would like to shut
6584 down the scope. This may be used to hook manager logic into
6585 the shutdown logic of scope units. Also, scope units may now
8b7d0494 6586 be put in a special "abandoned" state, in which case the
e49b5aad
LP
6587 manager process which created them takes no further
6588 responsibilities for it.
6589
1e190502 6590 * When reading unit files, systemd will now verify
e49b5aad
LP
6591 the access mode of these files, and warn about certain
6592 suspicious combinations. This has been added to make it
6593 easier to track down packaging bugs where unit files are
6594 marked executable or world-writable.
6595
6596 * systemd-nspawn gained a new "--setenv=" switch to set
8b7d0494 6597 container-wide environment variables. The similar option in
1e190502
ZJS
6598 systemd-activate was renamed from "--environment=" to
6599 "--setenv=" for consistency.
e49b5aad
LP
6600
6601 * systemd-nspawn has been updated to create a new kdbus domain
6602 for each container that is invoked, thus allowing each
b9761003 6603 container to have its own set of system and user buses,
8b7d0494 6604 independent of the host.
e49b5aad
LP
6605
6606 * systemd-nspawn gained a new --drop-capability= switch to run
6607 the container with less capabilities than the default. Both
b9761003 6608 --drop-capability= and --capability= now take the special
e49b5aad
LP
6609 string "all" for dropping or keeping all capabilities.
6610
6611 * systemd-nspawn gained new switches for executing containers
6612 with specific SELinux labels set.
6613
6614 * systemd-nspawn gained a new --quiet switch to not generate
6615 any additional output but the container's own console
6616 output.
6617
6618 * systemd-nspawn gained a new --share-system switch to run a
6619 container without PID namespacing enabled.
6620
6621 * systemd-nspawn gained a new --register= switch to control
1e190502 6622 whether the container is registered with systemd-machined or
8e420494 6623 not. This is useful for containers that do not run full
e49b5aad
LP
6624 OS images, but only specific apps.
6625
6626 * systemd-nspawn gained a new --keep-unit which may be used
8b7d0494 6627 when invoked as the only program from a service unit, and
e49b5aad 6628 results in registration of the unit service itself in
1e190502 6629 systemd-machined, instead of a newly opened scope unit.
e49b5aad
LP
6630
6631 * systemd-nspawn gained a new --network-interface= switch for
6632 moving arbitrary interfaces to the container. The new
4c2413bf 6633 --network-veth switch creates a virtual Ethernet connection
8b7d0494
JSJ
6634 between host and container. The new --network-bridge=
6635 switch then allows assigning the host side of this virtual
6636 Ethernet connection to a bridge device.
e49b5aad 6637
6afc95b7
LP
6638 * systemd-nspawn gained a new --personality= switch for
6639 setting the kernel personality for the container. This is
70a44afe 6640 useful when running a 32-bit container on a 64-bit host. A
b8bde116
JE
6641 similar option Personality= is now also available for service
6642 units to use.
6afc95b7 6643
e49b5aad
LP
6644 * logind will now also track a "Desktop" identifier for each
6645 session which encodes the desktop environment of it. This is
6646 useful for desktop environments that want to identify
6647 multiple running sessions of itself easily.
6648
6649 * A new SELinuxContext= setting for service units has been
6650 added that allows setting a specific SELinux execution
6651 context for a service.
6652
6653 * Most systemd client tools will now honour $SYSTEMD_LESS for
6654 settings of the "less" pager. By default, these tools will
8b7d0494
JSJ
6655 override $LESS to allow certain operations to work, such as
6656 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
e49b5aad
LP
6657 influence this logic.
6658
6659 * systemd's "seccomp" hook-up has been changed to make use of
6660 the libseccomp library instead of using its own
6661 implementation. This has benefits for portability among
6662 other things.
6663
4c2413bf 6664 * For usage together with SystemCallFilter=, a new
8b7d0494 6665 SystemCallErrorNumber= setting has been introduced that
b8bde116
JE
6666 allows configuration of a system error number to be returned
6667 on filtered system calls, instead of immediately killing the
e49b5aad
LP
6668 process. Also, SystemCallArchitectures= has been added to
6669 limit access to system calls of a particular architecture
6670 (in order to turn off support for unused secondary
4c2413bf 6671 architectures). There is also a global
8b7d0494 6672 SystemCallArchitectures= setting in system.conf now to turn
e49b5aad
LP
6673 off support for non-native system calls system-wide.
6674
210054d7
KS
6675 * systemd requires a kernel with a working name_to_handle_at(),
6676 please see the kernel config requirements in the README file.
6677
e49b5aad
LP
6678 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
6679 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
6680 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
6681 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
6682 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
6683 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
6684 Elia Pinto, Florian Weimer, George McCollister, Goffredo
6685 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
6686 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
6687 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
6688 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
6689 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
6690 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
6691 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
6692 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
6693 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
6694 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
6695 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
6696 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
6697 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
6698 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
6699 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
6700 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
6701 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
6702
ccddd104 6703 — Berlin, 2014-02-20
e49b5aad 6704
cd4010b3
LP
6705CHANGES WITH 208:
6706
6707 * logind has gained support for facilitating privileged input
6708 and drm device access for unprivileged clients. This work is
6709 useful to allow Wayland display servers (and similar
6710 programs, such as kmscon) to run under the user's ID and
6711 access input and drm devices which are normally
6712 protected. When this is used (and the kernel is new enough)
6713 logind will "mute" IO on the file descriptors passed to
6714 Wayland as long as it is in the background and "unmute" it
6715 if it returns into the foreground. This allows secure
6716 session switching without allowing background sessions to
6717 eavesdrop on input and display data. This also introduces
6718 session switching support if VT support is turned off in the
6719 kernel, and on seats that are not seat0.
6720
6721 * A new kernel command line option luks.options= is understood
06b643e7 6722 now which allows specifying LUKS options for usage for LUKS
cd4010b3
LP
6723 encrypted partitions specified with luks.uuid=.
6724
6725 * tmpfiles.d(5) snippets may now use specifier expansion in
6726 path names. More specifically %m, %b, %H, %v, are now
6727 replaced by the local machine id, boot id, hostname, and
6728 kernel version number.
6729
6730 * A new tmpfiles.d(5) command "m" has been introduced which
6731 may be used to change the owner/group/access mode of a file
d28315e4 6732 or directory if it exists, but do nothing if it does not.
cd4010b3
LP
6733
6734 * This release removes high-level support for the
6735 MemorySoftLimit= cgroup setting. The underlying kernel
6736 cgroup attribute memory.soft_limit= is currently badly
6737 designed and likely to be removed from the kernel API in its
d28315e4 6738 current form, hence we should not expose it for now.
cd4010b3
LP
6739
6740 * The memory.use_hierarchy cgroup attribute is now enabled for
6741 all cgroups systemd creates in the memory cgroup
6742 hierarchy. This option is likely to be come the built-in
cc98b302
TH
6743 default in the kernel anyway, and the non-hierarchical mode
6744 never made much sense in the intrinsically hierarchical
cd4010b3
LP
6745 cgroup system.
6746
6747 * A new field _SYSTEMD_SLICE= is logged along with all journal
6748 messages containing the slice a message was generated
6749 from. This is useful to allow easy per-customer filtering of
6750 logs among other things.
6751
6752 * systemd-journald will no longer adjust the group of journal
6753 files it creates to the "systemd-journal" group. Instead we
6754 rely on the journal directory to be owned by the
6755 "systemd-journal" group, and its setgid bit set, so that the
6756 kernel file system layer will automatically enforce that
6757 journal files inherit this group assignment. The reason for
6758 this change is that we cannot allow NSS look-ups from
6759 journald which would be necessary to resolve
6760 "systemd-journal" to a numeric GID, because this might
6761 create deadlocks if NSS involves synchronous queries to
6762 other daemons (such as nscd, or sssd) which in turn are
6763 logging clients of journald and might block on it, which
6764 would then dead lock. A tmpfiles.d(5) snippet included in
6765 systemd will make sure the setgid bit and group are
6766 properly set on the journal directory if it exists on every
6767 boot. However, we recommend adjusting it manually after
6768 upgrades too (or from RPM scriptlets), so that the change is
6769 not delayed until next reboot.
6770
6771 * Backlight and random seed files in /var/lib/ have moved into
6772 the /var/lib/systemd/ directory, in order to centralize all
6773 systemd generated files in one directory.
6774
6775 * Boot time performance measurements (as displayed by
6776 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
6777 performance information if that's available to determine how
6778 much time BIOS and boot loader initialization required. With
6779 a sufficiently new BIOS you hence no longer need to boot
6780 with Gummiboot to get access to such information.
6781
6782 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
6783 Cristian Rodríguez, Dave Reisner, David Herrmann, David
6784 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
6785 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
6786 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
6787 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
6788 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
6789
ccddd104 6790 — Berlin, 2013-10-02
cd4010b3 6791
4f0be680
LP
6792CHANGES WITH 207:
6793
6794 * The Restart= option for services now understands a new
f3a165b0 6795 on-watchdog setting, which will restart the service
4f0be680
LP
6796 automatically if the service stops sending out watchdog keep
6797 alive messages (as configured with WatchdogSec=).
6798
6799 * The getty generator (which is responsible for bringing up a
6800 getty on configured serial consoles) will no longer only
6801 start a getty on the primary kernel console but on all
6802 others, too. This makes the order in which console= is
6803 specified on the kernel command line less important.
6804
6805 * libsystemd-logind gained a new sd_session_get_vt() call to
6806 retrieve the VT number of a session.
6807
6808 * If the option "tries=0" is set for an entry of /etc/crypttab
6809 its passphrase is queried indefinitely instead of any
6810 maximum number of tries.
6811
6812 * If a service with a configure PID file terminates its PID
6813 file will now be removed automatically if it still exists
6814 afterwards. This should put an end to stale PID files.
6815
6816 * systemd-run will now also take relative binary path names
6817 for execution and no longer insists on absolute paths.
6818
6819 * InaccessibleDirectories= and ReadOnlyDirectories= now take
6820 paths that are optionally prefixed with "-" to indicate that
d28315e4 6821 it should not be considered a failure if they do not exist.
4f0be680 6822
f3a165b0
KS
6823 * journalctl -o (and similar commands) now understands a new
6824 output mode "short-precise", it is similar to "short" but
4f0be680
LP
6825 shows timestamps with usec accuracy.
6826
6827 * The option "discard" (as known from Debian) is now
6828 synonymous to "allow-discards" in /etc/crypttab. In fact,
387abf80 6829 "discard" is preferred now (since it is easier to remember
4f0be680
LP
6830 and type).
6831
f3a165b0 6832 * Some licensing clean-ups were made, so that more code is now
4f0be680
LP
6833 LGPL-2.1 licensed than before.
6834
6835 * A minimal tool to save/restore the display backlight
6836 brightness across reboots has been added. It will store the
f3a165b0 6837 backlight setting as late as possible at shutdown, and
4f0be680
LP
6838 restore it as early as possible during reboot.
6839
6840 * A logic to automatically discover and enable home and swap
6841 partitions on GPT disks has been added. With this in place
6842 /etc/fstab becomes optional for many setups as systemd can
6843 discover certain partitions located on the root disk
6844 automatically. Home partitions are recognized under their
6845 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
6846 partitions are recognized under their GPT type ID
6847 0657fd6da4ab43c484e50933c84b4f4f.
6848
6849 * systemd will no longer pass any environment from the kernel
6850 or initrd to system services. If you want to set an
6851 environment for all services, do so via the kernel command
6852 line systemd.setenv= assignment.
6853
387abf80
LP
6854 * The systemd-sysctl tool no longer natively reads the file
6855 /etc/sysctl.conf. If desired, the file should be symlinked
6856 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
6857 legacy support by a symlink rather than built-in code, it
6858 also makes the otherwise hidden order of application of the
6859 different files visible. (Note that this partly reverts to a
6860 pre-198 application order of sysctl knobs!)
04bf3c1a 6861
4f0be680
LP
6862 * The "systemctl set-log-level" and "systemctl dump" commands
6863 have been moved to systemd-analyze.
6864
6865 * systemd-run learned the new --remain-after-exit switch,
6866 which causes the scope unit not to be cleaned up
6867 automatically after the process terminated.
6868
6869 * tmpfiles learned a new --exclude-prefix= switch to exclude
6870 certain paths from operation.
6871
6872 * journald will now automatically flush all messages to disk
f47ad593
ZJS
6873 as soon as a message at the log level CRIT, ALERT or EMERG
6874 is received.
4f0be680
LP
6875
6876 Contributions from: Andrew Cook, Brandon Philips, Christian
6877 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
6878 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
6879 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
6880 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
6881 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
6882 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
6883 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
6884 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
6885 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
6886 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
6887 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
6888 William Giokas, Zbigniew Jędrzejewski-Szmek
6889
ccddd104 6890 — Berlin, 2013-09-13
4f0be680 6891
408f281b
LP
6892CHANGES WITH 206:
6893
6894 * The documentation has been updated to cover the various new
6895 concepts introduced with 205.
6896
6897 * Unit files now understand the new %v specifier which
6898 resolves to the kernel version string as returned by "uname
6899 -r".
6900
6901 * systemctl now supports filtering the unit list output by
6902 load state, active state and sub state, using the new
33b521be 6903 --state= parameter.
408f281b
LP
6904
6905 * "systemctl status" will now show the results of the
6906 condition checks (like ConditionPathExists= and similar) of
6907 the last start attempts of the unit. They are also logged to
6908 the journal.
6909
6910 * "journalctl -b" may now be used to look for boot output of a
6911 specific boot. Try "journalctl -b -1" for the previous boot,
6912 but the syntax is substantially more powerful.
6913
6914 * "journalctl --show-cursor" has been added which prints the
6915 cursor string the last shown log line. This may then be used
6916 with the new "journalctl --after-cursor=" switch to continue
6917 browsing logs from that point on.
6918
6919 * "journalctl --force" may now be used to force regeneration
6920 of an FSS key.
6921
251cc819
LP
6922 * Creation of "dead" device nodes has been moved from udev
6923 into kmod and tmpfiles. Previously, udev would read the kmod
6924 databases to pre-generate dead device nodes based on meta
6925 information contained in kernel modules, so that these would
6926 be auto-loaded on access rather then at boot. As this
d28315e4 6927 does not really have much to do with the exposing actual
251cc819
LP
6928 kernel devices to userspace this has always been slightly
6929 alien in the udev codebase. Following the new scheme kmod
6930 will now generate a runtime snippet for tmpfiles from the
6931 module meta information and it now is tmpfiles' job to the
6932 create the nodes. This also allows overriding access and
6933 other parameters for the nodes using the usual tmpfiles
6934 facilities. As side effect this allows us to remove the
6935 CAP_SYS_MKNOD capability bit from udevd entirely.
6936
6937 * logind's device ACLs may now be applied to these "dead"
6938 devices nodes too, thus finally allowing managed access to
ce830873 6939 devices such as /dev/snd/sequencer without loading the
251cc819 6940 backing module right-away.
408f281b
LP
6941
6942 * A new RPM macro has been added that may be used to apply
6943 tmpfiles configuration during package installation.
6944
6945 * systemd-detect-virt and ConditionVirtualization= now can
6946 detect User-Mode-Linux machines (UML).
6947
251cc819
LP
6948 * journald will now implicitly log the effective capabilities
6949 set of processes in the message metadata.
408f281b
LP
6950
6951 * systemd-cryptsetup has gained support for TrueCrypt volumes.
6952
6953 * The initrd interface has been simplified (more specifically,
6954 support for passing performance data via environment
6955 variables and fsck results via files in /run has been
6956 removed). These features were non-essential, and are
6957 nowadays available in a much nicer way by having systemd in
6958 the initrd serialize its state and have the hosts systemd
6959 deserialize it again.
6960
28f5c779
KS
6961 * The udev "keymap" data files and tools to apply keyboard
6962 specific mappings of scan to key codes, and force-release
6963 scan code lists have been entirely replaced by a udev
6964 "keyboard" builtin and a hwdb data file.
408f281b 6965
251cc819
LP
6966 * systemd will now honour the kernel's "quiet" command line
6967 argument also during late shutdown, resulting in a
6968 completely silent shutdown when used.
6969
6970 * There's now an option to control the SO_REUSEPORT socket
6971 option in .socket units.
6972
6973 * Instance units will now automatically get a per-template
6974 subslice of system.slice unless something else is explicitly
6975 configured. For example, instances of sshd@.service will now
6976 implicitly be placed in system-sshd.slice rather than
6977 system.slice as before.
6978
6979 * Test coverage support may now be enabled at build time.
6980
6981 Contributions from: Dave Reisner, Frederic Crozat, Harald
6982 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
6983 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
6984 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
6985 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
6986 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
6987 Giokas, Zbigniew Jędrzejewski-Szmek
6988
ccddd104 6989 — Berlin, 2013-07-23
4f0be680 6990
00aa832b
LP
6991CHANGES WITH 205:
6992
6993 * Two new unit types have been introduced:
6994
6995 Scope units are very similar to service units, however, are
ccddd104 6996 created out of pre-existing processes — instead of PID 1
00aa832b
LP
6997 forking off the processes. By using scope units it is
6998 possible for system services and applications to group their
6999 own child processes (worker processes) in a powerful way
7000 which then maybe used to organize them, or kill them
7001 together, or apply resource limits on them.
7002
7003 Slice units may be used to partition system resources in an
cc98b302 7004 hierarchical fashion and then assign other units to them. By
00aa832b
LP
7005 default there are now three slices: system.slice (for all
7006 system services), user.slice (for all user sessions),
7007 machine.slice (for VMs and containers).
7008
7009 Slices and scopes have been introduced primarily in
7010 context of the work to move cgroup handling to a
7011 single-writer scheme, where only PID 1
7012 creates/removes/manages cgroups.
7013
7014 * There's a new concept of "transient" units. In contrast to
7015 normal units these units are created via an API at runtime,
7016 not from configuration from disk. More specifically this
7017 means it is now possible to run arbitrary programs as
7018 independent services, with all execution parameters passed
7019 in via bus APIs rather than read from disk. Transient units
7020 make systemd substantially more dynamic then it ever was,
7021 and useful as a general batch manager.
7022
7023 * logind has been updated to make use of scope and slice units
7024 for managing user sessions. As a user logs in he will get
7025 his own private slice unit, to which all sessions are added
7026 as scope units. We also added support for automatically
7027 adding an instance of user@.service for the user into the
7028 slice. Effectively logind will no longer create cgroup
7029 hierarchies on its own now, it will defer entirely to PID 1
7030 for this by means of scope, service and slice units. Since
7031 user sessions this way become entities managed by PID 1
7032 the output of "systemctl" is now a lot more comprehensive.
7033
7034 * A new mini-daemon "systemd-machined" has been added which
7035 may be used by virtualization managers to register local
7036 VMs/containers. nspawn has been updated accordingly, and
7037 libvirt will be updated shortly. machined will collect a bit
7038 of meta information about the VMs/containers, and assign
7039 them their own scope unit (see above). The collected
7040 meta-data is then made available via the "machinectl" tool,
7041 and exposed in "ps" and similar tools. machined/machinectl
7042 is compile-time optional.
7043
7044 * As discussed earlier, the low-level cgroup configuration
7045 options ControlGroup=, ControlGroupModify=,
7046 ControlGroupPersistent=, ControlGroupAttribute= have been
7047 removed. Please use high-level attribute settings instead as
7048 well as slice units.
7049
7050 * A new bus call SetUnitProperties() has been added to alter
7051 various runtime parameters of a unit. This is primarily
7052 useful to alter cgroup parameters dynamically in a nice way,
7053 but will be extended later on to make more properties
7054 modifiable at runtime. systemctl gained a new set-properties
7055 command that wraps this call.
7056
7057 * A new tool "systemd-run" has been added which can be used to
7058 run arbitrary command lines as transient services or scopes,
7059 while configuring a number of settings via the command
7060 line. This tool is currently very basic, however already
7061 very useful. We plan to extend this tool to even allow
7062 queuing of execution jobs with time triggers from the
7063 command line, similar in fashion to "at".
7064
7065 * nspawn will now inform the user explicitly that kernels with
7066 audit enabled break containers, and suggest the user to turn
7067 off audit.
7068
7069 * Support for detecting the IMA and AppArmor security
7070 frameworks with ConditionSecurity= has been added.
7071
7072 * journalctl gained a new "-k" switch for showing only kernel
1fda0ab5
ZJS
7073 messages, mimicking dmesg output; in addition to "--user"
7074 and "--system" switches for showing only user's own logs
7075 and system logs.
00aa832b
LP
7076
7077 * systemd-delta can now show information about drop-in
7078 snippets extending unit files.
7079
7080 * libsystemd-bus has been substantially updated but is still
7081 not available as public API.
7082
7083 * systemd will now look for the "debug" argument on the kernel
499b604b 7084 command line and enable debug logging, similar to what
00aa832b
LP
7085 "systemd.log_level=debug" already did before.
7086
7087 * "systemctl set-default", "systemctl get-default" has been
7088 added to configure the default.target symlink, which
7089 controls what to boot into by default.
7090
1fda0ab5
ZJS
7091 * "systemctl set-log-level" has been added as a convenient
7092 way to raise and lower systemd logging threshold.
7093
00aa832b
LP
7094 * "systemd-analyze plot" will now show the time the various
7095 generators needed for execution, as well as information
7096 about the unit file loading.
7097
00aa832b
LP
7098 * libsystemd-journal gained a new sd_journal_open_files() call
7099 for opening specific journal files. journactl also gained a
7100 new switch to expose this new functionality. Previously we
7101 only supported opening all files from a directory, or all
7102 files from the system, as opening individual files only is
7103 racy due to journal file rotation.
7104
7105 * systemd gained the new DefaultEnvironment= setting in
7106 /etc/systemd/system.conf to set environment variables for
7107 all services.
7108
7109 * If a privileged process logs a journal message with the
7110 OBJECT_PID= field set, then journald will automatically
7111 augment this with additional OBJECT_UID=, OBJECT_GID=,
7112 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
7113 system services want to log events about specific client
7114 processes. journactl/systemctl has been updated to make use
7115 of this information if all log messages regarding a specific
7116 unit is requested.
7117
7118 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
7119 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
7120 Reisner, David Coppa, David King, David Strauss, Eelco
7121 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
7122 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
7123 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
7124 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
7125 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
7126 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
7127 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
7128 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
7129 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
7130 Łukasz Stelmach, 장동준
7131
606c24e3
LP
7132CHANGES WITH 204:
7133
7134 * The Python bindings gained some minimal support for the APIs
7135 exposed by libsystemd-logind.
7136
7137 * ConditionSecurity= gained support for detecting SMACK. Since
7138 this condition already supports SELinux and AppArmor we only
7139 miss IMA for this. Patches welcome!
7140
7141 Contributions from: Karol Lewandowski, Lennart Poettering,
7142 Zbigniew Jędrzejewski-Szmek
7143
2f3fcf85
LP
7144CHANGES WITH 203:
7145
7146 * systemd-nspawn will now create /etc/resolv.conf if
7147 necessary, before bind-mounting the host's file onto it.
7148
7149 * systemd-nspawn will now store meta information about a
7150 container on the container's cgroup as extended attribute
7151 fields, including the root directory.
7152
7153 * The cgroup hierarchy has been reworked in many ways. All
7154 objects any of the components systemd creates in the cgroup
b82eed9a 7155 tree are now suffixed. More specifically, user sessions are
2f3fcf85
LP
7156 now placed in cgroups suffixed with ".session", users in
7157 cgroups suffixed with ".user", and nspawn containers in
7158 cgroups suffixed with ".nspawn". Furthermore, all cgroup
7159 names are now escaped in a simple scheme to avoid collision
7160 of userspace object names with kernel filenames. This work
7161 is preparation for making these objects relocatable in the
7162 cgroup tree, in order to allow easy resource partitioning of
7163 these objects without causing naming conflicts.
7164
7165 * systemctl list-dependencies gained the new switches
7166 --plain, --reverse, --after and --before.
7167
7168 * systemd-inhibit now shows the process name of processes that
7169 have taken an inhibitor lock.
7170
7171 * nss-myhostname will now also resolve "localhost"
7172 implicitly. This makes /etc/hosts an optional file and
7173 nicely handles that on IPv6 ::1 maps to both "localhost" and
7174 the local hostname.
7175
7176 * libsystemd-logind.so gained a new call
7177 sd_get_machine_names() to enumerate running containers and
7178 VMs (currently only supported by very new libvirt and
7179 nspawn). sd_login_monitor can now be used to watch
7180 VMs/containers coming and going.
7181
7182 * .include is not allowed recursively anymore, and only in
7183 unit files. Usually it is better to use drop-in snippets in
7184 .d/*.conf anyway, as introduced with systemd 198.
7185
7186 * systemd-analyze gained a new "critical-chain" command that
7187 determines the slowest chain of units run during system
7188 boot-up. It is very useful for tracking down where
7189 optimizing boot time is the most beneficial.
7190
7191 * systemd will no longer allow manipulating service paths in
7192 the name=systemd:/system cgroup tree using ControlGroup= in
7193 units. (But is still fine with it in all other dirs.)
7194
7195 * There's a new systemd-nspawn@.service service file that may
7196 be used to easily run nspawn containers as system
7197 services. With the container's root directory in
7198 /var/lib/container/foobar it is now sufficient to run
7199 "systemctl start systemd-nspawn@foobar.service" to boot it.
7200
7201 * systemd-cgls gained a new parameter "--machine" to list only
7202 the processes within a certain container.
7203
7204 * ConditionSecurity= now can check for "apparmor". We still
7205 are lacking checks for SMACK and IMA for this condition
7206 check though. Patches welcome!
7207
7208 * A new configuration file /etc/systemd/sleep.conf has been
7209 added that may be used to configure which kernel operation
7210 systemd is supposed to execute when "suspend", "hibernate"
7211 or "hybrid-sleep" is requested. This makes the new kernel
7212 "freeze" state accessible to the user.
7213
7214 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
7215 the passed argument if applicable.
7216
7217 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
7218 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
7219 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
7220 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
7221 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
7222 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
7223 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
7224 Jędrzejewski-Szmek
7225
ef3b5246
LP
7226CHANGES WITH 202:
7227
7228 * The output of 'systemctl list-jobs' got some polishing. The
7229 '--type=' argument may now be passed more than once. A new
7230 command 'systemctl list-sockets' has been added which shows
7231 a list of kernel sockets systemd is listening on with the
7232 socket units they belong to, plus the units these socket
7233 units activate.
7234
7235 * The experimental libsystemd-bus library got substantial
7236 updates to work in conjunction with the (also experimental)
7237 kdbus kernel project. It works well enough to exchange
7238 messages with some sophistication. Note that kdbus is not
7239 ready yet, and the library is mostly an elaborate test case
7240 for now, and not installable.
7241
7242 * systemd gained a new unit 'systemd-static-nodes.service'
7243 that generates static device nodes earlier during boot, and
7244 can run in conjunction with udev.
7245
7246 * libsystemd-login gained a new call sd_pid_get_user_unit()
7247 to retrieve the user systemd unit a process is running
7248 in. This is useful for systems where systemd is used as
7249 session manager.
7250
7251 * systemd-nspawn now places all containers in the new /machine
7252 top-level cgroup directory in the name=systemd
7253 hierarchy. libvirt will soon do the same, so that we get a
7254 uniform separation of /system, /user and /machine for system
7255 services, user processes and containers/virtual
7256 machines. This new cgroup hierarchy is also useful to stick
7257 stable names to specific container instances, which can be
7c04ad2d 7258 recognized later this way (this name may be controlled
ef3b5246
LP
7259 via systemd-nspawn's new -M switch). libsystemd-login also
7260 gained a new call sd_pid_get_machine_name() to retrieve the
7261 name of the container/VM a specific process belongs to.
7262
7263 * bootchart can now store its data in the journal.
7264
7265 * libsystemd-journal gained a new call
7266 sd_journal_add_conjunction() for AND expressions to the
7267 matching logic. This can be used to express more complex
7268 logical expressions.
7269
7270 * journactl can now take multiple --unit= and --user-unit=
7271 switches.
7272
7273 * The cryptsetup logic now understands the "luks.key=" kernel
7274 command line switch for specifying a file to read the
7c04ad2d 7275 decryption key from. Also, if a configured key file is not
ef3b5246
LP
7276 found the tool will now automatically fall back to prompting
7277 the user.
7278
cbeabcfb
ZJS
7279 * Python systemd.journal module was updated to wrap recently
7280 added functions from libsystemd-journal. The interface was
7281 changed to bring the low level interface in s.j._Reader
7282 closer to the C API, and the high level interface in
7283 s.j.Reader was updated to wrap and convert all data about
7284 an entry.
7285
ef3b5246
LP
7286 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
7287 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
7288 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
7289 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
7290 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
7291 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
7292
d3a86981
LP
7293CHANGES WITH 201:
7294
7295 * journalctl --update-catalog now understands a new --root=
7296 option to operate on catalogs found in a different root
7297 directory.
7298
7299 * During shutdown after systemd has terminated all running
7300 services a final killing loop kills all remaining left-over
7301 processes. We will now print the name of these processes
7302 when we send SIGKILL to them, since this usually indicates a
7303 problem.
7304
7305 * If /etc/crypttab refers to password files stored on
7306 configured mount points automatic dependencies will now be
7307 generated to ensure the specific mount is established first
7308 before the key file is attempted to be read.
7309
7310 * 'systemctl status' will now show information about the
7311 network sockets a socket unit is listening on.
7312
7313 * 'systemctl status' will also shown information about any
7314 drop-in configuration file for units. (Drop-In configuration
7315 files in this context are files such as
7316 /etc/systemd/systemd/foobar.service.d/*.conf)
7317
7318 * systemd-cgtop now optionally shows summed up CPU times of
7319 cgroups. Press '%' while running cgtop to switch between
7320 percentage and absolute mode. This is useful to determine
7321 which cgroups use up the most CPU time over the entire
7322 runtime of the system. systemd-cgtop has also been updated
7323 to be 'pipeable' for processing with further shell tools.
7324
7325 * 'hostnamectl set-hostname' will now allow setting of FQDN
7326 hostnames.
7327
7328 * The formatting and parsing of time span values has been
7329 changed. The parser now understands fractional expressions
7330 such as "5.5h". The formatter will now output fractional
7331 expressions for all time spans under 1min, i.e. "5.123456s"
7332 rather than "5s 123ms 456us". For time spans under 1s
7333 millisecond values are shown, for those under 1ms
7334 microsecond values are shown. This should greatly improve
7335 all time-related output of systemd.
7336
7337 * libsystemd-login and libsystemd-journal gained new
7338 functions for querying the poll() events mask and poll()
7339 timeout value for integration into arbitrary event
7340 loops.
7341
7342 * localectl gained the ability to list available X11 keymaps
7343 (models, layouts, variants, options).
7344
7345 * 'systemd-analyze dot' gained the ability to filter for
7346 specific units via shell-style globs, to create smaller,
d28315e4 7347 more useful graphs. I.e. it is now possible to create simple
d3a86981
LP
7348 graphs of all the dependencies between only target units, or
7349 of all units that Avahi has dependencies with.
7350
7351 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
7352 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
7353 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
7354 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
7355 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
7356 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
7357 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
7358
9ca3c17f
LP
7359CHANGES WITH 200:
7360
7361 * The boot-time readahead implementation for rotating media
7362 will now read the read-ahead data in multiple passes which
7363 consist of all read requests made in equidistant time
7364 intervals. This means instead of strictly reading read-ahead
7365 data in its physical order on disk we now try to find a
7366 middle ground between physical and access time order.
7367
7368 * /etc/os-release files gained a new BUILD_ID= field for usage
7369 on operating systems that provide continuous builds of OS
7370 images.
7371
7372 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
7373 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
7374 William Douglas, Zbigniew Jędrzejewski-Szmek
7375
35911459
LP
7376CHANGES WITH 199:
7377
7378 * systemd-python gained an API exposing libsystemd-daemon.
7379
7380 * The SMACK setup logic gained support for uploading CIPSO
7381 security policy.
7382
7383 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
7384 ReadOnlyDirectories= and InaccessibleDirectories= has
7385 changed. The private /tmp and /var/tmp directories are now
7386 shared by all processes of a service (which means
7387 ExecStartPre= may now leave data in /tmp that ExecStart= of
7388 the same service can still access). When a service is
7389 stopped its temporary directories are immediately deleted
a87197f5 7390 (normal clean-up with tmpfiles is still done in addition to
35911459
LP
7391 this though).
7392
7393 * By default, systemd will now set a couple of sysctl
7394 variables in the kernel: the safe sysrq options are turned
7395 on, IP route verification is turned on, and source routing
7396 disabled. The recently added hardlink and softlink
7397 protection of the kernel is turned on. These settings should
7398 be reasonably safe, and good defaults for all new systems.
7399
7400 * The predictable network naming logic may now be turned off
a87197f5 7401 with a new kernel command line switch: net.ifnames=0.
35911459
LP
7402
7403 * A new libsystemd-bus module has been added that implements a
7404 pretty complete D-Bus client library. For details see:
7405
56cadcb6 7406 https://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
35911459 7407
c20d8298 7408 * journald will now explicitly flush the journal files to disk
a87197f5
ZJS
7409 at the latest 5min after each write. The file will then also
7410 be marked offline until the next write. This should increase
7411 reliability in case of a crash. The synchronization delay
7412 can be configured via SyncIntervalSec= in journald.conf.
35911459
LP
7413
7414 * There's a new remote-fs-setup.target unit that can be used
7415 to pull in specific services when at least one remote file
7416 system is to be mounted.
7417
7418 * There are new targets timers.target and paths.target as
7419 canonical targets to pull user timer and path units in
7420 from. This complements sockets.target with a similar
7421 purpose for socket units.
7422
6a7d3d68
LP
7423 * libudev gained a new call udev_device_set_attribute_value()
7424 to set sysfs attributes of a device.
7425
a87197f5
ZJS
7426 * The udev daemon now sets the default number of worker
7427 processes executed in parallel based on the number of available
c20d8298 7428 CPUs instead of the amount of available RAM. This is supposed
ab06eef8 7429 to provide a more reliable default and limit a too aggressive
ce830873 7430 parallelism for setups with 1000s of devices connected.
c20d8298 7431
35911459
LP
7432 Contributions from: Auke Kok, Colin Walters, Cristian
7433 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
7434 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
7435 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
7436 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
7437 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
7438 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
7439 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
7440 Zbigniew Jędrzejewski-Szmek
7441
85d68397
LP
7442CHANGES WITH 198:
7443
7444 * Configuration of unit files may now be extended via drop-in
7445 files without having to edit/override the unit files
7446 themselves. More specifically, if the administrator wants to
7447 change one value for a service file foobar.service he can
7448 now do so by dropping in a configuration snippet into
ad88e758 7449 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
85d68397
LP
7450 will load all these snippets and apply them on top of the
7451 main unit configuration file, possibly extending or
7452 overriding its settings. Using these drop-in snippets is
40e21da8
KS
7453 generally nicer than the two earlier options for changing
7454 unit files locally: copying the files from
85d68397
LP
7455 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
7456 them there; or creating a new file in /etc/systemd/system/
7457 that incorporates the original one via ".include". Drop-in
7458 snippets into these .d/ directories can be placed in any
fd868975 7459 directory systemd looks for units in, and the usual
85d68397
LP
7460 overriding semantics between /usr/lib, /etc and /run apply
7461 for them too.
7462
7463 * Most unit file settings which take lists of items can now be
6aa8d43a 7464 reset by assigning the empty string to them. For example,
85d68397
LP
7465 normally, settings such as Environment=FOO=BAR append a new
7466 environment variable assignment to the environment block,
7467 each time they are used. By assigning Environment= the empty
7468 string the environment block can be reset to empty. This is
7469 particularly useful with the .d/*.conf drop-in snippets
156f7d09
KS
7470 mentioned above, since this adds the ability to reset list
7471 settings from vendor unit files via these drop-ins.
85d68397
LP
7472
7473 * systemctl gained a new "list-dependencies" command for
7474 listing the dependencies of a unit recursively.
7475
40e21da8 7476 * Inhibitors are now honored and listed by "systemctl
85d68397
LP
7477 suspend", "systemctl poweroff" (and similar) too, not only
7478 GNOME. These commands will also list active sessions by
7479 other users.
7480
7481 * Resource limits (as exposed by the various control group
7482 controllers) can now be controlled dynamically at runtime
7483 for all units. More specifically, you can now use a command
7484 like "systemctl set-cgroup-attr foobar.service cpu.shares
7485 2000" to alter the CPU shares a specific service gets. These
6aa8d43a 7486 settings are stored persistently on disk, and thus allow the
85d68397
LP
7487 administrator to easily adjust the resource usage of
7488 services with a few simple commands. This dynamic resource
6aa8d43a 7489 management logic is also available to other programs via the
85d68397
LP
7490 bus. Almost any kernel cgroup attribute and controller is
7491 supported.
7492
7493 * systemd-vconsole-setup will now copy all font settings to
6aa8d43a
LP
7494 all allocated VTs, where it previously applied them only to
7495 the foreground VT.
85d68397
LP
7496
7497 * libsystemd-login gained the new sd_session_get_tty() API
7498 call.
7499
6aa8d43a
LP
7500 * This release drops support for a few legacy or
7501 distribution-specific LSB facility names when parsing init
7502 scripts: $x-display-manager, $mail-transfer-agent,
85d68397
LP
7503 $mail-transport-agent, $mail-transfer-agent, $smtp,
7504 $null. Also, the mail-transfer-agent.target unit backing
7505 this has been removed. Distributions which want to retain
6aa8d43a
LP
7506 compatibility with this should carry the burden for
7507 supporting this themselves and patch support for these back
7508 in, if they really need to. Also, the facilities $syslog and
7509 $local_fs are now ignored, since systemd does not support
7510 early-boot LSB init scripts anymore, and these facilities
7511 are implied anyway for normal services. syslog.target has
7512 also been removed.
85d68397 7513
40e21da8 7514 * There are new bus calls on PID1's Manager object for
6aa8d43a 7515 cancelling jobs, and removing snapshot units. Previously,
85d68397
LP
7516 both calls were only available on the Job and Snapshot
7517 objects themselves.
7518
7519 * systemd-journal-gatewayd gained SSL support.
7520
7521 * The various "environment" files, such as /etc/locale.conf
7522 now support continuation lines with a backslash ("\") as
499b604b 7523 last character in the line, similarly in style (but different)
85d68397
LP
7524 to how this is supported in shells.
7525
7526 * For normal user processes the _SYSTEMD_USER_UNIT= field is
7527 now implicitly appended to every log entry logged. systemctl
7528 has been updated to filter by this field when operating on a
7529 user systemd instance.
7530
7531 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
7532 CAP_AUDIT_CONTROL capabilities to the capabilities set for
7533 the container. This makes it easier to boot unmodified
7534 Fedora systems in a container, which however still requires
7535 audit=0 to be passed on the kernel command line. Auditing in
7536 kernel and userspace is unfortunately still too broken in
7537 context of containers, hence we recommend compiling it out
7538 of the kernel or using audit=0. Hopefully this will be fixed
7539 one day for good in the kernel.
7540
7541 * nspawn gained the new --bind= and --bind-ro= parameters to
7542 bind mount specific directories from the host into the
7543 container.
7544
40e21da8 7545 * nspawn will now mount its own devpts file system instance
6aa8d43a 7546 into the container, in order not to leak pty devices from
85d68397
LP
7547 the host into the container.
7548
7549 * systemd will now read the firmware boot time performance
6aa8d43a
LP
7550 information from the EFI variables, if the used boot loader
7551 supports this, and takes it into account for boot performance
7552 analysis via "systemd-analyze". This is currently supported
7553 only in conjunction with Gummiboot, but could be supported
7554 by other boot loaders too. For details see:
85d68397 7555
56cadcb6 7556 https://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
85d68397
LP
7557
7558 * A new generator has been added that automatically mounts the
7559 EFI System Partition (ESP) to /boot, if that directory
6aa8d43a
LP
7560 exists, is empty, and no other file system has been
7561 configured to be mounted there.
85d68397
LP
7562
7563 * logind will now send out PrepareForSleep(false) out
7564 unconditionally, after coming back from suspend. This may be
7565 used by applications as asynchronous notification for
7566 system resume events.
7567
7568 * "systemctl unlock-sessions" has been added, that allows
7569 unlocking the screens of all user sessions at once, similar
499b604b 7570 to how "systemctl lock-sessions" already locked all users
40e21da8 7571 sessions. This is backed by a new D-Bus call UnlockSessions().
85d68397
LP
7572
7573 * "loginctl seat-status" will now show the master device of a
7574 seat. (i.e. the device of a seat that needs to be around for
7575 the seat to be considered available, usually the graphics
7576 card).
7577
7578 * tmpfiles gained a new "X" line type, that allows
7579 configuration of files and directories (with wildcards) that
7580 shall be excluded from automatic cleanup ("aging").
7581
bf933560
KS
7582 * udev default rules set the device node permissions now only
7583 at "add" events, and do not change them any longer with a
7584 later "change" event.
85d68397
LP
7585
7586 * The log messages for lid events and power/sleep keypresses
7587 now carry a message ID.
7588
7589 * We now have a substantially larger unit test suite, but this
7590 continues to be work in progress.
7591
7592 * udevadm hwdb gained a new --root= parameter to change the
7593 root directory to operate relative to.
7594
40e21da8
KS
7595 * logind will now issue a background sync() request to the kernel
7596 early at shutdown, so that dirty buffers are flushed to disk early
85d68397
LP
7597 instead of at the last moment, in order to optimize shutdown
7598 times a little.
7599
7600 * A new bootctl tool has been added that is an interface for
7601 certain boot loader operations. This is currently a preview
7602 and is likely to be extended into a small mechanism daemon
7603 like timedated, localed, hostnamed, and can be used by
7604 graphical UIs to enumerate available boot options, and
7605 request boot into firmware operations.
7606
7607 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
7608 the rest of the package. It also has been updated to work
7609 correctly in initrds.
7610
d35f51ea
ZJS
7611 * polkit previously has been runtime optional, and is now also
7612 compile time optional via a configure switch.
85d68397
LP
7613
7614 * systemd-analyze has been reimplemented in C. Also "systemctl
7615 dot" has moved into systemd-analyze.
7616
7617 * "systemctl status" with no further parameters will now print
7618 the status of all active or failed units.
7619
7620 * Operations such as "systemctl start" can now be executed
7621 with a new mode "--irreversible" which may be used to queue
7622 operations that cannot accidentally be reversed by a later
6aa8d43a 7623 job queuing. This is by default used to make shutdown
85d68397
LP
7624 requests more robust.
7625
7626 * The Python API of systemd now gained a new module for
7627 reading journal files.
7628
7629 * A new tool kernel-install has been added that can install
7630 kernel images according to the Boot Loader Specification:
7631
56cadcb6 7632 https://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
85d68397
LP
7633
7634 * Boot time console output has been improved to provide
6aa8d43a 7635 animated boot time output for hanging jobs.
85d68397
LP
7636
7637 * A new tool systemd-activate has been added which can be used
7638 to test socket activation with, directly from the command
7639 line. This should make it much easier to test and debug
7640 socket activation in daemons.
7641
7642 * journalctl gained a new "--reverse" (or -r) option to show
7643 journal output in reverse order (i.e. newest line first).
7644
43447fb7
LP
7645 * journalctl gained a new "--pager-end" (or -e) option to jump
7646 to immediately jump to the end of the journal in the
7647 pager. This is only supported in conjunction with "less".
7648
85d68397 7649 * journalctl gained a new "--user-unit=" option, that works
499b604b 7650 similarly to "--unit=" but filters for user units rather than
85d68397
LP
7651 system units.
7652
7653 * A number of unit files to ease adoption of systemd in
7654 initrds has been added. This moves some minimal logic from
7655 the various initrd implementations into systemd proper.
7656
7657 * The journal files are now owned by a new group
7658 "systemd-journal", which exists specifically to allow access
7659 to the journal, and nothing else. Previously, we used the
6aa8d43a 7660 "adm" group for that, which however possibly covers more
85d68397
LP
7661 than just journal/log file access. This new group is now
7662 already used by systemd-journal-gatewayd to ensure this
7663 daemon gets access to the journal files and as little else
7664 as possible. Note that "make install" will also set FS ACLs
7665 up for /var/log/journal to give "adm" and "wheel" read
7666 access to it, in addition to "systemd-journal" which owns
7667 the journal files. We recommend that packaging scripts also
6aa8d43a 7668 add read access to "adm" + "wheel" to /var/log/journal, and
85d68397
LP
7669 all existing/future journal files. To normal users and
7670 administrators little changes, however packagers need to
7671 ensure to create the "systemd-journal" system group at
7672 package installation time.
7673
7674 * The systemd-journal-gatewayd now runs as unprivileged user
7675 systemd-journal-gateway:systemd-journal-gateway. Packaging
7676 scripts need to create these system user/group at
7677 installation time.
7678
7679 * timedated now exposes a new boolean property CanNTP that
7680 indicates whether a local NTP service is available or not.
7681
7682 * systemd-detect-virt will now also detect xen PVs
7683
40e21da8
KS
7684 * The pstore file system is now mounted by default, if it is
7685 available.
85d68397 7686
1aed4590
LP
7687 * In addition to the SELinux and IMA policies we will now also
7688 load SMACK policies at early boot.
7689
85d68397
LP
7690 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
7691 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
7692 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
7693 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
7694 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
7695 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
7696 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
7697 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
7698 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
7699 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
7700 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
7701 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
7702 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
7703 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
7704
8ad26859
LP
7705CHANGES WITH 197:
7706
7707 * Timer units now support calendar time events in addition to
7708 monotonic time events. That means you can now trigger a unit
7709 based on a calendar time specification such as "Thu,Fri
7710 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
7711 or fifth day of any month of the year 2013, given that it is
7712 a thursday or friday. This brings timer event support
7713 considerably closer to cron's capabilities. For details on
7714 the supported calendar time specification language see
7715 systemd.time(7).
7716
7717 * udev now supports a number of different naming policies for
7718 network interfaces for predictable names, and a combination
7719 of these policies is now the default. Please see this wiki
7720 document for details:
7721
56cadcb6 7722 https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
8ad26859
LP
7723
7724 * Auke Kok's bootchart implementation has been added to the
d28315e4
JE
7725 systemd tree. It is an optional component that can graph the
7726 boot in quite some detail. It is one of the best bootchart
8ad26859
LP
7727 implementations around and minimal in its code and
7728 dependencies.
7729
7730 * nss-myhostname has been integrated into the systemd source
7731 tree. nss-myhostname guarantees that the local hostname
7732 always stays resolvable via NSS. It has been a weak
7733 requirement of systemd-hostnamed since a long time, and
7734 since its code is actually trivial we decided to just
7735 include it in systemd's source tree. It can be turned off
7736 with a configure switch.
7737
7738 * The read-ahead logic is now capable of properly detecting
7739 whether a btrfs file system is on SSD or rotating media, in
7740 order to optimize the read-ahead scheme. Previously, it was
7741 only capable of detecting this on traditional file systems
7742 such as ext4.
7743
7744 * In udev, additional device properties are now read from the
7745 IAB in addition to the OUI database. Also, Bluetooth company
7746 identities are attached to the devices as well.
7747
7748 * In service files %U may be used as specifier that is
7749 replaced by the configured user name of the service.
7750
7751 * nspawn may now be invoked without a controlling TTY. This
7752 makes it suitable for invocation as its own service. This
7753 may be used to set up a simple containerized server system
7754 using only core OS tools.
7755
7756 * systemd and nspawn can now accept socket file descriptors
7757 when they are started for socket activation. This enables
7758 implementation of socket activated nspawn
7759 containers. i.e. think about autospawning an entire OS image
7760 when the first SSH or HTTP connection is received. We expect
7761 that similar functionality will also be added to libvirt-lxc
7762 eventually.
7763
7764 * journalctl will now suppress ANSI color codes when
7765 presenting log data.
7766
7767 * systemctl will no longer show control group information for
ce830873 7768 a unit if the control group is empty anyway.
8ad26859
LP
7769
7770 * logind can now automatically suspend/hibernate/shutdown the
7771 system on idle.
7772
7773 * /etc/machine-info and hostnamed now also expose the chassis
7774 type of the system. This can be used to determine whether
7775 the local system is a laptop, desktop, handset or
7776 tablet. This information may either be configured by the
7777 user/vendor or is automatically determined from ACPI and DMI
7778 information if possible.
7779
d35f51ea
ZJS
7780 * A number of polkit actions are now bound together with "imply"
7781 rules. This should simplify creating UIs because many actions
7782 will now authenticate similar ones as well.
8ad26859
LP
7783
7784 * Unit files learnt a new condition ConditionACPower= which
7785 may be used to conditionalize a unit depending on whether an
7786 AC power source is connected or not, of whether the system
7787 is running on battery power.
7788
7789 * systemctl gained a new "is-failed" verb that may be used in
7790 shell scripts and suchlike to check whether a specific unit
7791 is in the "failed" state.
7792
7793 * The EnvironmentFile= setting in unit files now supports file
7794 globbing, and can hence be used to easily read a number of
7795 environment files at once.
7796
7797 * systemd will no longer detect and recognize specific
7798 distributions. All distribution-specific #ifdeffery has been
7799 removed, systemd is now fully generic and
7800 distribution-agnostic. Effectively, not too much is lost as
7801 a lot of the code is still accessible via explicit configure
7802 switches. However, support for some distribution specific
7803 legacy configuration file formats has been dropped. We
7804 recommend distributions to simply adopt the configuration
7805 files everybody else uses now and convert the old
7806 configuration from packaging scripts. Most distributions
7807 already did that. If that's not possible or desirable,
7808 distributions are welcome to forward port the specific
7809 pieces of code locally from the git history.
7810
7811 * When logging a message about a unit systemd will now always
7812 log the unit name in the message meta data.
7813
7814 * localectl will now also discover system locale data that is
7815 not stored in locale archives, but directly unpacked.
7816
7817 * logind will no longer unconditionally use framebuffer
7818 devices as seat masters, i.e. as devices that are required
7819 to be existing before a seat is considered preset. Instead,
7820 it will now look for all devices that are tagged as
b938cb90
JE
7821 "seat-master" in udev. By default, framebuffer devices will
7822 be marked as such, but depending on local systems, other
8ad26859
LP
7823 devices might be marked as well. This may be used to
7824 integrate graphics cards using closed source drivers (such
7825 as NVidia ones) more nicely into logind. Note however, that
7826 we recommend using the open source NVidia drivers instead,
7827 and no udev rules for the closed-source drivers will be
7828 shipped from us upstream.
7829
7830 Contributions from: Adam Williamson, Alessandro Crismani, Auke
7831 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
7832 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
7833 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
7834 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
7835 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
7836 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
7837 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
7838 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
7839 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
7840 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
7841 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
7842 Jędrzejewski-Szmek
7843
0428ddb7
LP
7844CHANGES WITH 196:
7845
7846 * udev gained support for loading additional device properties
7847 from an indexed database that is keyed by vendor/product IDs
7848 and similar device identifiers. For the beginning this
7849 "hwdb" is populated with data from the well-known PCI and
7850 USB database, but also includes PNP, ACPI and OID data. In
7851 the longer run this indexed database shall grow into
7852 becoming the one central database for non-essential
7853 userspace device metadata. Previously, data from the PCI/USB
96ec33c0 7854 database was only attached to select devices, since the
0428ddb7 7855 lookup was a relatively expensive operation due to O(n) time
96ec33c0
LP
7856 complexity (with n being the number of entries in the
7857 database). Since this is now O(1), we decided to add in this
7858 data for all devices where this is available, by
0428ddb7
LP
7859 default. Note that the indexed database needs to be rebuilt
7860 when new data files are installed. To achieve this you need
7861 to update your packaging scripts to invoke "udevadm hwdb
7862 --update" after installation of hwdb data files. For
7863 RPM-based distributions we introduced the new
7864 %udev_hwdb_update macro for this purpose.
7865
7866 * The Journal gained support for the "Message Catalog", an
7867 indexed database to link up additional information with
7868 journal entries. For further details please check:
7869
56cadcb6 7870 https://www.freedesktop.org/wiki/Software/systemd/catalog
0428ddb7
LP
7871
7872 The indexed message catalog database also needs to be
7873 rebuilt after installation of message catalog files. Use
7874 "journalctl --update-catalog" for this. For RPM-based
7875 distributions we introduced the %journal_catalog_update
7876 macro for this purpose.
7877
7878 * The Python Journal bindings gained support for the standard
7879 Python logging framework.
7880
7881 * The Journal API gained new functions for checking whether
7882 the underlying file system of a journal file is capable of
7883 properly reporting file change notifications, or whether
7884 applications that want to reflect journal changes "live"
ab06eef8 7885 need to recheck journal files continuously in appropriate
0428ddb7
LP
7886 time intervals.
7887
7888 * It is now possible to set the "age" field for tmpfiles
7889 entries to 0, indicating that files matching this entry
7890 shall always be removed when the directories are cleaned up.
7891
7892 * coredumpctl gained a new "gdb" verb which invokes gdb
7893 right-away on the selected coredump.
7894
7895 * There's now support for "hybrid sleep" on kernels that
7896 support this, in addition to "suspend" and "hibernate". Use
7897 "systemctl hybrid-sleep" to make use of this.
7898
7899 * logind's HandleSuspendKey= setting (and related settings)
7900 now gained support for a new "lock" setting to simply
7901 request the screen lock on all local sessions, instead of
7902 actually executing a suspend or hibernation.
7903
7904 * systemd will now mount the EFI variables file system by
7905 default.
7906
7907 * Socket units now gained support for configuration of the
7908 SMACK security label.
7909
7910 * timedatectl will now output the time of the last and next
7911 daylight saving change.
7912
7913 * We dropped support for various legacy and distro-specific
7914 concepts, such as insserv, early-boot SysV services
7915 (i.e. those for non-standard runlevels such as 'b' or 'S')
7916 or ArchLinux /etc/rc.conf support. We recommend the
7917 distributions who still need support this to either continue
7918 to maintain the necessary patches downstream, or find a
7919 different solution. (Talk to us if you have questions!)
7920
d35f51ea
ZJS
7921 * Various systemd components will now bypass polkit checks for
7922 root and otherwise handle properly if polkit is not found to
7923 be around. This should fix most issues for polkit-less
7924 systems. Quite frankly this should have been this way since
7925 day one. It is absolutely our intention to make systemd work
7926 fine on polkit-less systems, and we consider it a bug if
7927 something does not work as it should if polkit is not around.
0428ddb7
LP
7928
7929 * For embedded systems it is now possible to build udev and
7930 systemd without blkid and/or kmod support.
7931
7932 * "systemctl switch-root" is now capable of switching root
7933 more than once. I.e. in addition to transitions from the
7934 initrd to the host OS it is now possible to transition to
7935 further OS images from the host. This is useful to implement
7936 offline updating tools.
7937
7938 * Various other additions have been made to the RPM macros
7939 shipped with systemd. Use %udev_rules_update() after
7940 installing new udev rules files. %_udevhwdbdir,
7941 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
7942 %_sysctldir are now available which resolve to the right
7943 directories for packages to place various data files in.
7944
7945 * journalctl gained the new --full switch (in addition to
7946 --all, to disable ellipsation for long messages.
7947
7948 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
7949 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
7950 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
7951 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
7952 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
7953 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
7954 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
7955 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
7956 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
7957
139ee8cc
LP
7958CHANGES WITH 195:
7959
6827101a 7960 * journalctl gained new --since= and --until= switches to
139ee8cc
LP
7961 filter by time. It also now supports nice filtering for
7962 units via --unit=/-u.
7963
6827101a 7964 * Type=oneshot services may use ExecReload= and do the
139ee8cc
LP
7965 right thing.
7966
7967 * The journal daemon now supports time-based rotation and
7968 vacuuming, in addition to the usual disk-space based
7969 rotation.
7970
7971 * The journal will now index the available field values for
7972 each field name. This enables clients to show pretty drop
7973 downs of available match values when filtering. The bash
7974 completion of journalctl has been updated
7975 accordingly. journalctl gained a new switch -F to list all
7976 values a certain field takes in the journal database.
7977
7978 * More service events are now written as structured messages
7979 to the journal, and made recognizable via message IDs.
7980
7981 * The timedated, localed and hostnamed mini-services which
7982 previously only provided support for changing time, locale
7983 and hostname settings from graphical DEs such as GNOME now
7984 also have a minimal (but very useful) text-based client
7985 utility each. This is probably the nicest way to changing
7986 these settings from the command line now, especially since
7987 it lists available options and is fully integrated with bash
7988 completion.
7989
7990 * There's now a new tool "systemd-coredumpctl" to list and
7991 extract coredumps from the journal.
7992
7993 * We now install a README each in /var/log/ and
7994 /etc/rc.d/init.d explaining where the system logs and init
7995 scripts went. This hopefully should help folks who go to
7996 that dirs and look into the otherwise now empty void and
7997 scratch their heads.
7998
7999 * When user-services are invoked (by systemd --user) the
8000 $MANAGERPID env var is set to the PID of systemd.
8001
8002 * SIGRTMIN+24 when sent to a --user instance will now result
8003 in immediate termination of systemd.
8004
8005 * gatewayd received numerous feature additions such as a
8006 "follow" mode, for live syncing and filtering.
8007
8008 * browse.html now allows filtering and showing detailed
8009 information on specific entries. Keyboard navigation and
8010 mouse screen support has been added.
8011
8012 * gatewayd/journalctl now supports HTML5/JSON
8013 Server-Sent-Events as output.
8014
1cb88f2c 8015 * The SysV init script compatibility logic will now
139ee8cc
LP
8016 heuristically determine whether a script supports the
8017 "reload" verb, and only then make this available as
8018 "systemctl reload".
8019
15f47220 8020 * "systemctl status --follow" has been removed, use "journalctl
139ee8cc
LP
8021 -u" instead.
8022
8023 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
8024 have been removed since they are hardly useful to be
8025 configured.
8026
8027 * And I'd like to take the opportunity to specifically mention
8028 Zbigniew for his great contributions. Zbigniew, you rock!
8029
8030 Contributions from: Andrew Eikum, Christian Hesse, Colin
8031 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
4d92e078
LP
8032 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
8033 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
8034 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
8035 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
8036 Jędrzejewski-Szmek, Сковорода Никита Андреевич
139ee8cc 8037
f9b55720
LP
8038CHANGES WITH 194:
8039
8040 * If /etc/vconsole.conf is non-existent or empty we will no
8041 longer load any console font or key map at boot by
8042 default. Instead the kernel defaults will be left
8043 intact. This is definitely the right thing to do, as no
8044 configuration should mean no configuration, and hard-coding
8045 font names that are different on all archs is probably a bad
8046 idea. Also, the kernel default key map and font should be
8047 good enough for most cases anyway, and mostly identical to
8048 the userspace fonts/key maps we previously overloaded them
8049 with. If distributions want to continue to default to a
8050 non-kernel font or key map they should ship a default
8051 /etc/vconsole.conf with the appropriate contents.
8052
8053 Contributions from: Colin Walters, Daniel J Walsh, Dave
8054 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
8055 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
8056
597c52cf
LP
8057CHANGES WITH 193:
8058
8059 * journalctl gained a new --cursor= switch to show entries
8060 starting from the specified location in the journal.
8061
8062 * We now enforce a size limit on journal entry fields exported
8063 with "-o json" in journalctl. Fields larger than 4K will be
8064 assigned null. This can be turned off with --all.
8065
8066 * An (optional) journal gateway daemon is now available as
8067 "systemd-journal-gatewayd.service". This service provides
8068 access to the journal via HTTP and JSON. This functionality
8069 will be used to implement live log synchronization in both
8070 pull and push modes, but has various other users too, such
8071 as easy log access for debugging of embedded devices. Right
8072 now it is already useful to retrieve the journal via HTTP:
8073
8074 # systemctl start systemd-journal-gatewayd.service
8075 # wget http://localhost:19531/entries
8076
8077 This will download the journal contents in a
8078 /var/log/messages compatible format. The same as JSON:
8079
8080 # curl -H"Accept: application/json" http://localhost:19531/entries
8081
8082 This service is also accessible via a web browser where a
8083 single static HTML5 app is served that uses the JSON logic
8084 to enable the user to do some basic browsing of the
8085 journal. This will be extended later on. Here's an example
8086 screenshot of this app in its current state:
8087
8088 http://0pointer.de/public/journal-gatewayd
8089
8090 Contributions from: Kay Sievers, Lennart Poettering, Robert
8091 Milasan, Tom Gundersen
8092
075d4ecb
LP
8093CHANGES WITH 192:
8094
8095 * The bash completion logic is now available for journalctl
8096 too.
8097
d28315e4 8098 * We do not mount the "cpuset" controller anymore together with
075d4ecb
LP
8099 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
8100 started if no parameters are assigned to it. "cpuset" hence
61233823 8101 broke code that assumed it could create "cpu" groups and
075d4ecb
LP
8102 just start them.
8103
8104 * journalctl -f will now subscribe to terminal size changes,
8105 and line break accordingly.
8106
597c52cf
LP
8107 Contributions from: Dave Reisner, Kay Sievers, Lennart
8108 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
075d4ecb 8109
b6a86739
LP
8110CHANGES WITH 191:
8111
8112 * nspawn will now create a symlink /etc/localtime in the
8113 container environment, copying the host's timezone
8114 setting. Previously this has been done via a bind mount, but
8115 since symlinks cannot be bind mounted this has now been
8116 changed to create/update the appropriate symlink.
8117
8118 * journalctl -n's line number argument is now optional, and
8119 will default to 10 if omitted.
8120
8121 * journald will now log the maximum size the journal files may
8122 take up on disk. This is particularly useful if the default
8123 built-in logic of determining this parameter from the file
8124 system size is used. Use "systemctl status
6563b535 8125 systemd-journald.service" to see this information.
b6a86739
LP
8126
8127 * The multi-seat X wrapper tool has been stripped down. As X
8128 is now capable of enumerating graphics devices via udev in a
8129 seat-aware way the wrapper is not strictly necessary
8130 anymore. A stripped down temporary stop-gap is still shipped
8131 until the upstream display managers have been updated to
8132 fully support the new X logic. Expect this wrapper to be
6563b535 8133 removed entirely in one of the next releases.
b6a86739
LP
8134
8135 * HandleSleepKey= in logind.conf has been split up into
8136 HandleSuspendKey= and HandleHibernateKey=. The old setting
6563b535 8137 is not available anymore. X11 and the kernel are
45afd519 8138 distinguishing between these keys and we should too. This
b6a86739
LP
8139 also means the inhibition lock for these keys has been split
8140 into two.
8141
597c52cf
LP
8142 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
8143 Poettering, Lukas Nykryn, Václav Pavlín
b6a86739 8144
0c11f949
LP
8145CHANGES WITH 190:
8146
d28315e4 8147 * Whenever a unit changes state we will now log this to the
0c11f949
LP
8148 journal and show along the unit's own log output in
8149 "systemctl status".
8150
8151 * ConditionPathIsMountPoint= can now properly detect bind
8152 mount points too. (Previously, a bind mount of one file
8d0256b7 8153 system to another place in the same file system could not be
0c11f949
LP
8154 detected as mount, since they shared struct stat's st_dev
8155 field.)
8156
8157 * We will now mount the cgroup controllers cpu, cpuacct,
8158 cpuset and the controllers net_cls, net_prio together by
8159 default.
8160
8161 * nspawn containers will now have a virtualized boot
8162 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
8163 over with a randomized ID at container initialization). This
8164 has the effect of making "journalctl -b" do the right thing
8165 in a container.
8166
8167 * The JSON output journal serialization has been updated not
8168 to generate "endless" list objects anymore, but rather one
8169 JSON object per line. This is more in line how most JSON
8170 parsers expect JSON objects. The new output mode
8171 "json-pretty" has been added to provide similar output, but
8172 neatly aligned for readability by humans.
8173
8174 * We dropped all explicit sync() invocations in the shutdown
8175 code. The kernel does this implicitly anyway in the kernel
8176 reboot() syscall. halt(8)'s -n option is now a compatibility
8177 no-op.
8178
8179 * We now support virtualized reboot() in containers, as
8180 supported by newer kernels. We will fall back to exit() if
8181 CAP_SYS_REBOOT is not available to the container. Also,
8182 nspawn makes use of this now and will actually reboot the
8183 container if the containerized OS asks for that.
8184
8185 * journalctl will only show local log output by default
8186 now. Use --merge (-m) to show remote log output, too.
8187
8188 * libsystemd-journal gained the new sd_journal_get_usage()
8189 call to determine the current disk usage of all journal
8190 files. This is exposed in the new "journalctl --disk-usage"
8191 command.
8192
8193 * journald gained a new configuration setting SplitMode= in
8194 journald.conf which may be used to control how user journals
8195 are split off. See journald.conf(5) for details.
8196
8197 * A new condition type ConditionFileNotEmpty= has been added.
8198
8199 * tmpfiles' "w" lines now support file globbing, to write
8200 multiple files at once.
8201
8202 * We added Python bindings for the journal submission
8203 APIs. More Python APIs for a number of selected APIs will
8204 likely follow. Note that we intend to add native bindings
8205 only for the Python language, as we consider it common
8206 enough to deserve bindings shipped within systemd. There are
8207 various projects outside of systemd that provide bindings
8208 for languages such as PHP or Lua.
8209
a98d5d64
LP
8210 * Many conditions will now resolve specifiers such as %i. In
8211 addition, PathChanged= and related directives of .path units
8212 now support specifiers as well.
0c11f949
LP
8213
8214 * There's now a new RPM macro definition for the system preset
8215 dir: %_presetdir.
8216
d28315e4 8217 * journald will now warn if it ca not forward a message to the
dca348bc 8218 syslog daemon because its socket is full.
0c11f949
LP
8219
8220 * timedated will no longer write or process /etc/timezone,
8221 except on Debian. As we do not support late mounted /usr
8222 anymore /etc/localtime always being a symlink is now safe,
8223 and hence the information in /etc/timezone is not necessary
8224 anymore.
8225
aaccc32c 8226 * logind will now always reserve one VT for a text getty (VT6
0c11f949
LP
8227 by default). Previously if more than 6 X sessions where
8228 started they took up all the VTs with auto-spawned gettys,
8229 so that no text gettys were available anymore.
8230
8231 * udev will now automatically inform the btrfs kernel logic
8232 about btrfs RAID components showing up. This should make
8233 simple hotplug based btrfs RAID assembly work.
8234
8235 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
8236 (but not for its children which will stay at the kernel
8237 default). This should allow setups with a lot more listening
8238 sockets.
8239
8240 * systemd will now always pass the configured timezone to the
8241 kernel at boot. timedated will do the same when the timezone
8242 is changed.
8243
8244 * logind's inhibition logic has been updated. By default,
8245 logind will now handle the lid switch, the power and sleep
8246 keys all the time, even in graphical sessions. If DEs want
8247 to handle these events on their own they should take the new
8248 handle-power-key, handle-sleep-key and handle-lid-switch
f131770b 8249 inhibitors during their runtime. A simple way to achieve
0c11f949
LP
8250 that is to invoke the DE wrapped in an invocation of:
8251
8252 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
8253
8254 * Access to unit operations is now checked via SELinux taking
8255 the unit file label and client process label into account.
8256
aad803af
LP
8257 * systemd will now notify the administrator in the journal
8258 when he over-mounts a non-empty directory.
8259
8260 * There are new specifiers that are resolved in unit files,
8261 for the host name (%H), the machine ID (%m) and the boot ID
8262 (%b).
8263
b6a86739 8264 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
0c11f949
LP
8265 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
8266 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
8267 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
8268 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
8269 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
8270 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
8271
38a60d71
LP
8272CHANGES WITH 189:
8273
8274 * Support for reading structured kernel messages from
8275 /dev/kmsg has now been added and is enabled by default.
8276
8277 * Support for reading kernel messages from /proc/kmsg has now
8278 been removed. If you want kernel messages in the journal
8279 make sure to run a recent kernel (>= 3.5) that supports
8280 reading structured messages from /dev/kmsg (see
8281 above). /proc/kmsg is now exclusive property of classic
8282 syslog daemons again.
8283
8284 * The libudev API gained the new
8285 udev_device_new_from_device_id() call.
8286
8287 * The logic for file system namespace (ReadOnlyDirectory=,
8288 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
8289 require pivot_root() anymore. This means fewer temporary
8290 directories are created below /tmp for this feature.
8291
8292 * nspawn containers will now see and receive all submounts
8293 made on the host OS below the root file system of the
8294 container.
8295
8296 * Forward Secure Sealing is now supported for Journal files,
8297 which provide cryptographical sealing of journal files so
8298 that attackers cannot alter log history anymore without this
8299 being detectable. Lennart will soon post a blog story about
8300 this explaining it in more detail.
8301
8302 * There are two new service settings RestartPreventExitStatus=
8303 and SuccessExitStatus= which allow configuration of exit
8304 status (exit code or signal) which will be excepted from the
8305 restart logic, resp. consider successful.
8306
8307 * journalctl gained the new --verify switch that can be used
8308 to check the integrity of the structure of journal files and
8309 (if Forward Secure Sealing is enabled) the contents of
8310 journal files.
8311
8312 * nspawn containers will now be run with /dev/stdin, /dev/fd/
8313 and similar symlinks pre-created. This makes running shells
8314 as container init process a lot more fun.
8315
8316 * The fstab support can now handle PARTUUID= and PARTLABEL=
8317 entries.
8318
8319 * A new ConditionHost= condition has been added to match
8320 against the hostname (with globs) and machine ID. This is
8321 useful for clusters where a single OS image is used to
8322 provision a large number of hosts which shall run slightly
8323 different sets of services.
8324
8325 * Services which hit the restart limit will now be placed in a
8326 failure state.
8327
b6a86739 8328 Contributions from: Bertram Poettering, Dave Reisner, Huang
38a60d71
LP
8329 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
8330 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
8331
c269cec3
LP
8332CHANGES WITH 188:
8333
8334 * When running in --user mode systemd will now become a
8335 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
8336 tree a lot more organized.
8337
8338 * A new PartOf= unit dependency type has been introduced that
8339 may be used to group services in a natural way.
8340
8341 * "systemctl enable" may now be used to enable instances of
8342 services.
8343
8344 * journalctl now prints error log levels in red, and
8345 warning/notice log levels in bright white. It also supports
8346 filtering by log level now.
8347
8348 * cgtop gained a new -n switch (similar to top), to configure
8349 the maximum number of iterations to run for. It also gained
8350 -b, to run in batch mode (accepting no input).
8351
ab06eef8 8352 * The suffix ".service" may now be omitted on most systemctl
c269cec3
LP
8353 command lines involving service unit names.
8354
8355 * There's a new bus call in logind to lock all sessions, as
8356 well as a loginctl verb for it "lock-sessions".
8357
8358 * libsystemd-logind.so gained a new call sd_journal_perror()
8359 that works similar to libc perror() but logs to the journal
8360 and encodes structured information about the error number.
8361
8362 * /etc/crypttab entries now understand the new keyfile-size=
8363 option.
8364
8365 * shutdown(8) now can send a (configurable) wall message when
8366 a shutdown is cancelled.
8367
8368 * The mount propagation mode for the root file system will now
8369 default to "shared", which is useful to make containers work
8370 nicely out-of-the-box so that they receive new mounts from
8371 the host. This can be undone locally by running "mount
8372 --make-rprivate /" if needed.
8373
8374 * The prefdm.service file has been removed. Distributions
8375 should maintain this unit downstream if they intend to keep
8376 it around. However, we recommend writing normal unit files
8377 for display managers instead.
8378
8379 * Since systemd is a crucial part of the OS we will now
8380 default to a number of compiler switches that improve
8381 security (hardening) such as read-only relocations, stack
8382 protection, and suchlike.
8383
8384 * The TimeoutSec= setting for services is now split into
8385 TimeoutStartSec= and TimeoutStopSec= to allow configuration
8386 of individual time outs for the start and the stop phase of
8387 the service.
8388
8389 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
8390 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
8391 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
8392 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
8393 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
8394 Gundersen, Zbigniew Jędrzejewski-Szmek
8395
c4f1b862
LP
8396CHANGES WITH 187:
8397
8398 * The journal and id128 C APIs are now fully documented as man
8399 pages.
8400
8401 * Extra safety checks have been added when transitioning from
8402 the initial RAM disk to the main system to avoid accidental
8403 data loss.
8404
c269cec3 8405 * /etc/crypttab entries now understand the new keyfile-offset=
c4f1b862
LP
8406 option.
8407
8408 * systemctl -t can now be used to filter by unit load state.
8409
8410 * The journal C API gained the new sd_journal_wait() call to
8411 make writing synchronous journal clients easier.
8412
8413 * journalctl gained the new -D switch to show journals from a
8414 specific directory.
8415
8416 * journalctl now displays a special marker between log
8417 messages of two different boots.
8418
8419 * The journal is now explicitly flushed to /var via a service
8420 systemd-journal-flush.service, rather than implicitly simply
8421 by seeing /var/log/journal to be writable.
8422
8423 * journalctl (and the journal C APIs) can now match for much
8424 more complex expressions, with alternatives and
8425 disjunctions.
8426
8427 * When transitioning from the initial RAM disk to the main
8428 system we will now kill all processes in a killing spree to
8429 ensure no processes stay around by accident.
8430
8431 * Three new specifiers may be used in unit files: %u, %h, %s
8432 resolve to the user name, user home directory resp. user
8433 shell. This is useful for running systemd user instances.
8434
8435 * We now automatically rotate journal files if their data
8436 object hash table gets a fill level > 75%. We also size the
8437 hash table based on the configured maximum file size. This
8438 together should lower hash collisions drastically and thus
8439 speed things up a bit.
8440
8441 * journalctl gained the new "--header" switch to introspect
8442 header data of journal files.
8443
8444 * A new setting SystemCallFilters= has been added to services
8445 which may be used to apply blacklists or whitelists to
8446 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
8447
8448 * nspawn gained a new --link-journal= switch (and quicker: -j)
8449 to link the container journal with the host. This makes it
8450 very easy to centralize log viewing on the host for all
8451 guests while still keeping the journal files separated.
8452
8453 * Many bugfixes and optimizations
8454
8455 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
8456 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
8457 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
8458 Jędrzejewski-Szmek
8459
b5b4c94a
LP
8460CHANGES WITH 186:
8461
8462 * Several tools now understand kernel command line arguments,
8463 which are only read when run in an initial RAM disk. They
8464 usually follow closely their normal counterparts, but are
8465 prefixed with rd.
8466
8467 * There's a new tool to analyze the readahead files that are
8468 automatically generated at boot. Use:
8469
8470 /usr/lib/systemd/systemd-readahead analyze /.readahead
8471
8472 * We now provide an early debug shell on tty9 if this enabled. Use:
8473
d1f9edaf 8474 systemctl enable debug-shell.service
b5b4c94a
LP
8475
8476 * All plymouth related units have been moved into the Plymouth
8477 package. Please make sure to upgrade your Plymouth version
8478 as well.
8479
8480 * systemd-tmpfiles now supports getting passed the basename of
8481 a configuration file only, in which case it will look for it
8482 in all appropriate directories automatically.
8483
8484 * udevadm info now takes a /dev or /sys path as argument, and
8485 does the right thing. Example:
8486
8487 udevadm info /dev/sda
8488 udevadm info /sys/class/block/sda
8489
8490 * systemctl now prints a warning if a unit is stopped but a
8491 unit that might trigger it continues to run. Example: a
8492 service is stopped but the socket that activates it is left
8493 running.
8494
8495 * "systemctl status" will now mention if the log output was
8496 shortened due to rotation since a service has been started.
8497
8498 * The journal API now exposes functions to determine the
8499 "cutoff" times due to rotation.
8500
8501 * journald now understands SIGUSR1 and SIGUSR2 for triggering
8502 immediately flushing of runtime logs to /var if possible,
8503 resp. for triggering immediate rotation of the journal
8504 files.
8505
8506 * It is now considered an error if a service is attempted to
8507 be stopped that is not loaded.
8508
8509 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
8510
8511 * systemd-analyze now supports Python 3
8512
8513 * tmpfiles now supports cleaning up directories via aging
8514 where the first level dirs are always kept around but
8515 directories beneath it automatically aged. This is enabled
8516 by prefixing the age field with '~'.
8517
8518 * Seat objects now expose CanGraphical, CanTTY properties
8519 which is required to deal with very fast bootups where the
8520 display manager might be running before the graphics drivers
8521 completed initialization.
8522
8523 * Seat objects now expose a State property.
8524
8525 * We now include RPM macros for service enabling/disabling
8526 based on the preset logic. We recommend RPM based
8527 distributions to make use of these macros if possible. This
8528 makes it simpler to reuse RPM spec files across
8529 distributions.
8530
8531 * We now make sure that the collected systemd unit name is
8532 always valid when services log to the journal via
8533 STDOUT/STDERR.
8534
8535 * There's a new man page kernel-command-line(7) detailing all
8536 command line options we understand.
8537
8538 * The fstab generator may now be disabled at boot by passing
8539 fstab=0 on the kernel command line.
8540
91ac7425 8541 * A new kernel command line option modules-load= is now understood
b5b4c94a
LP
8542 to load a specific kernel module statically, early at boot.
8543
8544 * Unit names specified on the systemctl command line are now
8545 automatically escaped as needed. Also, if file system or
8546 device paths are specified they are automatically turned
8547 into the appropriate mount or device unit names. Example:
8548
8549 systemctl status /home
8550 systemctl status /dev/sda
8551
8552 * The SysVConsole= configuration option has been removed from
8553 system.conf parsing.
8554
8555 * The SysV search path is no longer exported on the D-Bus
8556 Manager object.
8557
ce830873 8558 * The Names= option has been removed from unit file parsing.
b5b4c94a
LP
8559
8560 * There's a new man page bootup(7) detailing the boot process.
8561
8562 * Every unit and every generator we ship with systemd now
8563 comes with full documentation. The self-explanatory boot is
8564 complete.
8565
8566 * A couple of services gained "systemd-" prefixes in their
8567 name if they wrap systemd code, rather than only external
8568 code. Among them fsck@.service which is now
8569 systemd-fsck@.service.
8570
8571 * The HaveWatchdog property has been removed from the D-Bus
8572 Manager object.
8573
8574 * systemd.confirm_spawn= on the kernel command line should now
8575 work sensibly.
8576
8577 * There's a new man page crypttab(5) which details all options
8578 we actually understand.
8579
8580 * systemd-nspawn gained a new --capability= switch to pass
8581 additional capabilities to the container.
8582
8583 * timedated will now read known NTP implementation unit names
5b00c016 8584 from /usr/lib/systemd/ntp-units.d/*.list,
b5b4c94a
LP
8585 systemd-timedated-ntp.target has been removed.
8586
8587 * journalctl gained a new switch "-b" that lists log data of
8588 the current boot only.
8589
8590 * The notify socket is in the abstract namespace again, in
8591 order to support daemons which chroot() at start-up.
8592
8593 * There is a new Storage= configuration option for journald
8594 which allows configuration of where log data should go. This
8595 also provides a way to disable journal logging entirely, so
8596 that data collected is only forwarded to the console, the
8597 kernel log buffer or another syslog implementation.
8598
c4f1b862 8599 * Many bugfixes and optimizations
b5b4c94a 8600
2d938ac7
LP
8601 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
8602 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
8603 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
8604 Shawn Landden, Tom Gundersen
b5b4c94a 8605
2d197285 8606CHANGES WITH 185:
b6a86739 8607
2d197285
KS
8608 * "systemctl help <unit>" now shows the man page if one is
8609 available.
8610
8611 * Several new man pages have been added.
8612
b5b4c94a
LP
8613 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
8614 MaxLevelConsole= can now be specified in
8615 journald.conf. These options allow reducing the amount of
8616 data stored on disk or forwarded by the log level.
2d197285 8617
b5b4c94a
LP
8618 * TimerSlackNSec= can now be specified in system.conf for
8619 PID1. This allows system-wide power savings.
2d197285
KS
8620
8621 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
8622 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
8623 Matthias Clasen
8624
4c8cd173 8625CHANGES WITH 184:
b6a86739 8626
4c8cd173
LP
8627 * logind is now capable of (optionally) handling power and
8628 sleep keys as well as the lid switch.
8629
8630 * journalctl now understands the syntax "journalctl
8631 /usr/bin/avahi-daemon" to get all log output of a specific
8632 daemon.
8633
8634 * CapabilityBoundingSet= in system.conf now also influences
8635 the capability bound set of usermode helpers of the kernel.
8636
8637 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
8638 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
8639 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
8640 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
8641
ea5943d3 8642CHANGES WITH 183:
b6a86739 8643
187076d4
LP
8644 * Note that we skipped 139 releases here in order to set the
8645 new version to something that is greater than both udev's
8646 and systemd's most recent version number.
8647
194bbe33
KS
8648 * udev: all udev sources are merged into the systemd source tree now.
8649 All future udev development will happen in the systemd tree. It
8650 is still fully supported to use the udev daemon and tools without
8651 systemd running, like in initramfs or other init systems. Building
8652 udev though, will require the *build* of the systemd tree, but
ea5943d3 8653 udev can be properly *run* without systemd.
07cd4fc1 8654
91cf7e5c 8655 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
f13b388f
KS
8656 should be used to create dead device nodes as workarounds for broken
8657 subsystems.
64661ee7 8658
2d13da88
KS
8659 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
8660 no longer supported. udev_monitor_new_from_netlink() needs to be
8661 used to subscribe to events.
8662
194bbe33
KS
8663 * udev: when udevd is started by systemd, processes which are left
8664 behind by forking them off of udev rules, are unconditionally cleaned
8665 up and killed now after the event handling has finished. Services or
8666 daemons must be started as systemd services. Services can be
ea5943d3 8667 pulled-in by udev to get started, but they can no longer be directly
194bbe33
KS
8668 forked by udev rules.
8669
f13b388f
KS
8670 * udev: the daemon binary is called systemd-udevd now and installed
8671 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
8672 to adapt to that, create symlink, or rename the binary after building
8673 it.
8674
ea5943d3 8675 * libudev no longer provides these symbols:
c1959569
KS
8676 udev_monitor_from_socket()
8677 udev_queue_get_failed_list_entry()
8678 udev_get_{dev,sys,run}_path()
ea5943d3 8679 The versions number was bumped and symbol versioning introduced.
c1959569 8680
ea5943d3 8681 * systemd-loginctl and systemd-journalctl have been renamed
9ae9afce 8682 to loginctl and journalctl to match systemctl.
18b754d3
KS
8683
8684 * The config files: /etc/systemd/systemd-logind.conf and
8685 /etc/systemd/systemd-journald.conf have been renamed to
8686 logind.conf and journald.conf. Package updates should rename
8687 the files to the new names on upgrade.
8688
ea5943d3
LP
8689 * For almost all files the license is now LGPL2.1+, changed
8690 from the previous GPL2.0+. Exceptions are some minor stuff
8691 of udev (which will be changed to LGPL2.1 eventually, too),
8692 and the MIT licensed sd-daemon.[ch] library that is suitable
8693 to be used as drop-in files.
8694
8695 * systemd and logind now handle system sleep states, in
49f43d5f 8696 particular suspending and hibernating.
ea5943d3
LP
8697
8698 * logind now implements a sleep/shutdown/idle inhibiting logic
8699 suitable for a variety of uses. Soonishly Lennart will blog
8700 about this in more detail.
8701
8702 * var-run.mount and var-lock.mount are no longer provided
ce830873 8703 (which previously bind mounted these directories to their new
ea5943d3
LP
8704 places). Distributions which have not converted these
8705 directories to symlinks should consider stealing these files
8706 from git history and add them downstream.
8707
8708 * We introduced the Documentation= field for units and added
8709 this to all our shipped units. This is useful to make it
3943231c 8710 easier to explore the boot and the purpose of the various
ea5943d3
LP
8711 units.
8712
8713 * All smaller setup units (such as
8714 systemd-vconsole-setup.service) now detect properly if they
8715 are run in a container and are skipped when
8716 appropriate. This guarantees an entirely noise-free boot in
8717 Linux container environments such as systemd-nspawn.
8718
8719 * A framework for implementing offline system updates is now
8720 integrated, for details see:
c6749ba5 8721 https://www.freedesktop.org/wiki/Software/systemd/SystemUpdates
ea5943d3
LP
8722
8723 * A new service type Type=idle is available now which helps us
8724 avoiding ugly interleaving of getty output and boot status
8725 messages.
8726
439d6dfd
LP
8727 * There's now a system-wide CapabilityBoundingSet= option to
8728 globally reduce the set of capabilities for the
ea5943d3
LP
8729 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
8730 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
8731 even CAP_NET_ADMIN system-wide for secure systems.
8732
8733 * There are now system-wide DefaultLimitXXX= options to
8734 globally change the defaults of the various resource limits
8735 for all units started by PID 1.
8736
8737 * Harald Hoyer's systemd test suite has been integrated into
8738 systemd which allows easy testing of systemd builds in qemu
8739 and nspawn. (This is really awesome! Ask us for details!)
8740
3943231c
LP
8741 * The fstab parser is now implemented as generator, not inside
8742 of PID 1 anymore.
ea5943d3
LP
8743
8744 * systemctl will now warn you if .mount units generated from
8745 /etc/fstab are out of date due to changes in fstab that
d28315e4 8746 have not been read by systemd yet.
ea5943d3
LP
8747
8748 * systemd is now suitable for usage in initrds. Dracut has
8749 already been updated to make use of this. With this in place
8750 initrds get a slight bit faster but primarily are much
8751 easier to introspect and debug since "systemctl status" in
8752 the host system can be used to introspect initrd services,
8753 and the journal from the initrd is kept around too.
8754
8755 * systemd-delta has been added, a tool to explore differences
8756 between user/admin configuration and vendor defaults.
8757
8758 * PrivateTmp= now affects both /tmp and /var/tmp.
8759
8760 * Boot time status messages are now much prettier and feature
8761 proper english language. Booting up systemd has never been
8762 so sexy.
8763
8764 * Read-ahead pack files now include the inode number of all
8765 files to pre-cache. When the inode changes the pre-caching
8766 is not attempted. This should be nicer to deal with updated
8767 packages which might result in changes of read-ahead
8768 patterns.
8769
8770 * We now temporaritly lower the kernel's read_ahead_kb variable
8771 when collecting read-ahead data to ensure the kernel's
8772 built-in read-ahead does not add noise to our measurements
8773 of necessary blocks to pre-cache.
8774
8775 * There's now RequiresMountsFor= to add automatic dependencies
8776 for all mounts necessary for a specific file system path.
8777
8778 * MountAuto= and SwapAuto= have been removed from
8779 system.conf. Mounting file systems at boot has to take place
8780 in systemd now.
8781
8782 * nspawn now learned a new switch --uuid= to set the machine
8783 ID on the command line.
8784
f8c0a2cb 8785 * nspawn now learned the -b switch to automatically search
ea5943d3
LP
8786 for an init system.
8787
8788 * vt102 is now the default TERM for serial TTYs, upgraded from
8789 vt100.
8790
8791 * systemd-logind now works on VT-less systems.
8792
8793 * The build tree has been reorganized. The individual
3943231c 8794 components now have directories of their own.
ea5943d3
LP
8795
8796 * A new condition type ConditionPathIsReadWrite= is now available.
8797
8798 * nspawn learned the new -C switch to create cgroups for the
8799 container in other hierarchies.
8800
8801 * We now have support for hardware watchdogs, configurable in
8802 system.conf.
8803
8804 * The scheduled shutdown logic now has a public API.
8805
8806 * We now mount /tmp as tmpfs by default, but this can be
8807 masked and /etc/fstab can override it.
8808
d28315e4 8809 * Since udisks does not make use of /media anymore we are not
ea5943d3
LP
8810 mounting a tmpfs on it anymore.
8811
8812 * journalctl gained a new --local switch to only interleave
8813 locally generated journal files.
8814
8815 * We can now load the IMA policy at boot automatically.
8816
8817 * The GTK tools have been split off into a systemd-ui.
8818
79849bf9
LP
8819 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
8820 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
8821 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
8822 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
8823 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
8824 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
8825 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
8826 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
8827 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
8828 Gundersen
8829
16f1239e 8830CHANGES WITH 44:
b6a86739 8831
16f1239e
LP
8832 * This is mostly a bugfix release
8833
8834 * Support optional initialization of the machine ID from the
8835 KVM or container configured UUID.
8836
8837 * Support immediate reboots with "systemctl reboot -ff"
8838
8839 * Show /etc/os-release data in systemd-analyze output
8840
ab06eef8 8841 * Many bugfixes for the journal, including endianness fixes and
16f1239e
LP
8842 ensuring that disk space enforcement works
8843
ce830873 8844 * sd-login.h is C++ compatible again
16f1239e
LP
8845
8846 * Extend the /etc/os-release format on request of the Debian
8847 folks
8848
8849 * We now refuse non-UTF8 strings used in various configuration
d28315e4 8850 and unit files. This is done to ensure we do not pass invalid
16f1239e
LP
8851 data over D-Bus or expose it elsewhere.
8852
8853 * Register Mimo USB Screens as suitable for automatic seat
8854 configuration
8855
8856 * Read SELinux client context from journal clients in a race
8857 free fashion
8858
8859 * Reorder configuration file lookup order. /etc now always
8860 overrides /run in order to allow the administrator to always
b938cb90 8861 and unconditionally override vendor-supplied or
16f1239e
LP
8862 automatically generated data.
8863
8864 * The various user visible bits of the journal now have man
8865 pages. We still lack man pages for the journal API calls
8866 however.
8867
8868 * We now ship all man pages in HTML format again in the
8869 tarball.
8870
8871 Contributions from: Dave Reisner, Dirk Eibach, Frederic
8872 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
8873 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
8874 Reding
8875
437b7dee 8876CHANGES WITH 43:
b6a86739 8877
437b7dee
LP
8878 * This is mostly a bugfix release
8879
8880 * systems lacking /etc/os-release are no longer supported.
8881
8882 * Various functionality updates to libsystemd-login.so
8883
45afd519 8884 * Track class of PAM logins to distinguish greeters from
437b7dee
LP
8885 normal user logins.
8886
8887 Contributions from: Kay Sievers, Lennart Poettering, Michael
8888 Biebl
8889
204fa33c 8890CHANGES WITH 42:
b6a86739 8891
204fa33c
LP
8892 * This is an important bugfix release for v41.
8893
8894 * Building man pages is now optional which should be useful
8895 for those building systemd from git but unwilling to install
8896 xsltproc.
8897
8898 * Watchdog support for supervising services is now usable. In
8899 a future release support for hardware watchdogs
8900 (i.e. /dev/watchdog) will be added building on this.
8901
8902 * Service start rate limiting is now configurable and can be
8903 turned off per service. When a start rate limit is hit a
8904 reboot can automatically be triggered.
8905
8906 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
8907
8908 Contributions from: Benjamin Franzke, Bill Nottingham,
8909 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
8910 Schmidt, Michał Górny, Piotr Drąg
8911
e0d25329 8912CHANGES WITH 41:
b6a86739 8913
e0d25329
KS
8914 * The systemd binary is installed /usr/lib/systemd/systemd now;
8915 An existing /sbin/init symlink needs to be adapted with the
8916 package update.
8917
b13df964
LP
8918 * The code that loads kernel modules has been ported to invoke
8919 libkmod directly, instead of modprobe. This means we do not
8920 support systems with module-init-tools anymore.
8921
8922 * Watchdog support is now already useful, but still not
8923 complete.
8924
8925 * A new kernel command line option systemd.setenv= is
8926 understood to set system wide environment variables
8927 dynamically at boot.
8928
e9c1ea9d 8929 * We now limit the set of capabilities of systemd-journald.
ccd07a08 8930
353e12c2
LP
8931 * We now set SIGPIPE to ignore by default, since it only is
8932 useful in shell pipelines, and has little use in general
8933 code. This can be disabled with IgnoreSIPIPE=no in unit
8934 files.
8935
b13df964
LP
8936 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
8937 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
8938 William Douglas
8939
d26e4270 8940CHANGES WITH 40:
b6a86739 8941
d26e4270
LP
8942 * This is mostly a bugfix release
8943
8944 * We now expose the reason why a service failed in the
8945 "Result" D-Bus property.
8946
8947 * Rudimentary service watchdog support (will be completed over
8948 the next few releases.)
8949
8950 * When systemd forks off in order execute some service we will
8951 now immediately changes its argv[0] to reflect which process
8952 it will execute. This is useful to minimize the time window
8953 with a generic argv[0], which makes bootcharts more useful
8954
b13df964
LP
8955 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
8956 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
8957 Mike Kazantsev, Ray Strode
8958
220a21d3 8959CHANGES WITH 39:
b6a86739 8960
220a21d3
LP
8961 * This is mostly a test release, but incorporates many
8962 bugfixes.
8963
8964 * New systemd-cgtop tool to show control groups by their
8965 resource usage.
8966
8967 * Linking against libacl for ACLs is optional again. If
8968 disabled, support tracking device access for active logins
8969 goes becomes unavailable, and so does access to the user
8970 journals by the respective users.
8971
8972 * If a group "adm" exists, journal files are automatically
8973 owned by them, thus allow members of this group full access
8974 to the system journal as well as all user journals.
8975
8976 * The journal now stores the SELinux context of the logging
8977 client for all entries.
8978
8979 * Add C++ inclusion guards to all public headers
8980
8981 * New output mode "cat" in the journal to print only text
8982 messages, without any meta data like date or time.
8983
8984 * Include tiny X server wrapper as a temporary stop-gap to
8985 teach XOrg udev display enumeration. This is used by display
8986 managers such as gdm, and will go away as soon as XOrg
8987 learned native udev hotplugging for display devices.
8988
8989 * Add new systemd-cat tool for executing arbitrary programs
8990 with STDERR/STDOUT connected to the journal. Can also act as
8991 BSD logger replacement, and does so by default.
8992
8993 * Optionally store all locally generated coredumps in the
8994 journal along with meta data.
8995
8996 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
8997 writing short strings to files (for usage for /sys), and for
8998 creating symlinks, character and block device nodes.
8999
9000 * New unit file option ControlGroupPersistent= to make cgroups
9001 persistent, following the mechanisms outlined in
56cadcb6 9002 https://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
220a21d3
LP
9003
9004 * Support multiple local RTCs in a sane way
9005
9006 * No longer monopolize IO when replaying readahead data on
9007 rotating disks, since we might starve non-file-system IO to
9008 death, since fanotify() will not see accesses done by blkid,
9009 or fsck.
9010
d28315e4 9011 * Do not show kernel threads in systemd-cgls anymore, unless
220a21d3
LP
9012 requested with new -k switch.
9013
9014 Contributions from: Dan Horák, Kay Sievers, Lennart
9015 Poettering, Michal Schmidt
9016
9017CHANGES WITH 38:
b6a86739 9018
220a21d3
LP
9019 * This is mostly a test release, but incorporates many
9020 bugfixes.
9021
9022 * The git repository moved to:
9023 git://anongit.freedesktop.org/systemd/systemd
9024 ssh://git.freedesktop.org/git/systemd/systemd
9025
9026 * First release with the journal
9027 http://0pointer.de/blog/projects/the-journal.html
9028
9029 * The journal replaces both systemd-kmsg-syslogd and
9030 systemd-stdout-bridge.
9031
9032 * New sd_pid_get_unit() API call in libsystemd-logind
9033
9034 * Many systemadm clean-ups
9035
9036 * Introduce remote-fs-pre.target which is ordered before all
9037 remote mounts and may be used to start services before all
9038 remote mounts.
9039
9040 * Added Mageia support
9041
9042 * Add bash completion for systemd-loginctl
9043
9044 * Actively monitor PID file creation for daemons which exit in
9045 the parent process before having finished writing the PID
9046 file in the daemon process. Daemons which do this need to be
9047 fixed (i.e. PID file creation must have finished before the
9048 parent exits), but we now react a bit more gracefully to them.
9049
9050 * Add colourful boot output, mimicking the well-known output
9051 of existing distributions.
9052
9053 * New option PassCredentials= for socket units, for
9054 compatibility with a recent kernel ABI breakage.
9055
9056 * /etc/rc.local is now hooked in via a generator binary, and
9057 thus will no longer act as synchronization point during
9058 boot.
9059
9060 * systemctl list-unit-files now supports --root=.
9061
9062 * systemd-tmpfiles now understands two new commands: z, Z for
9063 relabelling files according to the SELinux database. This is
9064 useful to apply SELinux labels to specific files in /sys,
9065 among other things.
9066
9067 * Output of SysV services is now forwarded to both the console
9068 and the journal by default, not only just the console.
9069
9070 * New man pages for all APIs from libsystemd-login.
9071
ce830873 9072 * The build tree got reorganized and the build system is a
220a21d3
LP
9073 lot more modular allowing embedded setups to specifically
9074 select the components of systemd they are interested in.
9075
9076 * Support for Linux systems lacking the kernel VT subsystem is
9077 restored.
9078
9079 * configure's --with-rootdir= got renamed to
9080 --with-rootprefix= to follow the naming used by udev and
9081 kmod
9082
d28315e4 9083 * Unless specified otherwise we will now install to /usr instead
220a21d3
LP
9084 of /usr/local by default.
9085
9086 * Processes with '@' in argv[0][0] are now excluded from the
9087 final shut-down killing spree, following the logic explained
9088 in:
56cadcb6 9089 https://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
220a21d3
LP
9090
9091 * All processes remaining in a service cgroup when we enter
9092 the START or START_PRE states are now killed with
9093 SIGKILL. That means it is no longer possible to spawn
9094 background processes from ExecStart= lines (which was never
9095 supported anyway, and bad style).
9096
9097 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
9098 reloading of units together.
9099
4c8cd173 9100 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
220a21d3
LP
9101 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
9102 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
9103 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
9104 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek