]>
Commit | Line | Data |
---|---|---|
d657c51f | 1 | systemd System and Service Manager |
220a21d3 | 2 | |
28fd1ba3 | 3 | CHANGES WITH 256 in spe: |
880eaa3a LB |
4 | |
5 | Announcements of Future Feature Removals and Incompatible Changes: | |
6 | ||
03a46b5e YW |
7 | * Previously, systemd-networkd did not explicitly remove any bridge VLAN |
8 | IDs assigned on bridge master and ports. Since v256, if a .network | |
9 | file for an interface has at least one valid settings in [BridgeVLAN] | |
10 | section, then all assigned VLAN IDs on the interface that are not | |
11 | configured in the .network file are removed. | |
12 | ||
b4959550 RB |
13 | Network Management: |
14 | ||
15 | * systemd-networkd's proxy support gained a new option to configure | |
16 | a private VLAN variant of the proxy ARP supported by the kernel | |
17 | under the name IPv4ProxyARPPrivateVLAN=. | |
18 | ||
db11bab3 | 19 | CHANGES WITH 255: |
c405b91f LB |
20 | |
21 | Announcements of Future Feature Removals and Incompatible Changes: | |
22 | ||
23 | * Support for split-usr (/usr/ mounted separately during late boot, | |
24 | instead of being mounted by the initrd before switching to the rootfs) | |
25 | and unmerged-usr (parallel directories /bin/ and /usr/bin/, /lib/ and | |
26 | /usr/lib/, …) has been removed. For more details, see: | |
27 | https://lists.freedesktop.org/archives/systemd-devel/2022-September/048352.html | |
28 | ||
29 | * We intend to remove cgroup v1 support from a systemd release after | |
30 | the end of 2023. If you run services that make explicit use of | |
31 | cgroup v1 features (i.e. the "legacy hierarchy" with separate | |
32 | hierarchies for each controller), please implement compatibility with | |
33 | cgroup v2 (i.e. the "unified hierarchy") sooner rather than later. | |
34 | Most of Linux userspace has been ported over already. | |
35 | ||
36 | * Support for System V service scripts is now deprecated and will be | |
37 | removed in a future release. Please make sure to update your software | |
38 | *now* to include a native systemd unit file instead of a legacy | |
39 | System V script to retain compatibility with future systemd releases. | |
40 | ||
41 | * Support for the SystemdOptions EFI variable is deprecated. | |
42 | 'bootctl systemd-efi-options' will emit a warning when used. It seems | |
43 | that this feature is little-used and it is better to use alternative | |
44 | approaches like credentials and confexts. The plan is to drop support | |
45 | altogether at a later point, but this might be revisited based on | |
46 | user feedback. | |
47 | ||
48 | * systemd-run's switch --expand-environment= which currently is disabled | |
49 | by default when combined with --scope, will be changed in a future | |
50 | release to be enabled by default. | |
51 | ||
fcdd21ec | 52 | * "systemctl switch-root" is now restricted to initrd transitions only. |
4da159bc | 53 | |
427ddaf6 ZJS |
54 | Transitions between real systems should be done with |
55 | "systemctl soft-reboot" instead. | |
56 | ||
57 | * The "ip=off" and "ip=none" kernel command line options interpreted by | |
c24a8c6b | 58 | systemd-network-generator will now result in IPv6RA + link-local |
427ddaf6 | 59 | addressing being disabled, too. Previously DHCP was turned off, but |
c24a8c6b LP |
60 | IPv6RA and IPv6 link-local addressing was left enabled. |
61 | ||
f456764c LB |
62 | * The NAMING_BRIDGE_MULTIFUNCTION_SLOT naming scheme has been deprecated |
63 | and is now disabled. | |
64 | ||
ce3eccd8 LB |
65 | * SuspendMode=, HibernateState= and HybridSleepState= in the [Sleep] |
66 | section of systemd-sleep.conf are now deprecated and have no effect. | |
67 | They did not (and could not) take any value other than the respective | |
68 | default. HybridSleepMode= is also deprecated, and will now always use | |
69 | the 'suspend' disk mode. | |
70 | ||
f456764c LB |
71 | Service Manager: |
72 | ||
c2322b48 LP |
73 | * The way services are spawned has been overhauled. Previously, a |
74 | process was forked that shared all of the manager's memory (via | |
427ddaf6 | 75 | copy-on-write) while doing all the required setup (e.g.: mount |
c2322b48 LP |
76 | namespaces, CGroup configuration, etc.) before exec'ing the target |
77 | executable. This was problematic for various reasons: several glibc | |
78 | APIs were called that are not supposed to be used after a fork but | |
79 | before an exec, copy-on-write meant that if either process (the | |
80 | manager or the child) touched a memory page a copy was triggered, and | |
81 | also the memory footprint of the child process was that of the | |
427ddaf6 | 82 | manager, but with the memory limits of the service. From this version |
c2322b48 | 83 | onward, the new process is spawned using CLONE_VM and CLONE_VFORK |
427ddaf6 | 84 | semantics via posix_spawn(3), and it immediately execs a new internal |
c2322b48 LP |
85 | binary, systemd-executor, that receives the configuration to apply |
86 | via memfd, and sets up the process before exec'ing the target | |
054042f8 LB |
87 | executable. The systemd-executor binary is pinned by file descriptor |
88 | by each manager instance (system and users), and the reference is | |
89 | updated on daemon-reexec - it is thus important to reexec all running | |
90 | manager instances when the systemd-executor and/or libsystemd* | |
91 | libraries are updated on the filesystem. | |
f456764c | 92 | |
c2322b48 LP |
93 | * Most of the internal process tracking is being changed to use PIDFDs |
94 | instead of PIDs when the kernel supports it, to improve robustness | |
95 | and reliability. | |
f456764c | 96 | |
427ddaf6 ZJS |
97 | * A new option SurviveFinalKillSignal= can be used to configure the |
98 | unit to be skipped in the final SIGTERM/SIGKILL spree on shutdown. | |
99 | This is part of the required configuration to let a unit's processes | |
100 | survive a soft-reboot operation. | |
f456764c | 101 | |
c2322b48 LP |
102 | * System extension images (sysext) can now set |
103 | EXTENSION_RELOAD_MANAGER=1 in their extension-release files to | |
104 | automatically reload the service manager (PID 1) when | |
105 | merging/refreshing/unmerging on boot. Generally, while this can be | |
106 | used to ship services in system extension images it's recommended to | |
107 | do that via portable services instead. | |
f456764c LB |
108 | |
109 | * The ExtensionImages= and ExtensionDirectories= options now support | |
110 | confexts images/directories. | |
111 | ||
427ddaf6 ZJS |
112 | * A new option NFTSet= provides a method for integrating dynamic cgroup |
113 | IDs into firewall rules with NFT sets. The benefit of using this | |
114 | setting is to be able to use control group as a selector in firewall | |
115 | rules easily and this in turn allows more fine grained filtering. | |
116 | Also, NFT rules for cgroup matching use numeric cgroup IDs, which | |
117 | change every time a service is restarted, making them hard to use in | |
118 | systemd environment. | |
f456764c | 119 | |
c2322b48 LP |
120 | * A new option CoredumpReceive= can be set for service and scope units, |
121 | together with Delegate=yes, to make systemd-coredump on the host | |
427ddaf6 ZJS |
122 | forward core files from processes crashing inside the delegated |
123 | CGroup subtree to systemd-coredump running in the container. This new | |
124 | option is by default used by systemd-nspawn containers that use the | |
125 | "--boot" switch. | |
f456764c LB |
126 | |
127 | * A new ConditionSecurity=measured-uki option is now available, to ensure | |
128 | a unit can only run when the system has been booted from a measured UKI. | |
129 | ||
130 | * MemoryAvailable= now considers physical memory if there are no CGroup | |
131 | memory limits set anywhere in the tree. | |
132 | ||
133 | * The $USER environment variable is now always set for services, while | |
134 | previously it was only set if User= was specified. A new option | |
135 | SetLoginEnvironment= is now supported to determine whether to also set | |
427ddaf6 | 136 | $HOME, $LOGNAME, and $SHELL. |
f456764c | 137 | |
c2322b48 LP |
138 | * Socket units now support a new pair of |
139 | PollLimitBurst=/PollLimitInterval= options to configure a limit on | |
140 | how often polling events on the file descriptors backing this unit | |
141 | will be considered within a time window. | |
f456764c | 142 | |
427ddaf6 | 143 | * Scope units can now be created using PIDFDs instead of PIDs to select |
f456764c LB |
144 | the processes they should include. |
145 | ||
c2322b48 LP |
146 | * Sending SIGRTMIN+18 with 0x500 as sigqueue() value will now cause the |
147 | manager to dump the list of currently pending jobs. | |
f456764c | 148 | |
427ddaf6 ZJS |
149 | * If the kernel supports MOVE_MOUNT_BENEATH, the systemctl and |
150 | machinectl bind and mount-image verbs will now cause the new mount to | |
151 | replace the old mount (if any), instead of overmounting it. | |
f456764c | 152 | |
feed2910 | 153 | * Units now have MemoryPeak, MemorySwapPeak, MemorySwapCurrent and |
b0f96596 LP |
154 | MemoryZSwapCurrent properties, which respectively contain the values |
155 | of the cgroup v2's memory.peak, memory.swap.peak, memory.swap.current | |
7ba8260c | 156 | and memory.zswap.current properties. This information is also shown in |
b0f96596 | 157 | "systemctl status" output, if available. |
6c71db76 | 158 | |
f456764c LB |
159 | TPM2 Support + Disk Encryption & Authentication: |
160 | ||
c2322b48 LP |
161 | * systemd-cryptenroll now allows specifying a PCR bank and explicit hash |
162 | value in the --tpm2-pcrs= option. | |
f456764c | 163 | |
b0f96596 LP |
164 | * systemd-cryptenroll now allows specifying a TPM2 key handle (nv |
165 | index) to be used instead of the default SRK via the new | |
166 | --tpm2-seal-key-handle= option. | |
f456764c | 167 | |
b0f96596 LP |
168 | * systemd-cryptenroll now allows TPM2 enrollment using only a TPM2 |
169 | public key (in TPM2B_PUBLIC format) – without access to the TPM2 | |
170 | device itself – which enables offline sealing of LUKS images for a | |
171 | specific TPM2 chip, as long as the SRK public key is known. Pass the | |
172 | public to the tool via the new --tpm2-device-key= switch. | |
feed2910 | 173 | |
f456764c LB |
174 | * systemd-cryptsetup is now installed in /usr/bin/ and is no longer an |
175 | internal-only executable. | |
176 | ||
177 | * The TPM2 Storage Root Key will now be set up, if not already present, | |
b0f96596 LP |
178 | by a new systemd-tpm2-setup.service early boot service. The SRK will |
179 | be stored in PEM format and TPM2_PUBLIC format (the latter is useful | |
180 | for systemd-cryptenroll --tpm2-device-key=, as mentioned above) for | |
181 | easier access. A new "srk" verb has been added to systemd-analyze to | |
182 | allow extracting it on demand if it is already set up. | |
f456764c | 183 | |
c2322b48 LP |
184 | * The internal systemd-pcrphase executable has been renamed to |
185 | systemd-pcrextend. | |
f456764c | 186 | |
c2322b48 LP |
187 | * The systemd-pcrextend tool gained a new --pcr= switch to override |
188 | which PCR to measure into. | |
189 | ||
190 | * systemd-pcrextend now exposes a Varlink interface at | |
191 | io.systemd.PCRExtend that can be used to do measurements and event | |
192 | logging on demand. | |
f456764c LB |
193 | |
194 | * TPM measurements are now also written to an event log at | |
c2322b48 LP |
195 | /run/log/systemd/tpm2-measure.log, using a derivative of the TCG |
196 | Canonical Event Log format. Previously we'd only log them to the | |
197 | journal, where they however were subject to rotation and similar. | |
f456764c | 198 | |
c5c5f0fe | 199 | * A new component "systemd-pcrlock" has been added that allows managing |
7eff3e2c | 200 | local TPM2 PCR policies for PCRs 0-7 and similar, which are hard to |
c5c5f0fe LP |
201 | predict by the OS vendor because of the inherently local nature of |
202 | what measurements they contain, such as firmware versions of the | |
203 | system and extension cards and suchlike. pcrlock can predict PCR | |
204 | measurements ahead of time based on various inputs, such as the local | |
205 | TPM2 event log, GPT partition tables, PE binaries, UKI kernels, and | |
206 | various other things. It can then pre-calculate a TPM2 policy from | |
207 | this, which it stores in an TPM2 NV index. TPM2 objects (such as disk | |
208 | encryption keys) can be locked against this NV index, so that they | |
209 | are locked against a specific combination of system firmware and | |
210 | state. Alternatives for each component are supported to allowlist | |
211 | multiple kernel versions or boot loader version simultaneously | |
212 | without losing access to the disk encryption keys. The tool can also | |
837667a1 ZJS |
213 | be used to analyze and validate the local TPM2 event log. |
214 | systemd-cryptsetup, systemd-cryptenroll, systemd-repart have all been | |
215 | updated to support such policies. There's currently no support for | |
216 | locking the system's root disk against a pcrlock policy, this will be | |
217 | added soon. Moreover, it is currently not possible to combine a | |
218 | pcrlock policy with a signed PCR policy. This component is | |
0e5f89b5 | 219 | experimental and its public interface is subject to change. |
c5c5f0fe | 220 | |
f456764c LB |
221 | systemd-boot, systemd-stub, ukify, bootctl, kernel-install: |
222 | ||
f456764c LB |
223 | * bootctl will now show whether the system was booted from a UKI in its |
224 | status output. | |
225 | ||
226 | * systemd-boot and systemd-stub now use different project keys in their | |
227 | respective SBAT sections, so that they can be revoked individually if | |
228 | needed. | |
229 | ||
230 | * systemd-boot will no longer load unverified Devicetree blobs when UEFI | |
231 | SecureBoot is enabled. For more details see: | |
232 | https://github.com/systemd/systemd/security/advisories/GHSA-6m6p-rjcq-334c | |
233 | ||
c2322b48 LP |
234 | * systemd-boot gained new hotkeys to reboot and power off the system |
235 | from the boot menu ("B" and "O"). If the "auto-poweroff" and | |
236 | "auto-reboot" options in loader.conf are set these entries are also | |
237 | shown as menu items (which is useful on devices lacking a regular | |
238 | keyboard). | |
f456764c | 239 | |
c2322b48 | 240 | * systemd-boot gained a new configuration value "menu-disabled" for the |
f456764c LB |
241 | set-timeout option, to allow completely disabling the boot menu, |
242 | including the hotkey. | |
243 | ||
7eff3e2c ZJS |
244 | * systemd-boot will now measure the content of loader.conf in TPM2 |
245 | PCR 5. | |
f456764c | 246 | |
c2322b48 | 247 | * systemd-stub will now concatenate the content of all kernel |
7eff3e2c | 248 | command-line addons before measuring them in TPM2 PCR 12, in a single |
c2322b48 | 249 | measurement, instead of measuring them individually. |
f456764c | 250 | |
c2322b48 LP |
251 | * systemd-stub will now measure and load Devicetree Blob addons, which |
252 | are searched and loaded following the same model as the existing | |
253 | kernel command-line addons. | |
f456764c LB |
254 | |
255 | * systemd-stub will now ignore unauthenticated kernel command line options | |
256 | passed from systemd-boot when running inside Confidential VMs with UEFI | |
257 | SecureBoot enabled. | |
258 | ||
feed2910 LB |
259 | * systemd-stub will now load a Devicetree blob even if the firmware did |
260 | not load any beforehand (e.g.: for ACPI systems). | |
261 | ||
427ddaf6 ZJS |
262 | * ukify is no longer considered experimental, and now ships in /usr/bin/. |
263 | ||
264 | * ukify gained a new verb inspect to describe the sections of a UKI and | |
265 | print the contents of the well-known sections. | |
266 | ||
111df871 | 267 | * ukify gained a new verb genkey to generate a set of key pairs for |
427ddaf6 ZJS |
268 | signing UKIs and their PCR data. |
269 | ||
270 | * The 90-loaderentry kernel-install hook now supports installing device | |
271 | trees. | |
272 | ||
7eff3e2c ZJS |
273 | * kernel-install now supports the --json=, --root=, --image=, and |
274 | --image-policy= options for the inspect verb. | |
feed2910 | 275 | |
b0f96596 LP |
276 | * kernel-install now supports new list and add-all verbs. The former |
277 | lists all installed kernel images (if those are available in | |
278 | /usr/lib/modules/). The latter will install all the kernels it can | |
279 | find to the ESP. | |
feed2910 | 280 | |
f456764c LB |
281 | systemd-repart: |
282 | ||
427ddaf6 ZJS |
283 | * A new option --copy-from= has been added that synthesizes partition |
284 | definitions from the given image, which are then applied by the | |
285 | systemd-repart algorithm. | |
f456764c | 286 | |
68a5300f | 287 | * A new option --copy-source= has been added, which can be used to specify |
f456764c LB |
288 | a directory to which CopyFiles= is considered relative to. |
289 | ||
427ddaf6 ZJS |
290 | * New --make-ddi=confext, --make-ddi=sysext, and --make-ddi=portable |
291 | options have been added to make it easier to generate these types of | |
292 | DDIs, without having to provide repart.d definitions for them. | |
f456764c | 293 | |
c2322b48 LP |
294 | * The dm-verity salt and UUID will now be derived from the specified |
295 | seed value. | |
f456764c LB |
296 | |
297 | * New VerityDataBlockSizeBytes= and VerityHashBlockSizeBytes= can now be | |
c2322b48 | 298 | configured in repart.d/ configuration files. |
f456764c | 299 | |
c2322b48 | 300 | * A new Subvolumes= setting is now supported in repart.d/ configuration |
f456764c LB |
301 | files, to indicate which directories in the target partition should be |
302 | btrfs subvolumes. | |
303 | ||
b0f96596 LP |
304 | * A new --tpm2-device-key= option can be used to lock a disk against a |
305 | specific TPM2 public key. This matches the same switch the | |
306 | systemd-cryptenroll tool now supports (see above). | |
feed2910 | 307 | |
f456764c LB |
308 | Journal: |
309 | ||
68a5300f | 310 | * The journalctl --lines= parameter now accepts +N to show the oldest N |
f456764c LB |
311 | entries instead of the newest. |
312 | ||
feed2910 LB |
313 | * journald now ensures that sealing happens once per epoch, and sets a |
314 | new compatibility flag to distinguish old journal files that were | |
315 | created before this change, for backward compatibility. | |
316 | ||
7a05926f YW |
317 | Device Management: |
318 | ||
319 | * udev will now create symlinks to loopback block devices in the | |
320 | /dev/disk/by-loop-ref/ directory that are based on the .lo_file_name | |
321 | string field selected during allocation. The systemd-dissect tool and | |
427ddaf6 ZJS |
322 | the util-linux losetup command now supports a complementing new switch |
323 | --loop-ref= for selecting the string. This means a loopback block | |
324 | device may now be allocated under a caller-chosen reference and can | |
325 | subsequently be referenced without first having to look up the block | |
326 | device name the caller ended up with. | |
7a05926f YW |
327 | |
328 | * udev also creates symlinks to loopback block devices in the | |
329 | /dev/disk/by-loop-inode/ directory based on the .st_dev/st_ino fields | |
330 | of the inode attached to the loopback block device. This means that | |
331 | attaching a file to a loopback device will implicitly make a handle | |
332 | available to be found via that file's inode information. | |
333 | ||
c2322b48 | 334 | * udevadm info gained support for JSON output via a new --json= flag, and |
f456764c LB |
335 | for filtering output using the same mechanism that udevadm trigger |
336 | already implements. | |
337 | ||
64f2cf77 ZJS |
338 | * The predictable network interface naming logic is extended to include |
339 | the SR-IOV-R "representor" information in network interface names. | |
340 | This feature was intended for v254, but even though the code was | |
341 | merged, the part that actually enabled the feature was forgotten. | |
342 | It is now enabled by default and is part of the new "v255" naming | |
343 | scheme. | |
344 | ||
c2322b48 LP |
345 | * A new hwdb/rules file has been added that sets the |
346 | ID_NET_AUTO_LINK_LOCAL_ONLY=1 udev property on all network interfaces | |
347 | that should usually only be configured with link-local addressing | |
348 | (IPv4LL + IPv6LL), i.e. for PC-to-PC cables ("laplink") or | |
349 | Thunderbolt networking. systemd-networkd and NetworkManager (soon) | |
350 | will make use of this information to apply an appropriate network | |
351 | configuration by default. | |
352 | ||
353 | * The ID_NET_DRIVER property on network interfaces is now set | |
354 | relatively early in the udev rule set so that other rules may rely on | |
355 | its use. This is implemented in a new "net-driver" udev built-in. | |
356 | ||
c57ff623 YW |
357 | Network Management: |
358 | ||
359 | * The "duid-only" option for DHCPv4 client's ClientIdentifier= setting | |
360 | is now dropped, as it never worked, hence it should not be used by | |
361 | anyone. | |
362 | ||
427ddaf6 ZJS |
363 | * The 'prefixstable' ipv6 address generation mode now considers the SSID |
364 | when generating stable addresses, so that a different stable address | |
365 | is used when roaming between wireless networks. If you already use | |
366 | 'prefixstable' addresses with wireless networks, the stable address | |
367 | will be changed by the update. | |
06960d17 | 368 | |
427ddaf6 | 369 | * The DHCPv4 client gained a RapidCommit option, true by default, which |
808b65a0 RP |
370 | enables RFC4039 Rapid Commit behavior to obtain a lease in a |
371 | simplified 2-message exchange instead of the typical 4-message | |
427ddaf6 | 372 | exchange, if also supported by the DHCP server. |
808b65a0 | 373 | |
f456764c | 374 | * The DHCPv4 client gained new InitialCongestionWindow= and |
c57ff623 | 375 | InitialAdvertisedReceiveWindow= options for route configurations. |
f456764c LB |
376 | |
377 | * The DHCPv4 client gained a new RequestAddress= option that allows | |
378 | to send a preferred IP address in the initial DHCPDISCOVER message. | |
379 | ||
c57ff623 YW |
380 | * The DHCPv4 server and client gained support for IPv6-only mode |
381 | (RFC8925). | |
382 | ||
68a5300f | 383 | * The SendHostname= and Hostname= options are now available for the |
427ddaf6 | 384 | DHCPv6 client, independently of the DHCPv4= option, so that these |
b90480c8 RP |
385 | configuration values can be set independently for each client. |
386 | ||
f456764c LB |
387 | * The DHCPv4 and DHCPv6 client state can now be queried via D-Bus, |
388 | including lease information. | |
389 | ||
c57ff623 | 390 | * The DHCPv6 client can now be configured to use a custom DUID type. |
f456764c | 391 | |
c57ff623 YW |
392 | * .network files gained a new IPv4ReversePathFilter= setting in the |
393 | [Network] section, to control sysctl's rp_filter setting. | |
f456764c | 394 | |
c57ff623 YW |
395 | * .network files gaiend a new HopLimit= setting in the [Route] section, |
396 | to configure a per-route hop limit. | |
f456764c | 397 | |
c57ff623 YW |
398 | * .network files gained a new TCPRetransmissionTimeoutSec= setting in |
399 | the [Route] section, to configure a per-route TCP retransmission | |
400 | timeout. | |
f456764c | 401 | |
68a5300f | 402 | * A new directive NFTSet= provides a method for integrating network |
f456764c LB |
403 | configuration into firewall rules with NFT sets. The benefit of using |
404 | this setting is that static network configuration or dynamically | |
405 | obtained network addresses can be used in firewall rules with the | |
406 | indirection of NFT set types. | |
407 | ||
f456764c | 408 | * The [IPv6AcceptRA] section supports the following new options: |
427ddaf6 | 409 | UsePREF64=, UseHopLimit=, UseICMP6RateLimit=, and NFTSet=. |
f456764c LB |
410 | |
411 | * The [IPv6SendRA] section supports the following new options: | |
427ddaf6 | 412 | RetransmitSec=, HopLimit=, HomeAgent=, HomeAgentLifetimeSec=, and |
f456764c LB |
413 | HomeAgentPreference=. |
414 | ||
c57ff623 YW |
415 | * A new [IPv6PREF64Prefix] set of options, containing Prefix= and |
416 | LifetimeSec=, has been introduced to append pref64 options in router | |
417 | advertisements (RFC8781). | |
418 | ||
f456764c | 419 | * The network generator now configures the interfaces with only |
427ddaf6 | 420 | link-local addressing if "ip=link-local" is specified on the kernel |
f456764c LB |
421 | command line. |
422 | ||
08752450 YW |
423 | * The prefix of the configuration files generated by the network |
424 | generator from the kernel command line is now prefixed with '70-', | |
425 | to make them have higher precedence over the default configuration | |
426 | files. | |
427 | ||
28a8aac7 LB |
428 | * Added a new -Ddefault-network=BOOL meson option, that causes more |
429 | .network files to be installed as enabled by default. These configuration | |
430 | files will which match generic setups, e.g. 89-ethernet.network matches | |
431 | all Ethernet interfaces and enables both DHCPv4 and DHCPv6 clients. | |
08752450 | 432 | |
c2322b48 LP |
433 | * If a ID_NET_MANAGED_BY= udev property is set on a network device and |
434 | it is any other string than "io.systemd.Network" then networkd will | |
435 | not manage this device. This may be used to allow multiple network | |
436 | management services to run in parallel and assign ownership of | |
437 | specific devices explicitly. NetworkManager will soon implement a | |
438 | similar logic. | |
f456764c | 439 | |
c2322b48 | 440 | systemctl: |
f456764c LB |
441 | |
442 | * systemctl is-failed now checks the system state if no unit is | |
443 | specified. | |
444 | ||
427ddaf6 ZJS |
445 | * systemctl will now automatically soft-reboot if a new root file system |
446 | is found under /run/nextroot/ when a reboot operation is invoked. | |
f456764c | 447 | |
c2322b48 | 448 | Login management: |
f456764c | 449 | |
427ddaf6 | 450 | * Wall messages now work even when utmp support is disabled, using |
f456764c LB |
451 | systemd-logind to query the necessary information. |
452 | ||
453 | * systemd-logind now sends a new PrepareForShutdownWithMetadata D-Bus | |
427ddaf6 ZJS |
454 | signal before shutdown/reboot/soft-reboot that includes additional |
455 | information compared to the PrepareForShutdown signal. Currently the | |
456 | additional information is the type of operation that is about to be | |
457 | executed. | |
f456764c | 458 | |
c2322b48 LP |
459 | Hibernation & Suspend: |
460 | ||
461 | * The kernel and OS versions will no longer be checked on resume from | |
462 | hibernation. | |
463 | ||
464 | * Hibernation into swap files backed by btrfs are now | |
465 | supported. (Previously this was supported only for other file | |
466 | systems.) | |
467 | ||
468 | Other: | |
469 | ||
0e5f89b5 LB |
470 | * A new systemd-vmspawn tool has been added, that aims to provide for VMs |
471 | the same interfaces and functionality that systemd-nspawn provides for | |
472 | containers. For now it supports QEMU as a backend, and exposes some of | |
473 | its options to the user. This component is experimental and its public | |
474 | interface is subject to change. | |
bf88d041 | 475 | |
c2322b48 LP |
476 | * "systemd-analyze plot" has gained tooltips on each unit name with |
477 | related-unit information in its svg output, such as Before=, | |
478 | Requires=, and similar properties. | |
479 | ||
480 | * A new varlinkctl tool has been added to allow interfacing with | |
481 | Varlink services, and introspection has been added to all such | |
48696c2d LB |
482 | services. This component is experimental and its public interface is |
483 | subject to change. | |
c2322b48 LP |
484 | |
485 | * systemd-sysext and systemd-confext now expose a Varlink service | |
486 | at io.systemd.sysext. | |
487 | ||
d8e7e1b2 LB |
488 | * portable services now accept confexts as extensions. |
489 | ||
f456764c LB |
490 | * systemd-sysupdate now accepts directories in the MatchPattern= option. |
491 | ||
f456764c | 492 | * systemd-run will now output the invocation ID of the launched |
324ec6b5 | 493 | transient unit and its peak memory usage. |
f456764c LB |
494 | |
495 | * systemd-analyze, systemd-tmpfiles, systemd-sysusers, systemd-sysctl, | |
427ddaf6 ZJS |
496 | and systemd-binfmt gained a new --tldr option that can be used instead |
497 | of --cat-config to suppress uninteresting configuration lines, such as | |
498 | comments and whitespace. | |
f456764c | 499 | |
c2322b48 LP |
500 | * resolvectl gained a new "show-server-state" command that shows |
501 | current statistics of the resolver. This is backed by a new | |
502 | DumpStatistics() Varlink method provided by systemd-resolved. | |
f456764c LB |
503 | |
504 | * systemd-timesyncd will now emit a D-Bus signal when the LinkNTPServers | |
505 | property changes. | |
506 | ||
507 | * vconsole now supports KEYMAP=@kernel for preserving the kernel keymap | |
508 | as-is. | |
509 | ||
f456764c LB |
510 | * seccomp now supports the LoongArch64 architecture. |
511 | ||
feed2910 LB |
512 | * seccomp may now be enabled for services running as a non-root User= |
513 | without NoNewPrivileges=yes. | |
514 | ||
427ddaf6 ZJS |
515 | * systemd-id128 now supports a new -P option to show only values. The |
516 | combination of -P and --app options is also supported. | |
f456764c | 517 | |
427ddaf6 ZJS |
518 | * A new pam_systemd_loadkey.so PAM module is now available, which will |
519 | automatically fetch the passphrase used by cryptsetup to unlock the | |
520 | root file system and set it as the PAM authtok. This enables, among | |
521 | other things, configuring auto-unlock of the GNOME Keyring / KDE | |
522 | Wallet when autologin is configured. | |
f456764c LB |
523 | |
524 | * Many meson options now use the 'feature' type, which means they | |
525 | take enabled/disabled/auto as values. | |
526 | ||
427ddaf6 | 527 | * A new meson option -Dconfigfiledir= can be used to change where |
f456764c LB |
528 | configuration files with default values are installed to. |
529 | ||
43fe529e | 530 | * Options and verbs in man pages are now tagged with the version they |
f456764c LB |
531 | were first introduced in. |
532 | ||
08b939a6 LP |
533 | * A new component "systemd-storagetm" has been added, which exposes all |
534 | local block devices as NVMe-TCP devices, fully automatically. It's | |
535 | hooked into a new target unit storage-target-mode.target that is | |
536 | suppsoed to be booted into via | |
537 | rd.systemd.unit=storage-target-mode.target on the kernel command | |
538 | line. This is intended to be used for installers and debugging to | |
539 | quickly get access to the local disk. It's inspired by MacOS "target | |
48696c2d LB |
540 | disk mode". This component is experimental and its public interface is |
541 | subject to change. | |
08b939a6 | 542 | |
c2322b48 LP |
543 | * A new component "systemd-bsod" has been added, which can show logged |
544 | error messages full screen, if they have a log level of LOG_EMERG log | |
48696c2d LB |
545 | level. This component is experimental and its public interface is |
546 | subject to change. | |
c2322b48 LP |
547 | |
548 | * The systemd-dissect tool's --with command will now set the | |
549 | $SYSTEMD_DISSECT_DEVICE environment variable to the block device it | |
550 | operates on for the invoked process. | |
551 | ||
552 | * The systemd-mount tool gained a new --tmpfs switch for mounting a new | |
553 | 'tmpfs' instance. This is useful since it does so via .mount units | |
554 | and thus can be executed remotely or in containers. | |
555 | ||
556 | * The various tools in systemd that take "verbs" (such as systemctl, | |
557 | loginctl, machinectl, …) now will suggest a close verb name in case | |
558 | the user specified an unrecognized one. | |
559 | ||
560 | * libsystemd now exports a new function sd_id128_get_app_specific() | |
561 | that generates "app-specific" 128bit IDs from any ID. It's similar to | |
562 | sd_id128_get_machine_app_specific() and | |
563 | sd_id128_get_boot_app_specific() but takes the ID to base calculation | |
564 | on as input. This new functionality is also exposed in the | |
565 | "systemd-id128" tool where you can now combine --app= with `show`. | |
566 | ||
567 | * All tools that parse timestamps now can also parse RFC3339 style | |
568 | timestamps that include the "T" and Z" characters. | |
569 | ||
a9d942ae | 570 | * New documentation has been added: |
c2322b48 LP |
571 | |
572 | https://systemd.io/FILE_DESCRIPTOR_STORE | |
573 | https://systemd.io/TPM2_PCR_MEASUREMENTS | |
31a4796e | 574 | https://systemd.io/MOUNT_REQUIREMENTS |
c2322b48 LP |
575 | |
576 | * The codebase now recognizes the suffix .confext.raw and .sysext.raw | |
577 | as alternative to the .raw suffix generally accepted for DDIs. It is | |
578 | recommended to name configuration extensions and system extensions | |
579 | with such suffixes, to indicate their purpose in the name. | |
580 | ||
581 | * The sd-device API gained a new function | |
582 | sd_device_enumerator_add_match_property_required() which allows | |
583 | configuring matches on properties that are strictly required. This is | |
584 | different from the existing sd_device_enumerator_add_match_property() | |
585 | matches of which one one needs to apply. | |
586 | ||
e423b40d | 587 | * The MAC address the veth side of an nspawn container shall get |
c2322b48 LP |
588 | assigned may now be controlled via the $SYSTEMD_NSPAWN_NETWORK_MAC |
589 | environment variable. | |
590 | ||
28a8aac7 | 591 | * The libiptc dependency is now implemented via dlopen(), so that tools |
c2322b48 LP |
592 | such as networkd and nspawn no longer have a hard dependency on the |
593 | shared library when compiled with support for libiptc. | |
594 | ||
a4a047c3 ZJS |
595 | * New rpm macros have been added: %systemd_user_daemon_reexec does |
596 | daemon-reexec for all user managers, and %systemd_postun_with_reload | |
597 | and %systemd_user_postun_with_reload do a reload for system and user | |
598 | units on upgrades. | |
599 | ||
feed2910 LB |
600 | * coredumpctl now propagates SIGTERM to the debugger process. |
601 | ||
9a848052 LB |
602 | Contributions from: 김인수, Abderrahim Kitouni, Adam Goldman, |
603 | Adam Williamson, Alexandre Peixoto Ferreira, Alex Hudspith, | |
604 | Alvin Alvarado, André Paiusco, Antonio Alvarez Feijoo, | |
605 | Anton Lundin, Arian van Putten, Arseny Maslennikov, Arthur Shau, | |
606 | Balázs Úr, beh_10257, Benjamin Peterson, Bertrand Jacquin, | |
607 | Brian Norris, Charles Lee, Cheng-Chia Tseng, Chris Patterson, | |
608 | Christian Hergert, Christian Hesse, Christian Kirbach, | |
609 | Clayton Craft, commondservice, cunshunxia, Curtis Klein, cvlc12, | |
610 | Daan De Meyer, Daniele Medri, Daniel P. Berrangé, Daniel Rusek, | |
611 | Daniel Thompson, Dan Nicholson, Dan Streetman, David Rheinsberg, | |
612 | David Santamaría Rogado, David Tardon, dependabot[bot], | |
613 | Diego Viola, Dmitry V. Levin, Emanuele Giuseppe Esposito, | |
614 | Emil Renner Berthing, Emil Velikov, Etienne Dechamps, Fabian Vogt, | |
615 | felixdoerre, Felix Dörre, Florian Schmaus, Franck Bui, | |
616 | Frantisek Sumsal, G2-Games, Gioele Barabucci, Hugo Carvalho, | |
617 | huyubiao, Iago López Galeiras, IllusionMan1212, Jade Lovelace, | |
618 | janana, Jan Janssen, Jan Kuparinen, Jan Macku, Jeremy Fleischman, | |
619 | Jin Liu, jjimbo137, Joerg Behrmann, Johannes Segitz, Jordan Rome, | |
620 | Jordan Williams, Julien Malka, Juno Computers, Khem Raj, khm, | |
621 | Kingbom Dou, Kiran Vemula, Krzesimir Nowak, Laszlo Gombos, | |
622 | Lennart Poettering, linuxlion, Luca Boccassi, Lucas Adriano Salles, | |
623 | Lukas, Lukáš Nykrýn, Maanya Goenka, Maarten, Malte Poll, | |
624 | Marc Pervaz Boocha, Martin Beneš, Martin Joerg, Martin Wilck, | |
625 | Mathieu Tortuyaux, Matthias Schiffer, Maxim Mikityanskiy, | |
626 | Max Kellermann, Michael A Cassaniti, Michael Biebl, Michael Kuhn, | |
627 | Michael Vasseur, Michal Koutný, Michal Sekletár, Mike Yuan, | |
60142662 | 628 | Milton D. Miller II, mordner, msizanoen, NAHO, Nandakumar Raghavan, |
9a848052 LB |
629 | Neil Wilson, Nick Rosbrook, Nils K, NRK, Oğuz Ersen, |
630 | Omojola Joshua, onenowy, Paul Meyer, Paymon MARANDI, pelaufer, | |
631 | Peter Hutterer, PhylLu, Pierre GRASSER, Piotr Drąg, Priit Laes, | |
632 | Rahil Bhimjiani, Raito Bezarius, Raul Cheleguini, Reto Schneider, | |
e6e335b5 LB |
633 | Richard Maw, Robby Red, RoepLuke, Roland Hieber, Roland Singer, |
634 | Ronan Pigott, Sam James, Sam Leonard, Sergey A, Susant Sahani, | |
635 | Sven Joachim, Tad Fisher, Takashi Sakamoto, Thorsten Kukuk, Tj, | |
636 | Tomasz Świątek, Topi Miettinen, Valentin David, | |
637 | Valentin Lefebvre, Victor Westerhuis, Vincent Haupert, | |
638 | Vishal Chillara Srinivas, Vito Caputo, Warren, Weblate, | |
639 | Xiaotian Wu, xinpeng wang, Yaron Shahrabani, Yo-Jung Lin, | |
640 | Yu Watanabe, Zbigniew Jędrzejewski-Szmek, zeroskyx, | |
9a848052 | 641 | Дамјан Георгиевски, наб |
68ef4c57 | 642 | |
db11bab3 | 643 | — Edinburgh, 2023-12-06 |
0a917108 | 644 | |
994c7978 | 645 | CHANGES WITH 254: |
a1012609 | 646 | |
b4ff8ba0 | 647 | Announcements of Future Feature Removals and Incompatible Changes: |
648 | ||
d7b3c52c | 649 | * The next release (v255) will remove support for split-usr (/usr/ |
22900fa3 LB |
650 | mounted separately during late boot, instead of being mounted by the |
651 | initrd before switching to the rootfs) and unmerged-usr (parallel | |
33db1b90 | 652 | directories /bin/ and /usr/bin/, /lib/ and /usr/lib/, …). For more |
d7b3c52c | 653 | details, see: |
22900fa3 LB |
654 | https://lists.freedesktop.org/archives/systemd-devel/2022-September/048352.html |
655 | ||
3e938f1d LB |
656 | * We intend to remove cgroup v1 support from a systemd release after |
657 | the end of 2023. If you run services that make explicit use of | |
658 | cgroup v1 features (i.e. the "legacy hierarchy" with separate | |
659 | hierarchies for each controller), please implement compatibility with | |
660 | cgroup v2 (i.e. the "unified hierarchy") sooner rather than later. | |
661 | Most of Linux userspace has been ported over already. | |
b4ff8ba0 | 662 | |
7474097d LP |
663 | * Support for System V service scripts is now deprecated and will be |
664 | removed in a future release. Please make sure to update your software | |
665 | *now* to include a native systemd unit file instead of a legacy | |
666 | System V script to retain compatibility with future systemd releases. | |
667 | ||
7e90814f ZJS |
668 | * Support for the SystemdOptions EFI variable is deprecated. |
669 | 'bootctl systemd-efi-options' will emit a warning when used. It seems | |
670 | that this feature is little-used and it is better to use alternative | |
671 | approaches like credentials and confexts. The plan is to drop support | |
672 | altogether at a later point, but this might be revisited based on | |
673 | user feedback. | |
674 | ||
3e938f1d LB |
675 | * EnvironmentFile= now treats the line following a comment line |
676 | trailing with escape as a non comment line. For details, see: | |
677 | https://github.com/systemd/systemd/issues/27975 | |
678 | ||
6b2d576f ED |
679 | * PrivateNetwork=yes and NetworkNamespacePath= now imply |
680 | PrivateMounts=yes unless PrivateMounts=no is explicitly specified. | |
681 | ||
221332ee ZJS |
682 | * Behaviour of sandboxing options for the per-user service manager |
683 | units has changed. They now imply PrivateUsers=yes, which means user | |
684 | namespaces will be implicitly enabled when a sandboxing option is | |
b23c7e14 | 685 | enabled in a user unit. Enabling user namespaces has the drawback |
221332ee ZJS |
686 | that system users will no longer be visible (and processes/files will |
687 | appear as owned by 'nobody') in the user unit. | |
688 | ||
689 | By definition a sandboxed user unit should run with reduced | |
690 | privileges, so impact should be small. This will remove a great | |
691 | source of confusion that has been reported by users over the years, | |
692 | due to how these options require an extra setting to be manually | |
693 | enabled when used in the per-user service manager, which is not | |
694 | needed in the system service manager. For more details, see: | |
e6da1e04 LB |
695 | https://lists.freedesktop.org/archives/systemd-devel/2022-December/048682.html |
696 | ||
d0555379 LB |
697 | * systemd-run's switch --expand-environment= which currently is disabled |
698 | by default when combined with --scope, will be changed in a future | |
699 | release to be enabled by default. | |
700 | ||
d7b3c52c | 701 | Security Relevant Changes: |
a1012609 LP |
702 | |
703 | * pam_systemd will now by default pass the CAP_WAKE_ALARM ambient | |
704 | process capability to invoked session processes of regular users on | |
705 | local seats (as well as to systemd --user), unless configured | |
706 | otherwise via data from JSON user records, or via the PAM module's | |
707 | parameter list. This is useful in order allow desktop tools such as | |
708 | GNOME's Alarm Clock application to set a timer for | |
709 | CLOCK_REALTIME_ALARM that wakes up the system when it elapses. A | |
710 | per-user service unit file may thus use AmbientCapability= to pass | |
711 | the capability to invoked processes. Note that this capability is | |
712 | relatively narrow in focus (in particular compared to other process | |
713 | capabilities such as CAP_SYS_ADMIN) and we already — by default — | |
714 | permit more impactful operations such as system suspend to local | |
715 | users. | |
716 | ||
d7b3c52c LP |
717 | Service Manager: |
718 | ||
e9ae4187 ZJS |
719 | * Memory limits that apply while the unit is activating are now |
720 | supported. Previously IO and CPU settings were already supported via | |
721 | StartupCPUWeight= and similar. The same logic has been added for the | |
722 | various manager and unit memory settings (DefaultStartupMemoryLow=, | |
723 | StartupMemoryLow=, StartupMemoryHigh=, StartupMemoryMax=, | |
724 | StartupMemorySwapMax=, StartupMemoryZSwapMax=). | |
d7b3c52c LP |
725 | |
726 | * The service manager gained support for enqueuing POSIX signals to | |
727 | services that carry an additional integer value, exposing the | |
49bf8bd5 | 728 | sigqueue() system call. This is accessible via new D-Bus calls |
221332ee ZJS |
729 | org.freedesktop.systemd1.Manager.QueueSignalUnit() and |
730 | org.freedesktop.systemd1.Unit.QueueSignal(), as well as in systemctl | |
731 | via the new --kill-value= option. | |
d7b3c52c LP |
732 | |
733 | * systemctl gained a new "list-paths" verb, which shows all currently | |
221332ee | 734 | active .path units, similarly to how "systemctl list-timers" shows |
d7b3c52c LP |
735 | active timers, and "systemctl list-sockets" shows active sockets. |
736 | ||
d7b3c52c LP |
737 | * systemctl gained a new --when= switch which is honoured by the various |
738 | forms of shutdown (i.e. reboot, kexec, poweroff, halt) and allows | |
739 | scheduling these operations by time, similar in fashion to how this | |
740 | has been supported by SysV shutdown. | |
741 | ||
221332ee ZJS |
742 | * If MemoryDenyWriteExecute= is enabled for a service and the kernel |
743 | supports the new PR_SET_MDWE prctl() call, it is used instead of the | |
744 | seccomp()-based system call filter to achieve the same effect. | |
745 | ||
d7b3c52c LP |
746 | * A new set of kernel command line options is now understood: |
747 | systemd.tty.term.<name>=, systemd.tty.rows.<name>=, | |
748 | systemd.tty.columns.<name>= allow configuring the TTY type and | |
b1ee7474 LB |
749 | dimensions for the tty specified via <name>. When systemd invokes a |
750 | service on a tty (via TTYName=) it will look for these and configure | |
221332ee | 751 | the TTY accordingly. This is particularly useful in VM environments |
d7b3c52c LP |
752 | to propagate host terminal settings into the appropriate TTYs of the |
753 | guest. | |
754 | ||
755 | * A new RootEphemeral= setting is now understood in service units. It | |
756 | takes a boolean argument. If enabled for services that use RootImage= | |
757 | or RootDirectory= an ephemeral copy of the disk image or directory | |
758 | tree is made when the service is started. It is removed automatically | |
08423f6d | 759 | when the service is stopped. That ephemeral copy is made using |
dc3b5e04 | 760 | btrfs/xfs reflinks or btrfs snapshots, if available. |
d7b3c52c LP |
761 | |
762 | * The service activation logic gained new settings RestartSteps= and | |
221332ee | 763 | RestartMaxDelaySec= which allow exponentially-growing restart |
d7b3c52c LP |
764 | intervals for Restart=. |
765 | ||
305bea82 LB |
766 | * The service activation logic gained a new setting RestartMode= which |
767 | can be set to 'direct' to skip the inactive/failed states when | |
768 | restarting, so that dependent units are not notified until the service | |
769 | converges to a final (successful or failed) state. For example, this | |
770 | means that OnSuccess=/OnFailure= units will not be triggered until the | |
771 | service state has converged. | |
772 | ||
d7b3c52c LP |
773 | * PID 1 will now automatically load the virtio_console kernel module |
774 | during early initialization if running in a suitable VM. This is done | |
775 | so that early-boot logging can be written to the console if available. | |
776 | ||
221332ee ZJS |
777 | * Similarly, virtio-vsock support is loaded early in suitable VM |
778 | environments. PID 1 will send sd_notify() notifications via AF_VSOCK | |
779 | to the VMM if configured, thus loading this early is beneficial. | |
d7b3c52c LP |
780 | |
781 | * A new verb "fdstore" has been added to systemd-analyze to show the | |
782 | current contents of the file descriptor store of a unit. This is | |
783 | backed by a new D-Bus call DumpUnitFileDescriptorStore() provided by | |
784 | the service manager. | |
785 | ||
786 | * The service manager will now set a new $FDSTORE environment variable | |
787 | when invoking processes for services that have the file descriptor | |
788 | store enabled. | |
789 | ||
790 | * A new service option FileDescriptorStorePreserve= has been added that | |
55e40b0b ZJS |
791 | allows tuning the lifecycle of the per-service file descriptor store. |
792 | If set to "yes", the entries in the fd store are retained even after | |
793 | the service has been fully stopped. | |
d7b3c52c LP |
794 | |
795 | * The "systemctl clean" command may now be used to clear the fdstore of | |
796 | a service. | |
797 | ||
d7b3c52c | 798 | * Unit *.preset files gained a new directive "ignore", in addition to |
221332ee ZJS |
799 | the existing "enable" and "disable". As the name suggests, matching |
800 | units are left unchanged, i.e. neither enabled nor disabled. | |
d7b3c52c LP |
801 | |
802 | * Service units gained a new setting DelegateSubgroup=. It takes the | |
803 | name of a sub-cgroup to place any processes the service manager forks | |
804 | off in. Previously, the service manager would place all service | |
221332ee ZJS |
805 | processes directly in the top-level cgroup it created for the |
806 | service. This usually meant that main process in a service with | |
807 | delegation enabled would first have to create a subgroup and move | |
808 | itself down into it, in order to not conflict with the "no processes | |
809 | in inner cgroups" rule of cgroup v2. With this option, this step is | |
810 | now handled by PID 1. | |
811 | ||
812 | * The service manager will now look for .upholds/ directories, | |
813 | similarly to the existing support for .wants/ and .requires/ | |
814 | directories. Symlinks in this directory result in Upholds= | |
815 | dependencies. | |
816 | ||
817 | The [Install] section of unit files gained support for a new | |
818 | UpheldBy= directive to generate .upholds/ symlinks automatically when | |
819 | a unit is enabled. | |
d7b3c52c LP |
820 | |
821 | * The service manager now supports a new kernel command line option | |
822 | systemd.default_device_timeout_sec=, which may be used to override | |
823 | the default timeout for .device units. | |
824 | ||
221332ee ZJS |
825 | * A new "soft-reboot" mechanism has been added to the service manager. |
826 | A "soft reboot" is similar to a regular reboot, except that it | |
827 | affects userspace only: the service manager shuts down any running | |
d7b3c52c LP |
828 | services and other units, then optionally switches into a new root |
829 | file system (mounted to /run/nextroot/), and then passes control to a | |
49bf8bd5 | 830 | systemd instance in the new file system which then starts the system |
221332ee ZJS |
831 | up again. The kernel is not rebooted and neither is the hardware, |
832 | firmware or boot loader. This provides a fast, lightweight mechanism | |
833 | to quickly reset or update userspace, without the latency that a full | |
d7b3c52c LP |
834 | system reset involves. Moreover, open file descriptors may be passed |
835 | across the soft reboot into the new system where they will be passed | |
836 | back to the originating services. This allows pinning resources | |
7f0bf48d | 837 | across the reboot, thus minimizing grey-out time further. This new |
d7b3c52c LP |
838 | reboot mechanism is accessible via the new "systemctl soft-reboot" |
839 | command. | |
840 | ||
25c66abb LB |
841 | * Services using RootDirectory= or RootImage= will now have read-only |
842 | access to a copy of the host's os-release file under | |
843 | /run/host/os-release, which will be kept up-to-date on 'soft-reboot'. | |
844 | This was already the case for Portable Services, and the feature has | |
845 | now been extended to all services that do not run off the host's | |
846 | root filesystem. | |
847 | ||
221332ee ZJS |
848 | * A new service setting MemoryKSM= has been added to enable kernel |
849 | same-page merging individually for services. | |
d7b3c52c | 850 | |
08423f6d LP |
851 | * A new service setting ImportCredentials= has been added that augments |
852 | LoadCredential= and LoadCredentialEncrypted= and searches for | |
853 | credentials to import from the system, and supports globbing. | |
d7b3c52c | 854 | |
eade959b LP |
855 | * A new job mode "restart-dependencies" has been added to the service |
856 | manager (exposed via systemctl --job-mode=). It is only valid when | |
857 | used with "start" jobs, and has the effect that the "start" job will | |
858 | be propagated as "restart" jobs to currently running units that have | |
859 | a BindsTo= or Requires= dependency on the started unit. | |
860 | ||
861 | * A new verb "whoami" has been added to "systemctl" which determines as | |
862 | part of which unit the command is being invoked. It writes the unit | |
863 | name to standard output. If one or more PIDs are specified reports | |
864 | the unit names the processes referenced by the PIDs belong to. | |
865 | ||
866 | * The system and service credential logic has been improved: there's | |
867 | now a clearly defined place where system provisioning tools running | |
868 | in the initrd can place credentials that will be imported into the | |
869 | system's set of credentials during the initrd → host transition: the | |
870 | /run/credentials/@initrd/ directory. Once the credentials placed | |
871 | there are imported into the system credential set they are deleted | |
872 | from this directory, and the directory itself is deleted afterwards | |
873 | too. | |
874 | ||
875 | * A new kernel command line option systemd.set_credential_binary= has | |
876 | been added, that is similar to the pre-existing | |
877 | systemd.set_credential= but accepts arbitrary binary credential data, | |
878 | encoded in Base64. Note that the kernel command line is not a | |
879 | recommend way to transfer credentials into a system, since it is | |
880 | world-readable from userspace. | |
881 | ||
882 | * The default machine ID to use may now be configured via the | |
883 | system.machine_id system credential. It will only be used if no | |
884 | machine ID was set yet on the host. | |
885 | ||
886 | * On Linux kernel 6.4 and newer system and service credentials will now | |
887 | be placed in a tmpfs instance that has the "noswap" mount option | |
888 | set. Previously, a "ramfs" instance was used. By switching to tmpfs | |
889 | ACL support and overall size limits can now be enforced, without | |
890 | compromising on security, as the memory is never paged out either | |
891 | way. | |
892 | ||
305bea82 LB |
893 | * The service manager now can detect when it is running in a |
894 | 'Confidential Virtual Machine', and a corresponding 'cvm' value is now | |
895 | accepted by ConditionSecurity= for units that want to conditionalize | |
896 | themselves on this. systemd-detect-virt gained new 'cvm' and | |
897 | '--list-cvm' switches to respectively perform the detection or list | |
898 | all known flavours of confidential VM, depending on the vendor. The | |
899 | manager will publish a 'ConfidentialVirtualization' D-Bus property, | |
900 | and will also set a SYSTEMD_CONFIDENTIAL_VIRTUALIZATION= environment | |
901 | variable for unit generators. Finally, udev rules can match on a new | |
902 | 'cvm' key that will be set when in a confidential VM. | |
d4c124a2 LB |
903 | Additionally, when running in a 'Confidential Virtual Machine', SMBIOS |
904 | strings and QEMU's fw_cfg protocol will not be used to import | |
905 | credentials and kernel command line parameters by the system manager, | |
906 | systemd-boot and systemd-stub, because the hypervisor is considered | |
907 | untrusted in this particular setting. | |
305bea82 | 908 | |
d7b3c52c LP |
909 | Journal: |
910 | ||
221332ee ZJS |
911 | * The sd-journal API gained a new call sd_journal_get_seqnum() to |
912 | retrieve the current log record's sequence number and sequence number | |
913 | ID, which allows applications to order records the same way as | |
914 | journal does internally. The sequence number is now also exported in | |
915 | the JSON and "export" output of the journal. | |
d7b3c52c LP |
916 | |
917 | * journalctl gained a new switch --truncate-newline. If specified | |
918 | multi-line log records will be truncated at the first newline, | |
221332ee | 919 | i.e. only the first line of each log message will be shown. |
d7b3c52c | 920 | |
eade959b LP |
921 | * systemd-journal-upload gained support for --namespace=, similar to |
922 | the switch of the same name of journalctl. | |
923 | ||
d7b3c52c LP |
924 | systemd-repart: |
925 | ||
926 | * systemd-repart's drop-in files gained a new ExcludeFiles= option which | |
221332ee | 927 | may be used to exclude certain files from the effect of CopyFiles=. |
d7b3c52c LP |
928 | |
929 | * systemd-repart's Verity support now implements the Minimize= setting | |
930 | to minimize the size of the resulting partition. | |
931 | ||
932 | * systemd-repart gained a new --offline= switch, which may be used to | |
933 | control whether images shall be built "online" or "offline", | |
934 | i.e. whether to make use of kernel facilities such as loopback block | |
221332ee | 935 | devices and device mapper or not. |
d7b3c52c LP |
936 | |
937 | * If systemd-repart is told to populate a newly created ESP or XBOOTLDR | |
221332ee ZJS |
938 | partition with some files, it will now default to VFAT rather than |
939 | ext4. | |
d7b3c52c LP |
940 | |
941 | * systemd-repart gained a new --architecture= switch. If specified, the | |
942 | per-architecture GPT partition types (i.e. the root and /usr/ | |
943 | partitions) configured in the partition drop-in files are | |
08423f6d LP |
944 | automatically adjusted to match the specified CPU architecture, in |
945 | order to simplify cross-architecture DDI building. | |
d7b3c52c | 946 | |
739bbc58 LB |
947 | * systemd-repart will now default to a minimum size of 300MB for XFS |
948 | filesystems if no size parameter is specified. This matches what the | |
949 | XFS tools (xfsprogs) can support. | |
950 | ||
d7b3c52c LP |
951 | systemd-boot, systemd-stub, ukify, bootctl, kernel-install: |
952 | ||
51bb4221 LB |
953 | * gnu-efi is no longer required to build systemd-boot and systemd-stub. |
954 | Instead, pyelftools is now needed, and it will be used to perform the | |
955 | ELF -> PE relocations at build time. | |
956 | ||
221332ee | 957 | * bootctl gained a new switch --print-root-device/-R that prints the |
d5163f9c LP |
958 | block device the root file system is backed by. If specified twice, |
959 | it returns the whole disk block device (as opposed to partition block | |
960 | device) the root file system is on. It's useful for invocations such | |
961 | as "cfdisk $(bootctl -RR)" to quickly show the partition table of the | |
962 | running OS. | |
d7b3c52c LP |
963 | |
964 | * systemd-stub will now look for the SMBIOS Type 1 field | |
965 | "io.systemd.stub.kernel-cmdline-extra" and append its value to the | |
966 | kernel command line it invokes. This is useful for VMMs such as qemu | |
967 | to pass additional kernel command lines into the system even when | |
221332ee | 968 | booting via full UEFI. The contents of the field are measured into |
7eff3e2c | 969 | TPM PCR 12. |
d7b3c52c LP |
970 | |
971 | * The KERNEL_INSTALL_LAYOUT= setting for kernel-install gained a new | |
221332ee ZJS |
972 | value "auto". With this value, a kernel will be automatically |
973 | analyzed, and if it qualifies as UKI, it will be installed as if the | |
974 | setting was to set to "uki", otherwise as "bls". | |
d7b3c52c LP |
975 | |
976 | * systemd-stub can now optionally load UEFI PE "add-on" images that may | |
977 | contain additional kernel command line information. These "add-ons" | |
978 | superficially look like a regular UEFI executable, and are expected | |
979 | to be signed via SecureBoot/shim. However, they do not actually | |
980 | contain code, but instead a subset of the PE sections that UKIs | |
981 | support. They are supposed to provide a way to extend UKIs with | |
982 | additional resources in a secure and authenticated way. Currently, | |
983 | only the .cmdline PE section may be used in add-ons, in which case | |
984 | any specified string is appended to the command line embedded into | |
acf678de LB |
985 | the UKI itself. A new 'addon<EFI-ARCH>.efi.stub' is now provided that |
986 | can be used to trivially create addons, via 'ukify' or 'objcopy'. In | |
987 | the future we expect other sections to be made extensible like this as | |
988 | well. | |
d7b3c52c LP |
989 | |
990 | * ukify has been updated to allow building these UEFI PE "add-on" | |
acf678de | 991 | images, using the new 'addon<EFI-ARCH>.efi.stub'. |
d7b3c52c | 992 | |
eade959b | 993 | * ukify now accepts SBAT information to place in the .sbat PE section |
5bc9ea07 | 994 | of UKIs and addons. If a UKI is built the SBAT information from the |
eade959b LP |
995 | inner kernel is merged with any SBAT information associated with |
996 | systemd-stub and the SBAT data specified on the ukify command line. | |
997 | ||
d7b3c52c | 998 | * The kernel-install script has been rewritten in C, and reuses much of |
49bf8bd5 JB |
999 | the infrastructure of existing tools such as bootctl. It also gained |
1000 | --esp-path= and --boot-path= options to override the path to the ESP, | |
1001 | and the $BOOT partition. Options --make-entry-directory= and | |
1002 | --entry-token= have been added as well, similar to bootctl's options | |
1003 | of the same name. | |
d7b3c52c LP |
1004 | |
1005 | * A new kernel-install plugin 60-ukify has been added which will | |
5bc9ea07 | 1006 | combine kernel/initrd locally into a UKI and optionally sign them |
221332ee ZJS |
1007 | with a local key. This may be used to switch to UKI mode even on |
1008 | systems where a local kernel or initrd is used. (Typically UKIs are | |
1009 | built and signed by the vendor.) | |
d7b3c52c | 1010 | |
221332ee | 1011 | * The ukify tool now supports "pesign" in addition to the pre-existing |
d7b3c52c LP |
1012 | "sbsign" for signing UKIs. |
1013 | ||
221332ee ZJS |
1014 | * systemd-measure and systemd-stub now look for the .uname PE section |
1015 | that should contain the kernel's "uname -r" string. | |
d7b3c52c | 1016 | |
221332ee ZJS |
1017 | * systemd-measure and ukify now calculate expected PCR hashes for a UKI |
1018 | "offline", i.e. without access to a TPM (physical or | |
1019 | software-emulated). | |
d7b3c52c LP |
1020 | |
1021 | Memory Pressure & Control: | |
1022 | ||
1023 | * The sd-event API gained new calls sd_event_add_memory_pressure(), | |
1024 | sd_event_source_set_memory_pressure_type(), | |
221332ee ZJS |
1025 | sd_event_source_set_memory_pressure_period() to create and configure |
1026 | an event source that is called whenever the OS signals memory | |
1027 | pressure. Another call sd_event_trim_memory() is provided that | |
d7b3c52c | 1028 | compacts the process' memory use by releasing allocated but unused |
198aff76 LB |
1029 | malloc() memory back to the kernel. Services can also provide their |
1030 | own custom callback to do memory trimming. This should improve system | |
221332ee ZJS |
1031 | behaviour under memory pressure, as on Linux traditionally provided |
1032 | no mechanism to return process memory back to the kernel if the | |
1033 | kernel was under memory pressure. This makes use of the kernel's PSI | |
1034 | interface. Most long-running services in systemd have been hooked up | |
1035 | with this, and in particular systems with low memory should benefit | |
1036 | from this. | |
1037 | ||
1038 | * Service units gained new settings MemoryPressureWatch= and | |
1039 | MemoryPressureThresholdSec= to configure the PSI memory pressure | |
1040 | logic individually. If these options are used, the | |
d7b3c52c | 1041 | $MEMORY_PRESSURE_WATCH and $MEMORY_PRESSURE_WRITE environment |
49bf8bd5 JB |
1042 | variables will be set for the invoked processes to inform them about |
1043 | the requested memory pressure behaviour. (This is used by the | |
1044 | aforementioned sd-events API additions, if set.) | |
d7b3c52c LP |
1045 | |
1046 | * systemd-analyze gained a new "malloc" verb that shows the output | |
1047 | generated by glibc's malloc_info() on services that support it. Right | |
221332ee ZJS |
1048 | now, only the service manager has been updated accordingly. This |
1049 | call requires privileges. | |
d7b3c52c LP |
1050 | |
1051 | User & Session Management: | |
1052 | ||
221332ee ZJS |
1053 | * The sd-login API gained a new call sd_session_get_username() to |
1054 | return the user name of the owner of a login session. It also gained | |
1055 | a new call sd_session_get_start_time() to retrieve the time the login | |
1056 | session started. A new call sd_session_get_leader() has been added to | |
1057 | return the PID of the "leader" process of a session. A new call | |
1058 | sd_uid_get_login_time() returns the time since the specified user has | |
1059 | most recently been continuously logged in with at least one session. | |
d7b3c52c LP |
1060 | |
1061 | * JSON user records gained a new set of fields capabilityAmbientSet and | |
1062 | capabilityBoundingSet which contain a list of POSIX capabilities to | |
1063 | set for the logged in users in the ambient and bounding sets, | |
1064 | respectively. homectl gained the ability to configure these two sets | |
1065 | for users via --capability-bounding-set=/--capability-ambient-set=. | |
1066 | ||
1067 | * pam_systemd learnt two new module options | |
221332ee ZJS |
1068 | default-capability-bounding-set= and default-capability-ambient-set=, |
1069 | which configure the default bounding sets for users as they are | |
1070 | logging in, if the JSON user record doesn't specify this explicitly | |
1071 | (see above). The built-in default for the ambient set now contains | |
1072 | the CAP_WAKE_ALARM, thus allowing regular users who may log in | |
1073 | locally to resume from a system suspend via a timer. | |
d7b3c52c | 1074 | |
49bf8bd5 | 1075 | * The Session D-Bus objects systemd-logind gained a new SetTTY() method |
221332ee ZJS |
1076 | call to update the TTY of a session after it has been allocated. This |
1077 | is useful for SSH sessions which are typically allocated first, and | |
1078 | for which a TTY is added later. | |
d7b3c52c LP |
1079 | |
1080 | * The sd-login API gained a new call sd_pid_notifyf_with_fds() which | |
1081 | combines the various other sd_pid_notify() flavours into one: takes a | |
1082 | format string, an overriding PID, and a set of file descriptors to | |
221332ee ZJS |
1083 | send. It also gained a new call sd_pid_notify_barrier() call which is |
1084 | equivalent to sd_notify_barrier() but allows the originating PID to | |
1085 | be specified. | |
d7b3c52c LP |
1086 | |
1087 | * "loginctl list-users" and "loginctl list-sessions" will now show the | |
1088 | state of each logged in user/session in their tabular output. It will | |
1089 | also show the current idle state of sessions. | |
1090 | ||
1091 | DDIs: | |
1092 | ||
1093 | * systemd-dissect will now show the intended CPU architecture of an | |
1094 | inspected DDI. | |
1095 | ||
1096 | * systemd-dissect will now install itself as mount helper for the "ddi" | |
1097 | pseudo-file system type. This means you may now mount DDIs directly | |
1098 | via /bin/mount or /etc/fstab, making full use of embedded Verity | |
221332ee ZJS |
1099 | information and all other DDI features. |
1100 | ||
1101 | Example: mount -t ddi myimage.raw /some/where | |
d7b3c52c | 1102 | |
221332ee ZJS |
1103 | * The systemd-dissect tool gained the new switches --attach/--detach to |
1104 | attach/detach a DDI to a loopback block device without mounting it. | |
1105 | It will automatically derive the right sector size from the image | |
1106 | and set up Verity and similar, but not mount the file systems in it. | |
d7b3c52c | 1107 | |
221332ee ZJS |
1108 | * When systemd-gpt-auto-generator or the DDI mounting logic mount an |
1109 | ESP or XBOOTLDR partition the MS_NOSYMFOLLOW mount option is now | |
1110 | implied. Given that these file systems are typically untrusted, this | |
1111 | should make mounting them automatically have less of a security | |
1112 | impact. | |
d7b3c52c LP |
1113 | |
1114 | * All tools that parse DDIs (such as systemd-nspawn, systemd-dissect, | |
1115 | systemd-tmpfiles, …) now understand a new switch --image-policy= which | |
1116 | takes a string encoding image dissection policy. With this mechanism | |
1117 | automatic discovery and use of specific partition types and the | |
1118 | cryptographic requirements on the partitions (Verity, LUKS, …) can be | |
1119 | restricted, permitting better control of the exposed attack surfaces | |
1120 | when mounting disk images. systemd-gpt-auto-generator will honour such | |
1121 | an image policy too, configurable via the systemd.image_policy= kernel | |
1122 | command line option. Unit files gained the RootImagePolicy=, | |
1123 | MountImagePolicy= and ExtensionImagePolicy= to configure the same for | |
1124 | disk images a service runs off. | |
1125 | ||
221332ee ZJS |
1126 | * systemd-analyze gained a new verb "image-policy" to validate and |
1127 | parse image policy strings. | |
d7b3c52c | 1128 | |
221332ee ZJS |
1129 | * systemd-dissect gained support for a new --validate switch to |
1130 | superficially validate DDI structure, and check whether a specific | |
1131 | image policy allows the DDI. | |
d7b3c52c | 1132 | |
305bea82 LB |
1133 | * systemd-dissect gained support for a new --mtree-hash switch to |
1134 | optionally disable calculating mtree hashes, which can be slow on | |
1135 | large images. | |
1136 | ||
1137 | * systemd-dissect --copy-to, --copy-from, --list and --mtree switches | |
1138 | are now able to operate on directories too, other than images. | |
1139 | ||
d7b3c52c LP |
1140 | Network Management: |
1141 | ||
1142 | * networkd's GENEVE support as gained a new .network option | |
1143 | InheritInnerProtocol=. | |
1144 | ||
eade959b LP |
1145 | * The [Tunnel] section in .netdev files has gained a new setting |
1146 | IgnoreDontFragment for controlling the IPv4 "DF" flag of datagrams. | |
1147 | ||
1148 | * A new global IPv6PrivacyExtensions= setting has been added that | |
1149 | selects the default value of the per-network setting of the same | |
1150 | name. | |
1151 | ||
64f2cf77 ZJS |
1152 | * The predictable network interface naming logic was extended to |
1153 | include SR-IOV-R "representor" information in network interface | |
1154 | names. Unfortunately, this feature was not enabled by default and can | |
1155 | only be enabled at compilation time by setting | |
1156 | -Ddefault-net-naming-scheme=v254. | |
eade959b LP |
1157 | |
1158 | * The DHCPv4 + DHCPv6 + IPv6 RA logic in networkd gained support for | |
1159 | the RFC8910 captive portal option. | |
1160 | ||
d7b3c52c LP |
1161 | Device Management: |
1162 | ||
1163 | * udevadm gained the new "verify" verb for validating udev rules files | |
1164 | offline. | |
1165 | ||
d7b3c52c LP |
1166 | * udev gained a new tool "iocost" that can be used to configure QoS IO |
1167 | cost data based on hwdb information onto suitable block devices. Also | |
1168 | see https://github.com/iocost-benchmark/iocost-benchmarks. | |
1169 | ||
1170 | TPM2 Support + Disk Encryption & Authentication: | |
1171 | ||
1172 | * systemd-cryptenroll/systemd-cryptsetup will now install a TPM2 SRK | |
b1ee7474 | 1173 | ("Storage Root Key") as first step in the TPM2, and then use that |
d7b3c52c LP |
1174 | for binding FDE to, if TPM2 support is used. This matches |
1175 | recommendations of TCG (see | |
1176 | https://trustedcomputinggroup.org/wp-content/uploads/TCG-TPM-v2.0-Provisioning-Guidance-Published-v1r1.pdf) | |
1177 | ||
1178 | * systemd-cryptenroll and other tools that take TPM2 PCR parameters now | |
1179 | understand textual identifiers for these PCRs. | |
1180 | ||
1181 | * systemd-veritysetup + /etc/veritytab gained support for a series of | |
1182 | new options: hash-offset=, superblock=, format=, data-block-size=, | |
1183 | hash-block-size=, data-blocks=, salt=, uuid=, hash=, fec-device=, | |
1184 | fec-offset=, fec-roots= to configure various aspects of a Verity | |
1185 | volume. | |
1186 | ||
1187 | * systemd-cryptsetup + /etc/crypttab gained support for a new | |
1188 | veracrypt-pim= option for setting the Personal Iteration Multiplier | |
1189 | of veracrypt volumes. | |
1190 | ||
1191 | * systemd-integritysetup + /etc/integritytab gained support for a new | |
1192 | mode= setting for controlling the dm-integrity mode (journal, bitmap, | |
1193 | direct) for the volume. | |
1194 | ||
005bfe4e LP |
1195 | * systemd-analyze gained a new verb "pcrs" that shows the known TPM PCR |
1196 | registers, their symbolic names and current values. | |
1197 | ||
d7b3c52c LP |
1198 | systemd-tmpfiles: |
1199 | ||
1200 | * The ACL support in tmpfiles.d/ has been updated: if an uppercase "X" | |
1201 | access right is specified this is equivalent to "x" but only if the | |
1202 | inode in question already has the executable bit set for at least | |
1203 | some user/group. Otherwise the "x" bit will be turned off. | |
1204 | ||
1205 | * tmpfiles.d/'s C line type now understands a new modifier "+": a line | |
1206 | with C+ will result in a "merge" copy, i.e. all files of the source | |
1207 | tree are copied into the target tree, even if that tree already | |
1208 | exists, resulting in a combined tree of files already present in the | |
1209 | target tree and those copied in. | |
1210 | ||
1211 | * systemd-tmpfiles gained a new --graceful switch. If specified lines | |
1212 | with unknown users/groups will silently be skipped. | |
1213 | ||
1214 | systemd-notify: | |
1215 | ||
1216 | * systemd-notify gained two new options --fd= and --fdname= for sending | |
1217 | arbitrary file descriptors to the service manager (while specifying an | |
1218 | explicit name for it). | |
1219 | ||
1220 | * systemd-notify gained a new --exec switch, which makes it execute the | |
1221 | specified command line after sending the requested messages. This is | |
1222 | useful for sending out READY=1 first, and then continuing invocation | |
1223 | without changing process ID, so that the tool can be nicely used | |
ffe7ddb9 | 1224 | within an ExecStart= line of a unit file that uses Type=notify. |
d7b3c52c LP |
1225 | |
1226 | sd-event + sd-bus APIs: | |
1227 | ||
1228 | * The sd-event API gained a new call sd_event_source_leave_ratelimit() | |
1229 | which may be used to explicitly end a rate-limit state an event | |
1230 | source might be in, resetting all rate limiting counters. | |
1231 | ||
1232 | * When the sd-bus library is used to make connections to AF_UNIX D-Bus | |
221332ee ZJS |
1233 | sockets, it will now encode the "description" set via |
1234 | sd_bus_set_description() into the source socket address. It will also | |
d7b3c52c LP |
1235 | look for this information when accepting a connection. This is useful |
1236 | to track individual D-Bus connections on a D-Bus broker for debug | |
1237 | purposes. | |
1238 | ||
1239 | systemd-resolved: | |
1240 | ||
1241 | * systemd-resolved gained a new resolved.conf setting | |
1242 | StateRetentionSec= which may be used to retain cached DNS records | |
1243 | even after their nominal TTL, and use them in case upstream DNS | |
c23b07df | 1244 | servers cannot be reached. This can be used to make name resolution |
221332ee | 1245 | more resilient in case of network problems. |
d7b3c52c | 1246 | |
221332ee | 1247 | * resolvectl gained a new verb "show-cache" to show the current cache |
627cdcc7 | 1248 | contents of systemd-resolved. This verb communicates with the |
221332ee | 1249 | systemd-resolved daemon and requires privileges. |
d7b3c52c LP |
1250 | |
1251 | Other: | |
1252 | ||
6e522aa3 LB |
1253 | * Meson >= 0.60.0 is now required to build systemd. |
1254 | ||
d7b3c52c | 1255 | * The default keymap to apply may now be chosen at build-time via the |
221332ee | 1256 | new -Ddefault-keymap= meson option. |
d7b3c52c LP |
1257 | |
1258 | * Most of systemd's long-running services now have a generic handler of | |
1259 | the SIGRTMIN+18 signal handler which executes various operations | |
1260 | depending on the sigqueue() parameter sent along. For example, values | |
1261 | 0x100…0x107 allow changing the maximum log level of such | |
1262 | services. 0x200…0x203 allow changing the log target of such | |
221332ee ZJS |
1263 | services. 0x300 make the services trim their memory similarly to the |
1264 | automatic PSI-triggered action, see above. 0x301 make the services | |
d7b3c52c LP |
1265 | output their malloc_info() data to the logs. |
1266 | ||
1267 | * machinectl gained new "edit" and "cat" verbs for editing .nspawn | |
221332ee ZJS |
1268 | files, inspired by systemctl's verbs of the same name which edit unit |
1269 | files. Similarly, networkctl gained the same verbs for editing | |
d7b3c52c LP |
1270 | .network, .netdev, .link files. |
1271 | ||
1272 | * A new syscall filter group "@sandbox" has been added that contains | |
1273 | syscalls for sandboxing system calls such as those for seccomp and | |
1274 | Landlock. | |
1275 | ||
1276 | * New documentation has been added: | |
1277 | ||
1278 | https://systemd.io/COREDUMP | |
1279 | https://systemd.io/MEMORY_PRESSURE | |
eade959b | 1280 | smbios-type-11(7) |
d7b3c52c | 1281 | |
221332ee ZJS |
1282 | * systemd-firstboot gained a new --reset option. If specified, the |
1283 | settings in /etc/ it knows how to initialize are reset. | |
d7b3c52c | 1284 | |
221332ee ZJS |
1285 | * systemd-sysext is now a multi-call binary and is also installed under |
1286 | the systemd-confext alias name (via a symlink). When invoked that way | |
1287 | it will operate on /etc/ instead of /usr/ + /opt/. It thus becomes a | |
d7b3c52c LP |
1288 | powerful, atomic, secure configuration management of sorts, that |
1289 | locally can merge configuration from multiple confext configuration | |
1290 | images into a single immutable tree. | |
1291 | ||
1292 | * The --network-macvlan=, --network-ipvlan=, --network-interface= | |
1293 | switches of systemd-nspawn may now optionally take the intended | |
1294 | network interface inside the container. | |
1295 | ||
1296 | * All our programs will now send an sd_notify() message with their exit | |
1297 | status in the EXIT_STATUS= field when exiting, using the usual | |
1298 | protocol, including PID 1. This is useful for VMMs and container | |
1299 | managers to collect an exit status from a system as it shuts down, as | |
1300 | set via "systemctl exit …". This is particularly useful in test cases | |
1301 | and similar, as invocations via a VM can now nicely propagate an exit | |
1302 | status to the host, similar to local processes. | |
1303 | ||
1304 | * systemd-run gained a new switch --expand-environment=no to disable | |
49bf8bd5 | 1305 | server-side environment variable expansion in specified command |
d0555379 LB |
1306 | lines. Expansion defaults to enabled for all execution types except |
1307 | --scope, where it defaults to off (and prints a warning) for backward | |
b77ff21a ZJS |
1308 | compatibility reasons. --scope will be flipped to enabled by default |
1309 | too in a future release. If you are using --scope and passing a '$' | |
d0555379 LB |
1310 | character in the payload you should start explicitly using |
1311 | --expand-environment=yes/no according to the use case. | |
d7b3c52c | 1312 | |
221332ee | 1313 | * The systemd-system-update-generator has been updated to also look for |
d7b3c52c LP |
1314 | the special flag file /etc/system-update in addition to the existing |
1315 | support for /system-update to decide whether to enter system update | |
1316 | mode. | |
1317 | ||
08423f6d | 1318 | * The /dev/hugepages/ file system is now mounted with nosuid + nodev |
d7b3c52c LP |
1319 | mount options by default. |
1320 | ||
1321 | * systemd-fstab-generator now understands two new kernel command line | |
221332ee | 1322 | options systemd.mount-extra= and systemd.swap-extra=, which configure |
aca7c096 LB |
1323 | additional mounts or swaps in a format similar to /etc/fstab. 'fsck' |
1324 | will be ran on these block devices, like it already happens for | |
1325 | 'root='. It also now supports the new fstab.extra and | |
1326 | fstab.extra.initrd credentials that may contain additional /etc/fstab | |
1327 | lines to apply at boot. | |
eade959b LP |
1328 | |
1329 | * systemd-getty-generator now understands two new credentials | |
1330 | getty.ttys.container and getty.ttys.serial. These credentials may | |
1331 | contain a list of TTY devices – one per line – to instantiate | |
1332 | container-getty@.service and serial-getty@.service on. | |
d7b3c52c | 1333 | |
3e6b25eb LB |
1334 | * The getty/serial-getty/container-getty units now import the 'agetty.*' |
1335 | and 'login.*' credentials, which are consumed by the 'login' and | |
1336 | 'agetty' programs starting from util-linux v2.40. | |
1337 | ||
49bf8bd5 | 1338 | * systemd-sysupdate's sysupdate.d/ drop-ins gained a new setting |
d7b3c52c LP |
1339 | PathRelativeTo=, which can be set to "esp", "xbootldr", "boot", in |
1340 | which case the Path= setting is taken relative to the ESP or XBOOTLDR | |
1341 | partitions, rather than the system's root directory /. The relevant | |
1342 | directories are automatically discovered. | |
1343 | ||
1344 | * The systemd-ac-power tool gained a new switch --low, which reports | |
1345 | whether the battery charge is considered "low", similar to how the | |
1346 | s2h suspend logic checks this state to decide whether to enter system | |
1347 | suspend or hibernation. | |
1348 | ||
221332ee ZJS |
1349 | * The /etc/os-release file can now have two new optional fields |
1350 | VENDOR_NAME= and VENDOR_URL= to carry information about the vendor of | |
1351 | the OS. | |
d7b3c52c | 1352 | |
221332ee ZJS |
1353 | * When the system hibernates, information about the device and offset |
1354 | used is now written to a non-volatile EFI variable. On next boot the | |
1355 | system will attempt to resume from the location indicated in this EFI | |
1356 | variable. This should make hibernation a lot more robust, while | |
d7b3c52c | 1357 | requiring no manual configuration of the resume location. |
b0f02185 | 1358 | |
08423f6d LP |
1359 | * The $XDG_STATE_HOME environment variable (added in more recent |
1360 | versions of the XDG basedir specification) is now honoured to | |
1361 | implement the StateDirectory= setting in user services. | |
1362 | ||
1363 | * A new component "systemd-battery-check" has been added. It may run | |
1364 | during early boot (usually in the initrd), and checks the battery | |
1365 | charge level of the system. In case the charge level is very low the | |
1366 | user is notified (graphically via Plymouth – if available – as well | |
1367 | as in text form on the console), and the system is turned off after a | |
7cfef4bb YW |
1368 | 10s delay. The feature can be disabled by passing |
1369 | systemd.battery-check=0 through the kernel command line. | |
08423f6d | 1370 | |
305bea82 | 1371 | * The 'passwdqc' library is now supported as an alternative to the |
ffe7ddb9 | 1372 | 'pwquality' library and can be selected at build time. |
305bea82 | 1373 | |
eade959b LP |
1374 | Contributions from: 김인수, 07416, Addison Snelling, Adrian Vovk, |
1375 | Aidan Dang, Alexander Krabler, Alfred Klomp, Anatoli Babenia, | |
6f19cce9 LB |
1376 | Andrei Stepanov, Andrew Baxter, Antonio Alvarez Feijoo, |
1377 | Arian van Putten, Arthur Shau, A S Alam, | |
1378 | Asier Sarasua Garmendia, Balló György, Bastien Nocera, | |
eade959b LP |
1379 | Benjamin Herrenschmidt, Benjamin Raison, Bill Peterson, |
1380 | Brad Fitzpatrick, Brett Holman, bri, Chen Qi, Chitoku, | |
205c1da0 LB |
1381 | Christian Hesse, Christoph Anton Mitterer, Christopher Gurnee, |
1382 | Colin Walters, Cornelius Hoffmann, Cristian Rodríguez, cunshunxia, | |
1383 | cvlc12, Cyril Roelandt, Daan De Meyer, Daniele Medri, | |
0b5e5e4c | 1384 | Daniel P. Berrangé, Daniel Rusek, Dan Streetman, David Edmundson, |
6f19cce9 | 1385 | David Schroeder, David Tardon, dependabot[bot], |
eade959b LP |
1386 | Dimitri John Ledkov, Dmitrii Fomchenkov, Dmitry V. Levin, dmkUK, |
1387 | Dominique Martinet, don bright, drosdeck, Edson Juliano Drosdeck, | |
6f19cce9 | 1388 | Egor Ignatov, EinBaum, Emanuele Giuseppe Esposito, Eric Curtin, |
063e0279 LB |
1389 | Erik Sjölund, Evgeny Vereshchagin, Florian Klink, Franck Bui, |
1390 | François Rigault, Fran Diéguez, Franklin Yu, Frantisek Sumsal, | |
1391 | Fuminobu TAKEYAMA, Gaël PORTAY, Gerd Hoffmann, Gertalitec, | |
1392 | Gibeom Gwon, Gustavo Noronha Silva, Hannu Lounento, | |
1393 | Hans de Goede, Haochen Tong, HATAYAMA Daisuke, Henrik Holst, | |
1394 | Hoe Hao Cheng, Igor Tsiglyar, Ivan Vecera, James Hilliard, | |
1395 | Jan Engelhardt, Jan Janssen, Jan Luebbe, Jan Macku, Janne Sirén, | |
1396 | jcg, Jeidnx, Joan Bruguera, Joerg Behrmann, jonathanmetzman, | |
1397 | Jordan Rome, Josef Miegl, Joshua Goins, Joyce, Joyce Brum, | |
1398 | Juno Computers, Kai Lueke, Kevin P. Fleming, Kiran Vemula, Klaus, | |
1399 | Klaus Zipfel, Lawrence Thorpe, Lennart Poettering, licunlong, | |
1400 | Lily Foster, Luca Boccassi, Ludwig Nussel, Luna Jernberg, | |
1401 | maanyagoenka, Maanya Goenka, Maksim Kliazovich, Malte Poll, | |
1402 | Marko Korhonen, Masatake YAMATO, Mateusz Poliwczak, Matt Johnston, | |
1403 | Miao Wang, Micah Abbott, Michael A Cassaniti, Michal Koutný, | |
0b5e5e4c LB |
1404 | Michal Sekletár, Mike Yuan, mooo, Morten Linderud, msizanoen, |
1405 | Nick Rosbrook, nikstur, Olivier Gayot, Omojola Joshua, | |
1406 | Paolo Velati, Paul Barker, Pavel Borecki, Petr Menšík, | |
6f19cce9 | 1407 | Philipp Kern, Philip Withnall, Piotr Drąg, Quintin Hill, |
eade959b | 1408 | Rene Hollander, Richard Phibel, Robert Meijers, Robert Scheck, |
205c1da0 LB |
1409 | Roger Gammans, Romain Geissler, Ronan Pigott, Russell Harmon, |
1410 | saikat0511, Samanta Navarro, Sam James, Sam Morris, | |
1411 | Simon Braunschmidt, Sjoerd Simons, Sorah Fukumori, | |
1412 | Stanislaw Gruszka, Stefan Roesch, Steven Luo, Steve Ramage, | |
1413 | Susant Sahani, taniishkaaa, Tanishka, Temuri Doghonadze, | |
1414 | Thierry Martin, Thomas Blume, Thomas Genty, Thomas Weißschuh, | |
1415 | Thorsten Kukuk, Times-Z, Tobias Powalowski, tofylion, | |
1416 | Topi Miettinen, Uwe Kleine-König, Velislav Ivanov, | |
1417 | Vitaly Kuznetsov, Vít Zikmund, Weblate, Will Fancher, | |
1418 | William Roberts, Winterhuman, Wolfgang Müller, Xeonacid, | |
1419 | Xiaotian Wu, Xi Ruoyao, Yuri Chornoivan, Yu Watanabe, Yuxiang Zhu, | |
6f19cce9 LB |
1420 | Zbigniew Jędrzejewski-Szmek, zhmylove, ZjYwMj, |
1421 | Дамјан Георгиевски, наб | |
eade959b | 1422 | |
994c7978 | 1423 | — Edinburgh, 2023-07-28 |
eade959b | 1424 | |
477fdc5a | 1425 | CHANGES WITH 253: |
1d679b20 | 1426 | |
70879f6c LB |
1427 | Announcements of Future Feature Removals and Incompatible Changes: |
1428 | ||
1429 | * We intend to remove cgroup v1 support from systemd release after the | |
1430 | end of 2023. If you run services that make explicit use of cgroup v1 | |
1431 | features (i.e. the "legacy hierarchy" with separate hierarchies for | |
1432 | each controller), please implement compatibility with cgroup v2 (i.e. | |
1433 | the "unified hierarchy") sooner rather than later. Most of Linux | |
1434 | userspace has been ported over already. | |
1435 | ||
1436 | * We intend to remove support for split-usr (/usr mounted separately | |
1437 | during boot) and unmerged-usr (parallel directories /bin and | |
1438 | /usr/bin, /lib and /usr/lib, etc). This will happen in the second | |
1439 | half of 2023, in the first release that falls into that time window. | |
1440 | For more details, see: | |
1441 | https://lists.freedesktop.org/archives/systemd-devel/2022-September/048352.html | |
1442 | ||
318c2578 LB |
1443 | * We intend to change behaviour w.r.t. units of the per-user service |
1444 | manager and sandboxing options, so that they work without having to | |
1445 | manually enable PrivateUsers= as well, which is not required for | |
1446 | system units. To make this work, we will implicitly enable user | |
1447 | namespaces (PrivateUsers=yes) when a sandboxing option is enabled in a | |
1448 | user unit. The drawback is that system users will no longer be visible | |
1449 | (and appear as 'nobody') to the user unit when a sandboxing option is | |
1450 | enabled. By definition a sandboxed user unit should run with reduced | |
1451 | privileges, so impact should be small. This will remove a great source | |
1452 | of confusion that has been reported by users over the years, due to | |
1453 | how these options require an extra setting to be manually enabled when | |
1454 | used in the per-user service manager, as opposed as to the system | |
1455 | service manager. We plan to enable this change in the next release | |
1456 | later this year. For more details, see: | |
1457 | https://lists.freedesktop.org/archives/systemd-devel/2022-December/048682.html | |
1458 | ||
1ee3720e | 1459 | Deprecations and incompatible changes: |
3b288a2d | 1460 | |
1ee3720e LP |
1461 | * systemctl will now warn when invoked without /proc/ mounted |
1462 | (e.g. when invoked after chroot() into an directory tree without the | |
1463 | API mount points like /proc/ being set up.) Operation in such an | |
1464 | environment is not fully supported. | |
3b288a2d | 1465 | |
621f7615 ZJS |
1466 | * The return value of 'systemctl is-active|is-enabled|is-failed' for |
1467 | unknown units is changed: previously 1 or 3 were returned, but now 4 | |
1468 | (EXIT_PROGRAM_OR_SERVICES_STATUS_UNKNOWN) is used as documented. | |
1469 | ||
3b288a2d ZJS |
1470 | * 'udevadm hwdb' subcommand is deprecated and will emit a warning. |
1471 | systemd-hwdb (added in 2014) should be used instead. | |
1472 | ||
1ee3720e | 1473 | * 'bootctl --json' now outputs a single JSON array, instead of a stream |
3b288a2d ZJS |
1474 | of newline-separated JSON objects. |
1475 | ||
1ee3720e LP |
1476 | * Udev rules in 60-evdev.rules have been changed to load hwdb |
1477 | properties for all modalias patterns. Previously only the first | |
1478 | matching pattern was used. This could change what properties are | |
1479 | assigned if the user has more and less specific patterns that could | |
1480 | match the same device, but it is expected that the change will have | |
1481 | no effect for most users. | |
3b288a2d | 1482 | |
621f7615 ZJS |
1483 | * systemd-networkd-wait-online exits successfully when all interfaces |
1484 | are ready or unmanaged. Previously, if neither '--any' nor | |
1485 | '--interface=' options were used, at least one interface had to be in | |
ecf4be29 ZJS |
1486 | configured state. This change allows the case where systemd-networkd |
1487 | is enabled, but no interfaces are configured, to be handled | |
621f7615 ZJS |
1488 | gracefully. It may occur in particular when a different network |
1489 | manager is also enabled and used. | |
1490 | ||
1491 | * Some compatibility helpers were dropped: EmergencyAction= in the user | |
7eff3e2c | 1492 | manager, as well as measuring kernel command line into PCR 8 in |
1ee3720e LP |
1493 | systemd-stub, along with the -Defi-tpm-pcr-compat compile-time |
1494 | option. | |
621f7615 | 1495 | |
1ee3720e LP |
1496 | * The '-Dupdate-helper-user-timeout=' build-time option has been |
1497 | renamed to '-Dupdate-helper-user-timeout-sec=', and now takes an | |
1498 | integer as parameter instead of a string. | |
b67ea78f | 1499 | |
f9fdbd54 LP |
1500 | * The DDI image dissection logic (which backs RootImage= in service |
1501 | unit files, the --image= switch in various tools such as | |
1502 | systemd-nspawn, as well as systemd-dissect) will now only mount file | |
1503 | systems of types btrfs, ext4, xfs, erofs, squashfs, vfat. This list | |
bbcce4f8 | 1504 | can be overridden via the $SYSTEMD_DISSECT_FILE_SYSTEMS environment |
f9fdbd54 LP |
1505 | variable. These file systems are fairly well supported and maintained |
1506 | in current kernels, while others are usually more niche, exotic or | |
1507 | legacy and thus typically do not receive the same level of security | |
1508 | support and fixes. | |
1509 | ||
729045ff YW |
1510 | * The default per-link multicast DNS mode is changed to "yes" |
1511 | (that was previously "no"). As the default global multicast DNS mode | |
1512 | has been "yes" (but can be changed by the build option), now the | |
1513 | multicast DNS is enabled on all links by default. You can disable the | |
1514 | multicast DNS on all links by setting MulticastDNS= in resolved.conf, | |
1515 | or on an interface by calling "resolvectl mdns INTERFACE no". | |
1516 | ||
3b288a2d ZJS |
1517 | New components: |
1518 | ||
621f7615 ZJS |
1519 | * A tool 'ukify' tool to build, measure, and sign Unified Kernel Images |
1520 | (UKIs) has been added. This replaces functionality provided by | |
1ee3720e LP |
1521 | 'dracut --uefi' and extends it with automatic calculation of PE file |
1522 | offsets, insertion of signed PCR policies generated by | |
1523 | systemd-measure, support for initrd concatenation, signing of the | |
1524 | embedded Linux image and the combined image with sbsign, and | |
1525 | heuristics to autodetect the kernel uname and verify the splash | |
1526 | image. | |
621f7615 ZJS |
1527 | |
1528 | Changes in systemd and units: | |
3b288a2d | 1529 | |
1ee3720e | 1530 | * A new service type Type=notify-reload is defined. When such a unit is |
f9fdbd54 LP |
1531 | reloaded a UNIX process signal (typically SIGHUP) is sent to the main |
1532 | service process. The manager will then wait until it receives a | |
1533 | "RELOADING=1" followed by a "READY=1" notification from the unit as | |
1534 | response (via sd_notify()). Otherwise, this type is the same as | |
1535 | Type=notify. A new setting ReloadSignal= may be used to change the | |
1536 | signal to send from the default of SIGHUP. | |
621f7615 ZJS |
1537 | |
1538 | user@.service, systemd-networkd.service, systemd-udevd.service, and | |
1ee3720e LP |
1539 | systemd-logind have been updated to this type. |
1540 | ||
1541 | * Initrd environments which are not on a pure memory file system (e.g. | |
1542 | overlayfs combination as opposed to tmpfs) are now supported. With | |
1543 | this change, during the initrd → host transition ("switch root") | |
ecf4be29 | 1544 | systemd will erase all files of the initrd only when the initrd is |
1ee3720e LP |
1545 | backed by a memory file system such as tmpfs. |
1546 | ||
1547 | * New per-unit MemoryZSwapMax= option has been added to configure | |
1548 | memory.zswap.max cgroup properties (the maximum amount of zswap | |
1549 | used). | |
1550 | ||
1551 | * A new LogFilterPatterns= option has been added for units. It may be | |
1552 | used to specify accept/deny regular expressions for log messages | |
1553 | generated by the unit, that shall be enforced by systemd-journald. | |
621f7615 | 1554 | Rejected messages are neither stored in the journal nor forwarded. |
1ee3720e | 1555 | This option may be used to suppress noisy or uninteresting messages |
621f7615 ZJS |
1556 | from units. |
1557 | ||
1558 | * The manager has a new | |
1ee3720e LP |
1559 | org.freedesktop.systemd1.Manager.GetUnitByPIDFD() D-Bus method to |
1560 | query process ownership via a PIDFD, which is more resilient against | |
1561 | PID recycling issues. | |
621f7615 | 1562 | |
3b288a2d | 1563 | * Scope units now support OOMPolicy=. Login session scopes default to |
621f7615 | 1564 | OOMPolicy=continue, allowing login scopes to survive the OOM killer |
3b288a2d ZJS |
1565 | terminating some processes in the scope. |
1566 | ||
1567 | * systemd-fstab-generator now supports x-systemd.makefs option for | |
1ee3720e | 1568 | /sysroot/ (in the initrd). |
3b288a2d | 1569 | |
621f7615 ZJS |
1570 | * The maximum rate at which daemon reloads are executed can now be |
1571 | limited with the new ReloadLimitIntervalSec=/ReloadLimitBurst= | |
1572 | options. (Or the equivalent on the kernel command line: | |
ecf4be29 | 1573 | systemd.reload_limit_interval_sec=/systemd.reload_limit_burst=). In |
1ee3720e LP |
1574 | addition, systemd now logs the originating unit and PID when a reload |
1575 | request is received over D-Bus. | |
1576 | ||
1577 | * When enabling a swap device systemd will now reinitialize the device | |
1578 | when the page size of the swap space does not match the page size of | |
f5ebd2ef LB |
1579 | the running kernel. Note that this requires the 'swapon' utility to |
1580 | provide the '--fixpgsz' option, as implemented by util-linux, and it | |
1581 | is not supported by busybox at the time of writing. | |
1ee3720e LP |
1582 | |
1583 | * systemd now executes generator programs in a mount namespace | |
1584 | "sandbox" with most of the file system read-only and write access | |
1585 | restricted to the output directories, and with a temporary /tmp/ | |
1586 | mount provided. This provides a safeguard against programming errors | |
1587 | in the generators, but also fixes here-docs in shells, which | |
1588 | previously didn't work in early boot when /tmp/ wasn't available | |
1589 | yet. (This feature has no security implications, because the code is | |
1590 | still privileged and can trivially exit the sandbox.) | |
1591 | ||
30fd9a2d | 1592 | * The system manager will now parse a new "vmm.notify_socket" |
1ee3720e | 1593 | system credential, which may be supplied to a VM via SMBIOS. If |
ecf4be29 ZJS |
1594 | found, the manager will send a "READY=1" notification on the |
1595 | specified socket after boot is complete. This allows readiness | |
1596 | notification to be sent from a VM guest to the VM host over a VSOCK | |
1597 | socket. | |
621f7615 ZJS |
1598 | |
1599 | * The sample PAM configuration file for systemd-user@.service now | |
1600 | includes a call to pam_namespace. This puts children of user@.service | |
1601 | in the expected namespace. (Many distributions replace their file | |
1602 | with something custom, so this change has limited effect.) | |
1603 | ||
1ee3720e | 1604 | * A new environment variable $SYSTEMD_DEFAULT_MOUNT_RATE_LIMIT_BURST |
30fd9a2d | 1605 | can be used to override the mount units burst late limit for |
ecf4be29 ZJS |
1606 | parsing '/proc/self/mountinfo', which was introduced in v249. |
1607 | Defaults to 5. | |
be551917 | 1608 | |
1ee3720e | 1609 | * Drop-ins for init.scope changing control group resource limits are |
be551917 LB |
1610 | now applied, while they were previously ignored. |
1611 | ||
b67ea78f LB |
1612 | * New build-time configuration options '-Ddefault-timeout-sec=' and |
1613 | '-Ddefault-user-timeout-sec=' have been added, to let distributions | |
1614 | choose the default timeout for starting/stopping/aborting system and | |
1615 | user units respectively. | |
1616 | ||
f9fdbd54 LP |
1617 | * Service units gained a new setting OpenFile= which may be used to |
1618 | open arbitrary files in the file system (or connect to arbitrary | |
1619 | AF_UNIX sockets in the file system), and pass the open file | |
1620 | descriptor to the invoked process via the usual file descriptor | |
1621 | passing protocol. This is useful to give unprivileged services access | |
1622 | to select files which have restrictive access modes that would | |
1623 | normally not allow this. It's also useful in case RootDirectory= or | |
1624 | RootImage= is used to allow access to files from the host environment | |
1625 | (which is after all not visible from the service if these two options | |
1626 | are used.) | |
1627 | ||
3b288a2d ZJS |
1628 | Changes in udev: |
1629 | ||
1630 | * The new net naming scheme "v253" has been introduced. In the new | |
1631 | scheme, ID_NET_NAME_PATH is also set for USB devices not connected via | |
621f7615 | 1632 | a PCI bus. This extends the coverage of predictable interface names |
3b288a2d ZJS |
1633 | in some embedded systems. |
1634 | ||
1635 | The "amba" bus path is now included in ID_NET_NAME_PATH, resulting in | |
1636 | a more informative path on some embedded systems. | |
1637 | ||
1ee3720e | 1638 | * Partition block devices will now also get symlinks in |
621f7615 ZJS |
1639 | /dev/disk/by-diskseq/<seq>-part<n>, which may be used to reference |
1640 | block device nodes via the kernel's "diskseq" value. Previously those | |
1641 | symlinks were only created for the main block device. | |
1642 | ||
1643 | * A new operator '-=' is supported for SYMLINK variables. This allows | |
1644 | symlinks to be unconfigured even if an earlier rule added them. | |
1d679b20 | 1645 | |
621f7615 ZJS |
1646 | * 'udevadm --trigger --settle' now also works for network devices |
1647 | that are being renamed. | |
1d679b20 | 1648 | |
621f7615 | 1649 | Changes in sd-boot, bootctl, and the Boot Loader Specification: |
1d679b20 | 1650 | |
621f7615 ZJS |
1651 | * systemd-boot now passes its random seed directly to the kernel's RNG |
1652 | via the LINUX_EFI_RANDOM_SEED_TABLE_GUID configuration table, which | |
1653 | means the RNG gets seeded very early in boot before userspace has | |
1654 | started. | |
1d679b20 | 1655 | |
1ee3720e LP |
1656 | * systemd-boot will pass a disk-backed random seed – even when secure |
1657 | boot is enabled – if it can additionally get a random seed from EFI | |
1658 | itself (via EFI's RNG protocol), or a prior seed in | |
1659 | LINUX_EFI_RANDOM_SEED_TABLE_GUID from a preceding bootloader. | |
1d679b20 | 1660 | |
621f7615 | 1661 | * systemd-boot-system-token.service was renamed to |
1ee3720e LP |
1662 | systemd-boot-random-seed.service and extended to always save a random |
1663 | seed to ESP on every boot when a compatible boot loader is used. This | |
1664 | allows a refreshed random seed to be used in the boot loader. | |
da403fd3 | 1665 | |
621f7615 ZJS |
1666 | * systemd-boot handles various seed inputs using a domain- and |
1667 | field-separated hashing scheme. | |
da403fd3 | 1668 | |
621f7615 ZJS |
1669 | * systemd-boot's 'random-seed-mode' option has been removed. A system |
1670 | token is now always required to be present for random seeds to be | |
1671 | used. | |
da403fd3 | 1672 | |
1ee3720e LP |
1673 | * systemd-boot now supports being loaded from other locations than the |
1674 | ESP, for example for direct kernel boot under QEMU or when embedded | |
1675 | into the firmware. | |
3b288a2d | 1676 | |
1ee3720e LP |
1677 | * systemd-boot now parses SMBIOS information to detect |
1678 | virtualization. This information is used to skip some warnings which | |
1679 | are not useful in a VM and to conditionalize other aspects of | |
1680 | behaviour. | |
621f7615 | 1681 | |
b67ea78f LB |
1682 | * systemd-boot now supports a new 'if-safe' mode that will perform UEFI |
1683 | Secure Boot automated certificate enrollment from the ESP only if it | |
1ee3720e LP |
1684 | is considered 'safe' to do so. At the moment 'safe' means running in |
1685 | a virtual machine. | |
b67ea78f | 1686 | |
621f7615 | 1687 | * systemd-stub now processes random seeds in the same way as |
1ee3720e LP |
1688 | systemd-boot already does, in case a unified kernel image is being |
1689 | used from a different bootloader than systemd-boot, or without any | |
1690 | boot load at all. | |
621f7615 ZJS |
1691 | |
1692 | * bootctl will now generate a system token on all EFI systems, even | |
1693 | virtualized ones, and is activated in the case that the system token | |
1694 | is missing from either sd-boot and sd-stub booted systems. | |
1695 | ||
1696 | * bootctl now implements two new verbs: 'kernel-identify' prints the | |
1ee3720e LP |
1697 | type of a kernel image file, and 'kernel-inspect' provides |
1698 | information about the embedded command line and kernel version of | |
1699 | UKIs. | |
621f7615 ZJS |
1700 | |
1701 | * bootctl now honours $KERNEL_INSTALL_CONF_ROOT with the same meaning | |
1702 | as for kernel-install. | |
1703 | ||
f9fdbd54 LP |
1704 | * The JSON output of "bootctl list" will now contain two more fields: |
1705 | isDefault and isSelected are boolean fields set to true on the | |
1706 | default and currently booted boot menu entries. | |
1707 | ||
1708 | * bootctl gained a new verb "unlink" for removing a boot loader entry | |
1709 | type #1 file from disk in a safe and robust way. | |
1710 | ||
1711 | * bootctl also gained a new verb "cleanup" that automatically removes | |
1712 | all files from the ESP's and XBOOTLDR's "entry-token" directory, that | |
1713 | is not referenced anymore by any installed Type #1 boot loader | |
bbcce4f8 | 1714 | specification entry. This is particularly useful in environments where |
f9fdbd54 LP |
1715 | a large number of entries reference the same or partly the same |
1716 | resources (for example, for snapshot-based setups). | |
1717 | ||
3b288a2d ZJS |
1718 | Changes in kernel-install: |
1719 | ||
1ee3720e LP |
1720 | * A new "installation layout" can be configured as layout=uki. With |
1721 | this setting, a Boot Loader Specification Type#1 entry will not be | |
33db1b90 | 1722 | created. Instead, a new kernel-install plugin 90-uki-copy.install |
1ee3720e LP |
1723 | will copy any .efi files from the staging area into the boot |
1724 | partition. A plugin to generate the UKI .efi file must be provided | |
1725 | separately. | |
3b288a2d | 1726 | |
31853609 MY |
1727 | Changes in systemctl: |
1728 | ||
3b288a2d ZJS |
1729 | * 'systemctl reboot' has dropped support for accepting a positional |
1730 | argument as the argument to the reboot(2) syscall. Please use the | |
1ee3720e | 1731 | --reboot-argument= option instead. |
3b288a2d | 1732 | |
1ee3720e LP |
1733 | * 'systemctl disable' will now warn when called on units without |
1734 | install information. A new --no-warn option has been added that | |
1735 | silences this warning. | |
3b288a2d | 1736 | |
621f7615 | 1737 | * New option '--drop-in=' can be used to tell 'systemctl edit' the name |
1ee3720e LP |
1738 | of the drop-in to edit. (Previously, 'override.conf' was always |
1739 | used.) | |
621f7615 ZJS |
1740 | |
1741 | * 'systemctl list-dependencies' now respects --type= and --state=. | |
1742 | ||
1ee3720e | 1743 | * 'systemctl kexec' now supports XEN VMM environments. |
3b288a2d | 1744 | |
f9fdbd54 LP |
1745 | * 'systemctl edit' will now tell the invoked editor to jump into the |
1746 | first line with actual unit file data, skipping over synthesized | |
1747 | comments. | |
1748 | ||
3b288a2d ZJS |
1749 | Changes in systemd-networkd and related tools: |
1750 | ||
c9720268 | 1751 | * The [DHCPv4] section in .network file gained new SocketPriority= |
1ee3720e LP |
1752 | setting that assigns the Linux socket priority used by the DHCPv4 raw |
1753 | socket. This may be used in conjunction with the | |
1754 | EgressQOSMaps=setting in [VLAN] section of .netdev file to send the | |
1755 | desired ethernet 802.1Q frame priority for DHCPv4 initial | |
1756 | packets. This cannot be achieved with netfilter mangle tables because | |
1757 | of the raw socket bypass. | |
1758 | ||
1759 | * The [DHCPv4] and [IPv6AcceptRA] sections in .network file gained a | |
1760 | new QuickAck= boolean setting that enables the TCP quick ACK mode for | |
1761 | the routes configured by the acquired DHCPv4 lease or received router | |
1d00da37 YW |
1762 | advertisements (RAs). |
1763 | ||
3b288a2d ZJS |
1764 | * The RouteMetric= option (for DHCPv4, DHCPv6, and IPv6 advertised |
1765 | routes) now accepts three values, for high, medium, and low preference | |
1766 | of the router (which can be set with the RouterPreference=) setting. | |
1767 | ||
1ee3720e LP |
1768 | * systemd-networkd-wait-online now supports matching via alternative |
1769 | interface names. | |
3b288a2d | 1770 | |
b895aa5f | 1771 | * The [DHCPv6] section in .network file gained new SendRelease= |
1772 | setting which enables the DHCPv6 client to send release when | |
1773 | it stops. This is the analog of the [DHCPv4] SendRelease= setting. | |
1774 | It is enabled by default. | |
1775 | ||
58058629 YW |
1776 | * If the Address= setting in [Network] or [Address] sections in .network |
1777 | specified without its prefix length, then now systemd-networkd assumes | |
1778 | /32 for IPv4 or /128 for IPv6 addresses. | |
1779 | ||
621f7615 ZJS |
1780 | * networkctl shows network and link file dropins in status output. |
1781 | ||
3b288a2d ZJS |
1782 | Changes in systemd-dissect: |
1783 | ||
75438b2a | 1784 | * systemd-dissect gained a new option --list, to print the paths of |
1ee3720e | 1785 | all files and directories in a DDI. |
3b288a2d | 1786 | |
1ee3720e LP |
1787 | * systemd-dissect gained a new option --mtree, to generate a file |
1788 | manifest compatible with BSD mtree(5) of a DDI | |
3b288a2d | 1789 | |
1ee3720e LP |
1790 | * systemd-dissect gained a new option --with, to execute a command with |
1791 | the specified DDI temporarily mounted and used as working | |
1792 | directory. This is for example useful to convert a DDI to "tar" | |
1793 | simply by running it within a "systemd-dissect --with" invocation. | |
3b288a2d ZJS |
1794 | |
1795 | * systemd-dissect gained a new option --discover, to search for | |
1ee3720e LP |
1796 | Discoverable Disk Images (DDIs) in well-known directories of the |
1797 | system. This will list machine, portable service and system extension | |
1798 | disk images. | |
3b288a2d ZJS |
1799 | |
1800 | * systemd-dissect now understands 2nd stage initrd images stored as a | |
1801 | Discoverable Disk Image (DDI). | |
1802 | ||
f9fdbd54 LP |
1803 | * systemd-dissect will now display the main UUID of GPT DDIs (i.e. the |
1804 | disk UUID stored in the GPT header) among the other data it can show. | |
1805 | ||
1806 | * systemd-dissect gained a new --in-memory switch to operate on an | |
1807 | in-memory copy of the specified DDI file. This is useful to access a | |
1808 | DDI with write access without persisting any changes. It's also | |
1809 | useful for accessing a DDI without keeping the originating file | |
1810 | system busy. | |
1811 | ||
1812 | * The DDI dissection logic will now automatically detect the intended | |
1813 | sector size of disk images stored in files, based on the GPT | |
1814 | partition table arrangement. Loopback block devices for such DDIs | |
1815 | will then be configured automatically for the right sector size. This | |
1816 | is useful to make dealing with modern 4K sector size DDIs fully | |
1817 | automatic. The systemd-dissect tool will now show the detected sector | |
1818 | size among the other DDI information in its output. | |
1819 | ||
3b288a2d ZJS |
1820 | Changes in systemd-repart: |
1821 | ||
621f7615 ZJS |
1822 | * systemd-repart gained new options --include-partitions= and |
1823 | --exclude-partitions= to filter operation on partitions by type UUID. | |
3b288a2d ZJS |
1824 | This allows systemd-repart to be used to build images in which the |
1825 | type of one partition is set based on the contents of another | |
1826 | partition (for example when the boot partition shall include a verity | |
1827 | hash of the root partition). | |
1828 | ||
621f7615 ZJS |
1829 | * systemd-repart also gained a --defer-partitions= option that is |
1830 | similar to --exclude-partitions=, but the size of the partition is | |
1ee3720e LP |
1831 | still taken into account when sizing partitions, but without |
1832 | populating it. | |
621f7615 ZJS |
1833 | |
1834 | * systemd-repart gained a new --sector-size= option to specify what | |
1835 | sector size should be used when an image is created. | |
1836 | ||
1ee3720e LP |
1837 | * systemd-repart now supports generating erofs file systems via |
1838 | CopyFiles= (a read-only file system similar to squashfs). | |
3b288a2d | 1839 | |
621f7615 ZJS |
1840 | * The Minimize= option was extended to accept "best" (which means the |
1841 | most minimal image possible, but may require multiple attempts) and | |
1842 | "guess" (which means a reasonably small image). | |
1843 | ||
f9fdbd54 LP |
1844 | * The systemd-growfs binary now comes with a regular unit file template |
1845 | systemd-growfs@.service which can be instantiated directly for any | |
1846 | desired file system. (Previously, the unit was generated dynamically | |
1847 | by various generators, but no regular unit file template was | |
1848 | available.) | |
1849 | ||
621f7615 ZJS |
1850 | Changes in journal tools: |
1851 | ||
1852 | * Various systemd tools will append extra fields to log messages when | |
1853 | in debug mode, or when SYSTEMD_ENABLE_LOG_CONTEXT=1 is set. Currently | |
1854 | this includes information about D-Bus messages when sd-bus is used, | |
1855 | e.g. DBUS_SENDER=, DBUS_DESTINATION=, and DBUS_PATH=, and information | |
1856 | about devices when sd-device is used, e.g. DEVNAME= and DRIVER=. | |
1857 | Details of what is logged and when are subject to change. | |
1858 | ||
1ee3720e LP |
1859 | * The systemd-journald-audit.socket can now be disabled via the usual |
1860 | "systemctl disable" mechanism to stop collection of audit | |
1861 | messages. Please note that it is not enabled statically anymore and | |
1862 | must be handled by the preset/enablement logic in package | |
1863 | installation scripts. | |
621f7615 ZJS |
1864 | |
1865 | * New options MaxUse=, KeepFree=, MaxFileSize=, and MaxFiles= can | |
1866 | be used to curtail disk use by systemd-journal-remote. This is | |
1867 | similar to the options supported by systemd-journald. | |
1868 | ||
1869 | Changes in systemd-cryptenroll, systemd-cryptsetup, and related | |
1ee3720e | 1870 | components: |
621f7615 | 1871 | |
1ee3720e LP |
1872 | * When enrolling new keys systemd-cryptenroll now supports unlocking |
1873 | via FIDO2 tokens (option --unlock-fido2-device=). Previously, a | |
1874 | password was strictly required to be specified. | |
621f7615 | 1875 | |
ad901df9 PC |
1876 | * systemd-cryptsetup now supports pre-flight requests for FIDO2 tokens |
1877 | (except for tokens with user verification, UV) to identify tokens | |
1878 | before authentication. Multiple FIDO2 tokens can now be enrolled at | |
1879 | the same time, and systemd-cryptsetup will automatically select one | |
1880 | that corresponds to one of the available LUKS key slots. | |
1881 | ||
1ee3720e LP |
1882 | * systemd-cryptsetup now supports new options tpm2-measure-bank= and |
1883 | tpm2-measure-pcr= in crypttab(5). These allow specifying the TPM2 PCR | |
f9fdbd54 LP |
1884 | bank and number into which the volume key should be measured. This is |
1885 | automatically enabled for the encrypted root volume discovered and | |
1886 | activated by systemd-gpt-auto-generator. | |
621f7615 | 1887 | |
4a20ad15 | 1888 | * systemd-gpt-auto-generator mounts the ESP and XBOOTLDR partitions with |
621f7615 ZJS |
1889 | "noexec,nosuid,nodev". |
1890 | ||
f9fdbd54 LP |
1891 | * systemd-gpt-auto-generator will now honour the rootfstype= and |
1892 | rootflags= kernel command line switches for root file systems it | |
1893 | discovers, to match behaviour in case an explicit root fs is | |
1894 | specified via root=. | |
1895 | ||
621f7615 | 1896 | * systemd-pcrphase gained new options --machine-id and --file-system= |
7eff3e2c ZJS |
1897 | to measure the machine-id and mount point information into PCR 15. |
1898 | New service unit files systemd-pcrmachine.service and | |
1ee3720e LP |
1899 | systemd-pcrfs@.service have been added that invoke the tool with |
1900 | these switches during early boot. | |
621f7615 | 1901 | |
f9fdbd54 LP |
1902 | * systemd-pcrphase gained a --graceful switch will make it exit cleanly |
1903 | with a success exit code even if no TPM device is detected. | |
1904 | ||
621f7615 ZJS |
1905 | * systemd-cryptenroll now stores the user-supplied PIN with a salt, |
1906 | making it harder to brute-force. | |
1907 | ||
1908 | Changes in other tools: | |
3b288a2d ZJS |
1909 | |
1910 | * systemd-homed gained support for luksPbkdfForceIterations (the | |
1911 | intended number of iterations for the PBKDF operation on LUKS). | |
1912 | ||
621f7615 ZJS |
1913 | * Environment variables $SYSTEMD_HOME_MKFS_OPTIONS_BTRFS, |
1914 | $SYSTEMD_HOME_MKFS_OPTIONS_EXT4, and $SYSTEMD_HOME_MKFS_OPTIONS_XFS | |
1ee3720e | 1915 | may now be used to specify additional arguments for mkfs when |
621f7615 | 1916 | systemd-homed formats a file system. |
3b288a2d | 1917 | |
621f7615 | 1918 | * systemd-hostnamed now exports the contents of |
3b288a2d ZJS |
1919 | /sys/class/dmi/id/bios_vendor and /sys/class/dmi/id/bios_date via two |
1920 | new D-Bus properties: FirmwareVendor and FirmwareDate. This allows | |
1921 | unprivileged code to access those values. | |
1922 | ||
621f7615 | 1923 | systemd-hostnamed also exports the SUPPORT_END= field from |
1ee3720e | 1924 | os-release(5) as OperatingSystemSupportEnd. hostnamectl make uses of |
621f7615 ZJS |
1925 | this to show the status of the installed system. |
1926 | ||
1927 | * systemd-measure gained an --append= option to sign multiple phase | |
1928 | paths with different signing keys. This allows secrets to be | |
1929 | accessible only in certain parts of the boot sequence. Note that | |
1930 | 'ukify' provides similar functionality in a more accessible form. | |
1931 | ||
1932 | * systemd-timesyncd will now write a structured log message with | |
1933 | MESSAGE_ID set to SD_MESSAGE_TIME_BUMP when it bumps the clock based | |
1ee3720e | 1934 | on a on-disk timestamp, similarly to what it did when reaching |
621f7615 ZJS |
1935 | synchronization via NTP. |
1936 | ||
1ee3720e LP |
1937 | * systemd-timesyncd will now update the on-disk timestamp file on each |
1938 | boot at least once, making it more likely that the system time | |
1939 | increases in subsequent boots. | |
621f7615 | 1940 | |
1ee3720e | 1941 | * systemd-vconsole-setup gained support for system/service credentials: |
621f7615 ZJS |
1942 | vconsole.keymap/vconsole.keymap_toggle and |
1943 | vconsole.font/vconsole.font_map/vconsole.font_unimap are analogous | |
1944 | the similarly-named options in vconsole.conf. | |
1945 | ||
1946 | * systemd-localed will now save the XKB keyboard configuration to | |
1947 | /etc/vconsole.conf, and also read it from there with a higher | |
1948 | preference than the /etc/X11/xorg.conf.d/00-keyboard.conf config | |
1949 | file. Previously, this information was stored in the former file in | |
1950 | converted form, and only in latter file in the original form. Tools | |
1951 | which want to access keyboard configuration can now do so from a | |
1952 | standard location. | |
1953 | ||
1954 | * systemd-resolved gained support for configuring the nameservers and | |
1955 | search domains via kernel command line (nameserver=, domain=) and | |
1956 | credentials (network.dns, network.search_domains). | |
1957 | ||
f9fdbd54 LP |
1958 | * systemd-resolved will now synthesize host names for the DNS stub |
1959 | addresses it supports. Specifically when "_localdnsstub" is resolved, | |
1960 | 127.0.0.53 is returned, and if "_localdnsproxy" is resolved | |
1961 | 127.0.0.54 is returned. | |
1962 | ||
621f7615 ZJS |
1963 | * systemd-notify will now send a "RELOADING=1" notification when called |
1964 | with --reloading, and "STOPPING=1" when called with --stopping. This | |
1965 | can be used to implement notifications from units where it's easier | |
1966 | to call a program than to use the sd-daemon library. | |
1967 | ||
f9fdbd54 LP |
1968 | * systemd-analyze's 'plot' command can now output its information in |
1969 | JSON, controlled via the --json= switch. Also, new --table, and | |
1970 | --no-legend options have been added. | |
621f7615 ZJS |
1971 | |
1972 | * 'machinectl enable' will now automatically enable machines.target | |
1973 | unit in addition to adding the machine unit to the target. | |
1974 | ||
1975 | Similarly, 'machinectl start|stop' gained a --now option to enable or | |
1976 | disable the machine unit when starting or stopping it. | |
1977 | ||
1ee3720e | 1978 | * systemd-sysusers will now create /etc/ if it is missing. |
b67ea78f LB |
1979 | |
1980 | * systemd-sleep 'HibernateDelaySec=' setting is changed back to | |
1981 | pre-v252's behaviour, and a new 'SuspendEstimationSec=' setting is | |
1982 | added to provide the new initial value for the new automated battery | |
1983 | estimation functionality. If 'HibernateDelaySec=' is set to any value, | |
1984 | the automated estimate (and thus the automated hibernation on low | |
1985 | battery to avoid data loss) functionality will be disabled. | |
1986 | ||
1987 | * Default tmpfiles.d/ configuration will now automatically create | |
1988 | credentials storage directory '/etc/credstore/' with the appropriate, | |
1989 | secure permissions. If '/run/credstore/' exists, its permissions will | |
1990 | be fixed too in case they are not correct. | |
1991 | ||
3b288a2d ZJS |
1992 | Changes in libsystemd and shared code: |
1993 | ||
1994 | * sd-bus gained new convenience functions sd_bus_emit_signal_to(), | |
1995 | sd_bus_emit_signal_tov(), and sd_bus_message_new_signal_to(). | |
1996 | ||
621f7615 | 1997 | * sd-id128 functions now return -EUCLEAN (instead of -EIO) when the |
da890466 | 1998 | 128-bit ID in files such as /etc/machine-id has an invalid |
1ee3720e LP |
1999 | format. They also accept NULL as output parameter in more places, |
2000 | which is useful when the caller only wants to validate the inputs and | |
2001 | does not need the output value. | |
621f7615 ZJS |
2002 | |
2003 | * sd-login gained new functions sd_pidfd_get_session(), | |
2004 | sd_pidfd_get_owner_uid(), sd_pidfd_get_unit(), | |
2005 | sd_pidfd_get_user_unit(), sd_pidfd_get_slice(), | |
2006 | sd_pidfd_get_user_slice(), sd_pidfd_get_machine_name(), and | |
2007 | sd_pidfd_get_cgroup(), that are analogous to sd_pid_get_*(), | |
2008 | but accept a PIDFD instead of a PID. | |
2009 | ||
2010 | * sd-path (and systemd-path) now export four new paths: | |
2011 | SD_PATH_SYSTEMD_SYSTEM_ENVIRONMENT_GENERATOR, | |
2012 | SD_PATH_SYSTEMD_USER_ENVIRONMENT_GENERATOR, | |
2013 | SD_PATH_SYSTEMD_SEARCH_SYSTEM_ENVIRONMENT_GENERATOR, and | |
2014 | SD_PATH_SYSTEMD_SEARCH_USER_ENVIRONMENT_GENERATOR, | |
2015 | ||
f9fdbd54 LP |
2016 | * sd_notify() now supports AF_VSOCK as transport for notification |
2017 | messages (in addition to the existing AF_UNIX support). This is | |
2018 | enabled if $NOTIFY_SOCKET is set in a "vsock:CID:port" format. | |
be551917 | 2019 | |
1ee3720e | 2020 | * Detection of chroot() environments now works if /proc/ is not |
33db1b90 | 2021 | mounted. This affects systemd-detect-virt --chroot, but also means |
1ee3720e LP |
2022 | that systemd tools will silently skip various operations in such an |
2023 | environment. | |
3b288a2d | 2024 | |
8ad6e519 | 2025 | * "Lockheed Martin Hardened Security for Intel Processors" (HS SRE) |
3b288a2d ZJS |
2026 | virtualization is now detected. |
2027 | ||
2028 | Changes in the build system: | |
2029 | ||
f9fdbd54 LP |
2030 | * Standalone variants of systemd-repart and systemd-shutdown may now be |
2031 | built (if -Dstandalone=true). | |
3b288a2d | 2032 | |
1ee3720e LP |
2033 | * systemd-ac-power has been moved from /usr/lib/ to /usr/bin/, to, for |
2034 | example, allow scripts to conditionalize execution on AC power | |
2035 | supply. | |
3b288a2d | 2036 | |
621f7615 ZJS |
2037 | * The libp11kit library is now loaded through dlopen(3). |
2038 | ||
3b288a2d ZJS |
2039 | Changes in the documentation: |
2040 | ||
2041 | * Specifications that are not closely tied to systemd have moved to | |
bdc11d26 | 2042 | https://uapi-group.org/specifications/: the Boot Loader Specification |
3b288a2d ZJS |
2043 | and the Discoverable Partitions Specification. |
2044 | ||
8ca20dfa LB |
2045 | Contributions from: 김인수, 13r0ck, Aidan Dang, Alberto Planas, |
2046 | Alvin Šipraga, Andika Triwidada, AndyChi, angus-p, Anita Zhang, | |
aff998ae LB |
2047 | Antonio Alvarez Feijoo, Arsen Arsenović, asavah, Benjamin Fogle, |
2048 | Benjamin Tissoires, berenddeschouwer, BerndAdameit, | |
2049 | Bernd Steinhauser, blutch112, cake03, Callum Farmer, Carlo Teubner, | |
2050 | Charles Hardin, chris, Christian Brauner, Christian Göttsche, | |
2051 | Cristian Rodríguez, Daan De Meyer, Dan Streetman, DaPigGuy, | |
4142ea8d LB |
2052 | Darrell Kavanagh, David Tardon, dependabot[bot], Dirk Su, |
2053 | Dmitry V. Levin, drosdeck, Edson Juliano Drosdeck, edupont, | |
2054 | Eric DeVolder, Erik Moqvist, Evgeny Vereshchagin, Fabian Gurtner, | |
2055 | Felix Riemann, Franck Bui, Frantisek Sumsal, Geert Lorang, | |
8ca20dfa | 2056 | Gerd Hoffmann, Gio, Hannoskaj, Hans de Goede, Hugo Carvalho, |
aff998ae LB |
2057 | igo95862, Ilya Leoshkevich, Ivan Shapovalov, Jacek Migacz, |
2058 | Jade Lovelace, Jan Engelhardt, Jan Janssen, Jan Macku, January, | |
1233a7cf LB |
2059 | Jason A. Donenfeld, jcg, Jean-Tiare Le Bigot, Jelle van der Waa, |
2060 | Jeremy Linton, Jian Zhang, Jiayi Chen, Jia Zhang, Joerg Behrmann, | |
2061 | Jörg Thalheim, Joshua Goins, joshuazivkovic, Joshua Zivkovic, | |
2062 | Kai-Chuan Hsieh, Khem Raj, Koba Ko, Lennart Poettering, lichao, | |
2063 | Li kunyu, Luca Boccassi, Luca BRUNO, Ludwig Nussel, | |
2064 | Łukasz Stelmach, Lycowolf, marcel151, Marcus Schäfer, Marek Vasut, | |
2065 | Mark Laws, Michael Biebl, Michał Kotyla, Michal Koutný, | |
2066 | Michal Sekletár, Mike Gilbert, Mike Yuan, MkfsSion, ml, | |
2067 | msizanoen1, mvzlb, MVZ Ludwigsburg, Neil Moore, Nick Rosbrook, | |
2068 | noodlejetski, Pasha Vorobyev, Peter Cai, p-fpv, Phaedrus Leeds, | |
2069 | Philipp Jungkamp, Quentin Deslandes, Raul Tambre, Ray Strode, | |
8ca20dfa | 2070 | reuben olinsky, Richard E. van der Luit, Richard Phibel, |
aff998ae | 2071 | Ricky Tigg, Robin Humble, rogg, Rudi Heitbaum, Sam James, |
4142ea8d LB |
2072 | Samuel Cabrero, Samuel Thibault, Siddhesh Poyarekar, Simon Brand, |
2073 | Space Meyer, Spindle Security, Steve Ramage, Takashi Sakamoto, | |
2074 | Thomas Haller, Tonći Galić, Topi Miettinen, Torsten Hilbrich, | |
aff998ae LB |
2075 | Tuetuopay, uerdogan, Ulrich Ölmann, Valentin David, |
2076 | Vitaly Kuznetsov, Vito Caputo, Waltibaba, Will Fancher, | |
2077 | William Roberts, wouter bolsterlee, Youfu Zhang, Yu Watanabe, | |
8ca20dfa LB |
2078 | Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, |
2079 | наб | |
31853609 | 2080 | |
477fdc5a | 2081 | — Warsaw, 2023-02-15 |
903dd65b | 2082 | |
e8dc5276 | 2083 | CHANGES WITH 252 🎃: |
b98445cd | 2084 | |
02380e19 | 2085 | Announcements of Future Feature Removals: |
b98445cd | 2086 | |
02380e19 ZJS |
2087 | * We intend to remove cgroup v1 support from systemd release after the |
2088 | end of 2023. If you run services that make explicit use of cgroup v1 | |
2089 | features (i.e. the "legacy hierarchy" with separate hierarchies for | |
2090 | each controller), please implement compatibility with cgroup v2 (i.e. | |
2091 | the "unified hierarchy") sooner rather than later. Most of Linux | |
b98445cd LP |
2092 | userspace has been ported over already. |
2093 | ||
02380e19 ZJS |
2094 | * We intend to remove support for split-usr (/usr mounted separately |
2095 | during boot) and unmerged-usr (parallel directories /bin and | |
2096 | /usr/bin, /lib and /usr/lib, etc). This will happen in the second | |
2097 | half of 2023, in the first release that falls into that time window. | |
2098 | For more details, see: | |
f77c0840 LB |
2099 | https://lists.freedesktop.org/archives/systemd-devel/2022-September/048352.html |
2100 | ||
10736074 LP |
2101 | Compatibility Breaks: |
2102 | ||
02380e19 ZJS |
2103 | * ConditionKernelVersion= checks that use the '=' or '!=' operators |
2104 | will now do simple string comparisons (instead of version comparisons | |
3ff1721c | 2105 | à la stverscmp()). Version comparisons are still done for the |
02380e19 ZJS |
2106 | ordering operators '<', '>', '<=', '>='. Moreover, if no operator is |
2107 | specified, a shell-style glob match is now done. This creates a minor | |
2108 | incompatibility compared to older systemd versions when the '*', '?', | |
2109 | '[', ']' characters are used, as these will now match as shell globs | |
2110 | instead of literally. Given that kernel version strings typically do | |
2111 | not include these characters we expect little breakage through this | |
2112 | change. | |
2113 | ||
2114 | * The service manager will now read the SELinux label used for SELinux | |
2115 | access checks from the unit file at the time it loads the file. | |
2116 | Previously, the label would be read at the moment of the access | |
2117 | check, which was problematic since at that time the unit file might | |
2118 | already have been updated or removed. | |
a0769ee4 | 2119 | |
10736074 | 2120 | New Features: |
e49d111b | 2121 | |
8d3b7d2f ZJS |
2122 | * systemd-measure is a new tool for calculating and signing expected |
2123 | TPM2 PCR values for a given unified kernel image (UKI) booted via | |
2124 | sd-stub. The public key used for the signature and the signed | |
2125 | expected PCR information can be embedded inside the UKI. This | |
2126 | information can be extracted from the UKI by external tools and code | |
2127 | in the image itself and is made available to userspace in the booted | |
2128 | kernel. | |
25d615eb | 2129 | |
8d3b7d2f | 2130 | systemd-cryptsetup, systemd-cryptenroll, and systemd-creds have been |
25d615eb | 2131 | updated to make use of this information if available in the booted |
29818c4e | 2132 | kernel: when locking an encrypted volume/credential to the TPM |
8d3b7d2f ZJS |
2133 | systemd-cryptenroll/systemd-creds will use the public key to bind the |
2134 | volume/credential to any kernel that carries PCR information signed | |
2135 | by the same key pair. When unlocking such volumes/credentials | |
2136 | systemd-cryptsetup/systemd-creds will use the signature embedded in | |
2137 | the booted UKI to gain access. | |
25d615eb LP |
2138 | |
2139 | Binding TPM-based disk encryption to public keys/signatures of PCR | |
2140 | values — instead of literal PCR values — addresses the inherent | |
2141 | "brittleness" of traditional PCR-bound TPM disk encryption schemes: | |
29818c4e LP |
2142 | disks remain accessible even if the UKI is updated, without any TPM |
2143 | specific preparation during the OS update — as long as each UKI | |
2144 | carries the necessary PCR signature information. | |
2145 | ||
2146 | Net effect: if you boot a properly prepared kernel, TPM-bound disk | |
2147 | encryption now defaults to be locked to kernels which carry PCR | |
8d3b7d2f ZJS |
2148 | signatures from the same key pair. Example: if a hypothetical distro |
2149 | FooOS prepares its UKIs like this, TPM-based disk encryption is now – | |
2150 | by default – bound to only FooOS kernels, and encrypted volumes bound | |
2151 | to the TPM cannot be unlocked on kernels from other sources. (But do | |
2152 | note this behaviour requires preparation/enabling in the UKI, and of | |
2153 | course users can always enroll non-TPM ways to unlock the volume.) | |
a0769ee4 | 2154 | |
9ca1efbc | 2155 | * systemd-pcrphase is a new tool that is invoked at six places during |
7eff3e2c | 2156 | system runtime, and measures additional words into TPM2 PCR 11, to |
a0769ee4 | 2157 | mark milestones of the boot process. This allows binding access to |
02380e19 | 2158 | specific TPM2-encrypted secrets to specific phases of the boot |
29818c4e | 2159 | process. (Example: LUKS2 disk encryption key only accessible in the |
02380e19 | 2160 | initrd, but not later.) |
e49d111b | 2161 | |
02380e19 | 2162 | Changes in systemd itself, i.e. the manager and units |
e49d111b | 2163 | |
a0769ee4 LP |
2164 | * The cpu controller is delegated to user manager units by default, and |
2165 | CPUWeight= settings are applied to the top-level user slice units | |
2166 | (app.slice, background.slice, session.slice). This provides a degree | |
2167 | of resource isolation between different user services competing for | |
2168 | the CPU. | |
e49d111b ZJS |
2169 | |
2170 | * Systemd can optionally do a full preset in the "first boot" condition | |
2171 | (instead of just enable-only). This behaviour is controlled by the | |
02380e19 | 2172 | compile-time option -Dfirst-boot-full-preset. Right now it defaults |
e49d111b ZJS |
2173 | to 'false', but the plan is to switch it to 'true' for the subsequent |
2174 | release. | |
2175 | ||
9ca1efbc ZJS |
2176 | * Drop-ins are now allowed for transient units too. |
2177 | ||
e49d111b | 2178 | * Systemd will set the taint flag 'support-ended' if it detects that |
02380e19 ZJS |
2179 | the OS image is past its end-of-support date. This date is declared |
2180 | in a new /etc/os-release field SUPPORT_END= described below. | |
e49d111b | 2181 | |
a0769ee4 | 2182 | * Two new settings ConditionCredential= and AssertCredential= can be |
4db5c45d LP |
2183 | used to skip or fail units if a certain system credential is not |
2184 | provided. | |
e49d111b | 2185 | |
02380e19 | 2186 | * ConditionMemory= accepts size suffixes (K, M, G, T, …). |
e49d111b | 2187 | |
a0769ee4 LP |
2188 | * DefaultSmackProcessLabel= can be used in system.conf and user.conf to |
2189 | specify the SMACK security label to use when not specified in a unit | |
2190 | file. | |
e49d111b | 2191 | |
02380e19 ZJS |
2192 | * DefaultDeviceTimeoutSec= can be used in system.conf and user.conf to |
2193 | specify the default timeout when waiting for device units to | |
2194 | activate. | |
e49d111b | 2195 | |
a0769ee4 LP |
2196 | * C.UTF-8 is used as the default locale if nothing else has been |
2197 | configured. | |
e49d111b | 2198 | |
02380e19 ZJS |
2199 | * [Condition|Assert]Firmware= have been extended to support certain |
2200 | SMBIOS fields. For example | |
2201 | ||
2202 | ConditionFirmware=smbios-field(board_name = "Custom Board") | |
2203 | ||
2204 | conditionalizes the unit to run only when | |
2205 | /sys/class/dmi/id/board_name contains "Custom Board" (without the | |
a0769ee4 | 2206 | quotes). |
bf07a125 | 2207 | |
f77c0840 | 2208 | * ConditionFirstBoot= now correctly evaluates as true only during the |
02380e19 ZJS |
2209 | boot phase of the first boot. A unit executed later, after booting |
2210 | has completed, will no longer evaluate this condition as true. | |
f77c0840 LB |
2211 | |
2212 | * Socket units will now create sockets in the SELinuxContext= of the | |
2213 | associated service unit, if any. | |
2214 | ||
a0769ee4 | 2215 | * Boot phase transitions (start initrd → exit initrd → boot complete → |
7eff3e2c | 2216 | shutdown) will be measured into TPM2 PCR 11, so that secrets can be |
02380e19 | 2217 | bound to a specific runtime phase. E.g.: a LUKS encryption key can be |
f77c0840 LB |
2218 | unsealed only in the initrd. |
2219 | ||
a0769ee4 LP |
2220 | * Service credentials (i.e. SetCredential=/LoadCredential=/…) will now |
2221 | also be provided to ExecStartPre= processes. | |
f77c0840 | 2222 | |
02380e19 ZJS |
2223 | * Various units are now correctly ordered against |
2224 | initrd-switch-root.target where previously a conflict without | |
2225 | ordering was configured. A stop job for those units would be queued, | |
2226 | but without the ordering it could be executed only after | |
2227 | initrd-switch-root.service, leading to units not being restarted in | |
2228 | the host system as expected. | |
f77c0840 LB |
2229 | |
2230 | * In order to fully support the IPMI watchdog driver, which has not yet | |
a0769ee4 LP |
2231 | been ported to the new common watchdog device interface, |
2232 | /dev/watchdog0 will be tried first and systemd will silently fallback | |
2233 | to /dev/watchdog if it is not found. | |
f77c0840 | 2234 | |
3af9dc77 LB |
2235 | * New watchdog-related D-Bus properties are now published by systemd: |
2236 | WatchdogDevice, WatchdogLastPingTimestamp, | |
2237 | WatchdogLastPingTimestampMonotonic. | |
f77c0840 | 2238 | |
02380e19 ZJS |
2239 | * At shutdown, API virtual files systems (proc, sys, etc.) will be |
2240 | unmounted lazily. | |
043ba6a1 | 2241 | |
02380e19 ZJS |
2242 | * At shutdown, systemd will now log about processes blocking unmounting |
2243 | of file systems. | |
a0769ee4 | 2244 | |
043ba6a1 | 2245 | * A new meson build option 'clock-valid-range-usec-max' was added to |
a0769ee4 | 2246 | allow disabling system time correction if RTC returns a timestamp far |
043ba6a1 LB |
2247 | in the future. |
2248 | ||
167420a3 LB |
2249 | * Propagated restart jobs will no longer be discarded while a unit is |
2250 | activating. | |
043ba6a1 | 2251 | |
a0769ee4 LP |
2252 | * PID 1 will now import system credentials from SMBIOS Type 11 fields |
2253 | ("OEM vendor strings"), in addition to qemu_fwcfg. This provides a | |
02380e19 ZJS |
2254 | simple, fast and generic path for supplying credentials to a VM, |
2255 | without involving external tools such as cloud-init/ignition. | |
a0769ee4 LP |
2256 | |
2257 | * The CPUWeight= setting of unit files now accepts a new special value | |
2258 | "idle", which configures "idle" level scheduling for the unit. | |
2259 | ||
2260 | * Service processes that are activated due to a .timer or .path unit | |
2261 | triggering will now receive information about this via environment | |
02380e19 ZJS |
2262 | variables. Note that this is information is lossy, as activation |
2263 | might be coalesced and only one of the activating triggers will be | |
2264 | reported. This is hence more suited for debugging or tracing rather | |
2265 | than for behaviour decisions. | |
a0769ee4 | 2266 | |
9ca1efbc ZJS |
2267 | * The riscv_flush_icache(2) system call has been added to the list of |
2268 | system calls allowed by default when SystemCallFilter= is used. | |
2269 | ||
2270 | * The selinux context derived from the target executable, instead of | |
2271 | 'init_t' used for the manager itself, is now used when creating | |
2272 | listening sockets for units that specify SELinuxContextFromNet=yes. | |
2273 | ||
e49d111b ZJS |
2274 | Changes in sd-boot, bootctl, and the Boot Loader Specification: |
2275 | ||
2276 | * The Boot Loader Specification has been cleaned up and clarified. | |
2277 | Various corner cases in version string comparisons have been fixed | |
2278 | (e.g. comparisons for empty strings). Boot counting is now part of | |
2279 | the main specification. | |
2280 | ||
7eff3e2c ZJS |
2281 | * New PCRs measurements are performed during boot: PCR 11 for the |
2282 | kernel+initrd combo, PCR 13 for any sysext images. If a measurement | |
a0769ee4 LP |
2283 | took place this is now reported to userspace via the new |
2284 | StubPcrKernelImage and StubPcrInitRDSysExts EFI variables. | |
2285 | ||
02380e19 | 2286 | * As before, systemd-stub will measure kernel parameters and system |
7eff3e2c | 2287 | credentials into PCR 12. It will now report this fact via the |
a0769ee4 | 2288 | StubPcrKernelParameters EFI variable to userspace. |
e49d111b | 2289 | |
a0769ee4 LP |
2290 | * The UEFI monotonic boot counter is now included in the updated random |
2291 | seed file maintained by sd-boot, providing some additional entropy. | |
e49d111b | 2292 | |
9ca1efbc ZJS |
2293 | * sd-stub will use LoadImage/StartImage to execute the kernel, instead |
2294 | of arranging the image manually and jumping to the kernel entry | |
2295 | point. sd-stub also installs a temporary UEFI SecurityOverride to | |
2296 | allow the (unsigned) nested image to be booted. This is safe because | |
2297 | the outer (signed) stub+kernel binary must have been verified before | |
2298 | the stub was executed. | |
2299 | ||
e49d111b | 2300 | * Booting in EFI mixed mode (a 64-bit kernel over 32-bit UEFI firmware) |
a0769ee4 | 2301 | is now supported by sd-boot. |
e49d111b | 2302 | |
a0769ee4 LP |
2303 | * bootctl gained a bunch of new options: --all-architectures to install |
2304 | binaries for all supported EFI architectures, --root= and --image= | |
2305 | options to operate on a directory or disk image, and | |
2306 | --install-source= to specify the source for binaries to install, | |
2307 | --efi-boot-option-description= to control the name of the boot entry. | |
e49d111b ZJS |
2308 | |
2309 | * The sd-boot stub exports a StubFeatures flag, which is used by | |
2310 | bootctl to show features supported by the stub that was used to boot. | |
2311 | ||
6e50cf38 JJ |
2312 | * The PE section offsets that are used by tools that assemble unified |
2313 | kernel images have historically been hard-coded. This may lead to | |
2314 | overlapping PE sections which may break on boot. The UKI will now try | |
2315 | to detect and warn about this. | |
2316 | ||
2317 | Any tools that assemble UKIs must update to calculate these offsets | |
2318 | dynamically. Future sd-stub versions may use offsets that will not | |
2319 | work with the currently used set of hard-coded offsets! | |
f77c0840 | 2320 | |
a0769ee4 LP |
2321 | * sd-stub now accepts (and passes to the initrd and then to the full |
2322 | OS) new PE sections '.pcrsig' and '.pcrkey' that can be used to embed | |
02380e19 ZJS |
2323 | signatures of expected PCR values, to allow sealing secrets via the |
2324 | TPM2 against pre-calculated PCR measurements. | |
f77c0840 | 2325 | |
e49d111b ZJS |
2326 | Changes in the hardware database: |
2327 | ||
a0769ee4 | 2328 | * 'systemd-hwdb query' now supports the --root= option. |
e49d111b ZJS |
2329 | |
2330 | Changes in systemctl: | |
2331 | ||
a0769ee4 | 2332 | * systemctl now supports --state= and --type= options for the 'show' |
e49d111b ZJS |
2333 | and 'status' verbs. |
2334 | ||
2335 | * systemctl gained a new verb 'list-automounts' to list automount | |
2336 | points. | |
2337 | ||
a0769ee4 LP |
2338 | * systemctl gained support for a new --image= switch to be able to |
2339 | operate on the specified disk image (similar to the existing --root= | |
2340 | which operates relative to some directory). | |
2341 | ||
e49d111b ZJS |
2342 | Changes in systemd-networkd: |
2343 | ||
2344 | * networkd can set Linux NetLabel labels for integration with the | |
2345 | network control in security modules via a new NetLabel= option. | |
2346 | ||
e49d111b ZJS |
2347 | * The RapidCommit= is (re-)introduced to enable faster configuration |
2348 | via DHCPv6 (RFC 3315). | |
2349 | ||
f77c0840 LB |
2350 | * networkd gained a new option TCPCongestionControlAlgorithm= that |
2351 | allows setting a per-route TCP algorithm. | |
2352 | ||
043ba6a1 LB |
2353 | * networkd gained a new option KeepFileDescriptor= to allow keeping a |
2354 | reference (file descriptor) open on TUN/TAP interfaces, which is | |
2355 | useful to avoid link flaps while the underlying service providing the | |
2356 | interface is being serviced. | |
2357 | ||
ea3e581d LB |
2358 | * RouteTable= now also accepts route table names. |
2359 | ||
e49d111b ZJS |
2360 | Changes in systemd-nspawn: |
2361 | ||
2362 | * The --bind= and --overlay= options now support relative paths. | |
2363 | ||
3af9dc77 | 2364 | * The --bind= option now supports a 'rootidmap' value, which will |
f77c0840 LB |
2365 | use id-mapped mounts to map the root user inside the container to the |
2366 | owner of the mounted directory on the host. | |
e49d111b | 2367 | |
9ca1efbc ZJS |
2368 | Changes in systemd-resolved: |
2369 | ||
2370 | * systemd-resolved now persists DNSOverTLS in its state file too. This | |
2371 | fixes a problem when used in combination with NetworkManager, which | |
2372 | sends the setting only once, causing it to be lost if resolved was | |
2373 | restarted at any point. | |
2374 | ||
68a5300f | 2375 | * systemd-resolved now exposes a Varlink socket at |
9ca1efbc ZJS |
2376 | /run/systemd/resolve/io.systemd.Resolve.Monitor, accessible only for |
2377 | root. Processed DNS requests in a JSON format will be published to | |
2378 | any clients connected to this socket. | |
2379 | ||
2380 | resolvectl gained a 'monitor' verb to make use of this. | |
2381 | ||
2382 | * systemd-resolved now treats unsupported DNSSEC algorithms as INSECURE | |
2383 | instead of returning SERVFAIL, as per RFC: | |
2384 | https://datatracker.ietf.org/doc/html/rfc6840#section-5.2 | |
2385 | ||
2386 | * OpenSSL is the default crypto backend for systemd-resolved. (gnutls | |
2387 | is still supported.) | |
2388 | ||
f77c0840 | 2389 | Changes in libsystemd and other libraries: |
e49d111b | 2390 | |
a0769ee4 LP |
2391 | * libsystemd now exports sd_bus_error_setfv() (a convenience function |
2392 | for setting bus errors), sd_id128_string_equal (a convenience | |
da890466 | 2393 | function for 128-bit ID string comparisons), and |
a0769ee4 LP |
2394 | sd_bus_message_read_strv_extend() (a function to incrementally read |
2395 | string arrays). | |
e49d111b | 2396 | |
a0769ee4 LP |
2397 | * libsystemd now exports sd_device_get_child_first()/_next() as a |
2398 | high-level interface for enumerating child devices. It also supports | |
2399 | sd_device_new_child() for opening a child device given a device | |
2400 | object. | |
f77c0840 | 2401 | |
a0769ee4 | 2402 | * libsystemd now exports sd_device_monitor_set()/get_description() |
9ca1efbc | 2403 | which allow setting a custom description that will be used in log |
a0769ee4 | 2404 | messages by sd_device_monitor*. |
46c41ade | 2405 | |
e49d111b ZJS |
2406 | * Private shared libraries (libsystemd-shared-nnn.so, |
2407 | libsystemd-core-nnn.so) are now installed into arch-specific | |
2408 | directories to allow multi-arch installs. | |
2409 | ||
f77c0840 LB |
2410 | * A new sd-gpt.h header is now published, listing GUIDs from the |
2411 | Discoverable Partitions specification. For more details see: | |
2412 | https://systemd.io/DISCOVERABLE_PARTITIONS/ | |
2413 | ||
a0769ee4 LP |
2414 | * A new function sd_hwdb_new_from_path() has been added to open a hwdb |
2415 | database given an explicit path to the file. | |
2416 | ||
2417 | * The signal number argument to sd_event_add_signal() now can now be | |
02380e19 ZJS |
2418 | ORed with the SD_EVENT_SIGNAL_PROCMASK flag, causing sigprocmask() to |
2419 | be automatically invoked to block the specified signal. This is | |
2420 | useful to simplify invocations as the caller doesn't have to do this | |
2421 | manually. | |
a0769ee4 LP |
2422 | |
2423 | * A new convenience call sd_event_set_signal_exit() has been added to | |
02380e19 | 2424 | sd-event to set up signal handling so that the event loop |
a0769ee4 LP |
2425 | automatically terminates cleanly on SIGTERM/SIGINT. |
2426 | ||
e49d111b ZJS |
2427 | Changes in other components: |
2428 | ||
02380e19 ZJS |
2429 | * systemd-sysusers, systemd-tmpfiles, and systemd-sysctl configuration |
2430 | can now be provided via the credential mechanism. | |
e49d111b | 2431 | |
9ca1efbc ZJS |
2432 | * systemd-analyze gained a new verb 'compare-versions' that implements |
2433 | comparisons for versions strings (similarly to 'rpmdev-vercmp' and | |
2434 | 'dpkg --compare-versions'). | |
2435 | ||
2436 | * 'systemd-analyze dump' is extended to accept glob patterns for unit | |
2437 | names to limit the output to matching units. | |
2438 | ||
02380e19 ZJS |
2439 | * tmpfiles.d/ lines can read file contents to write from a credential. |
2440 | The new modifier char '^' is used to specify that the argument is a | |
2441 | credential name. This mechanism is used to automatically populate | |
a0769ee4 | 2442 | /etc/motd, /etc/issue, and /etc/hosts from credentials. |
e49d111b | 2443 | |
a0769ee4 LP |
2444 | * tmpfiles.d/ may now be configured to avoid changing uid/gid/mode of |
2445 | an inode if the specification is prefixed with ':' and the inode | |
2446 | already exists. | |
f77c0840 | 2447 | |
02380e19 ZJS |
2448 | * Default tmpfiles.d/ configuration now carries a line to automatically |
2449 | use an 'ssh.authorized_keys.root' credential if provided to set up | |
a0769ee4 | 2450 | the SSH authorized_keys file for the root user. |
f77c0840 | 2451 | |
a0769ee4 LP |
2452 | * systemd-tmpfiles will now gracefully handle absent source of "C" copy |
2453 | lines. | |
2454 | ||
02380e19 ZJS |
2455 | * tmpfiles.d/ F/w lines now optionally permit encoding of the payload |
2456 | in base64. This is useful to write arbitrary binary data into files. | |
f77c0840 | 2457 | |
e49d111b | 2458 | * The pkgconfig and rpm macros files now export the directory for user |
02380e19 | 2459 | units as 'user_tmpfiles_dir' and '%_user_tmpfilesdir'. |
e49d111b | 2460 | |
02380e19 ZJS |
2461 | * Detection of Apple Virtualization and detection of Parallels and |
2462 | KubeVirt virtualization on non-x86 archs have been added. | |
e49d111b ZJS |
2463 | |
2464 | * os-release gained a new field SUPPORT_END=YYYY-MM-DD to inform the | |
2465 | user when their system will become unsupported. | |
2466 | ||
2467 | * When performing suspend-then-hibernate, the system will estimate the | |
02380e19 ZJS |
2468 | discharge rate and use that to set the delay until hibernation and |
2469 | hibernate immediately instead of suspending when running from a | |
e49d111b ZJS |
2470 | battery and the capacity is below 5%. |
2471 | ||
a0769ee4 | 2472 | * systemd-sysctl gained a --strict option to fail when a sysctl |
e49d111b ZJS |
2473 | setting is unknown to the kernel. |
2474 | ||
a0769ee4 | 2475 | * machinectl supports --force for the 'copy-to' and 'copy-from' |
e49d111b ZJS |
2476 | verbs. |
2477 | ||
9ca1efbc ZJS |
2478 | * coredumpctl gained the --root and --image options to look for journal |
2479 | files under the specified root directory, image, or block device. | |
e49d111b | 2480 | |
02380e19 ZJS |
2481 | * 'journalctl -o' and similar commands now implement a new output mode |
2482 | "short-delta". It is similar to "short-monotonic", but also shows the | |
2483 | time delta between subsequent messages. | |
893bcd3d | 2484 | |
02380e19 ZJS |
2485 | * journalctl now respects the --quiet flag when verifying consistency |
2486 | of journal files. | |
f77c0840 | 2487 | |
02380e19 ZJS |
2488 | * Journal log messages gained a new implicit field _RUNTIME_SCOPE= that |
2489 | will indicate whether a message was logged in the 'initrd' phase or | |
2490 | in the 'system' phase of the boot process. | |
f77c0840 | 2491 | |
02380e19 ZJS |
2492 | * Journal files gained a new compatibility flag |
2493 | 'HEADER_INCOMPATIBLE_COMPACT'. Files with this flag implement changes | |
2494 | to the storage format that allow reducing size on disk. As with other | |
2495 | compatibility flags, older journalctl versions will not be able to | |
2496 | read journal files using this new format. The environment variable | |
2497 | 'SYSTEMD_JOURNAL_COMPACT=0' can be passed to systemd-journald to | |
2498 | disable this functionality. It is enabled by default. | |
7dbbb393 | 2499 | |
a0769ee4 LP |
2500 | * systemd-run's --working-directory= switch now works when used in |
2501 | combination with --scope. | |
f77c0840 | 2502 | |
9ca1efbc ZJS |
2503 | * portablectl gained a --force flag to skip certain sanity checks. This |
2504 | is implemented using new flags accepted by systemd-portabled for the | |
2505 | *WithExtensions() D-Bus methods: SD_SYSTEMD_PORTABLE_FORCE_ATTACH | |
2506 | flag now means that the attach/detach checks whether the units are | |
2507 | already present and running will be skipped. Similarly, | |
2508 | SD_SYSTEMD_PORTABLE_FORCE_SYSEXT flag means that the check whether | |
2509 | image name matches the name declared inside of the image will be | |
2510 | skipped. Callers must be sure to do those checks themselves if | |
2511 | appropriate. | |
f77c0840 | 2512 | |
043ba6a1 LB |
2513 | * systemd-portabled will now use the original filename to check |
2514 | extension-release.NAME for correctness, in case it is passed a | |
2515 | symlink. | |
2516 | ||
46c41ade LB |
2517 | * systemd-portabled now uses PrivateTmp=yes in the 'trusted' profile |
2518 | too. | |
2519 | ||
a0769ee4 LP |
2520 | * sysext's extension-release files now support '_any' as a special |
2521 | value for the ID= field, to allow distribution-independent extensions | |
2522 | (e.g.: fully statically compiled binaries, scripts). It also gained | |
2523 | support for a new ARCHITECTURE= field that may be used to explicitly | |
2524 | restrict an image to hosts of a specific architecture. | |
043ba6a1 | 2525 | |
02380e19 ZJS |
2526 | * systemd-repart now supports creating squashfs partitions. This |
2527 | requires mksquashfs from squashfs-tools. | |
f77c0840 | 2528 | |
02380e19 | 2529 | * systemd-repart gained a --split flag to also generate split |
a0769ee4 | 2530 | artifacts, i.e. a separate file for each partition. This is useful in |
766c1eae | 2531 | conjunction with systemd-sysupdate or other tools, or to generate |
a0769ee4 | 2532 | split dm-verity artifacts. |
f77c0840 LB |
2533 | |
2534 | * systemd-repart is now able to generate dm-verity partitions, including | |
2535 | signatures. | |
2536 | ||
02380e19 ZJS |
2537 | * systemd-repart can now set a partition UUID to zero, allowing it to |
2538 | be filled in later, such as when using verity partitions. | |
f77c0840 | 2539 | |
46c41ade LB |
2540 | * systemd-repart now supports drop-ins for its configuration files. |
2541 | ||
02380e19 | 2542 | * Package metadata logged by systemd-coredump in the system journal is |
f77c0840 LB |
2543 | now more compact. |
2544 | ||
2545 | * xdg-autostart-service now expands 'tilde' characters in Exec lines. | |
2546 | ||
2547 | * systemd-oomd now automatically links against libatomic, if available. | |
2548 | ||
46c41ade LB |
2549 | * systemd-oomd now sends out a 'Killed' D-Bus signal when a cgroup is |
2550 | killed. | |
2551 | ||
2552 | * scope units now also provide oom-kill status. | |
2553 | ||
a0769ee4 LP |
2554 | * systemd-pstore will now try to load only the efi_pstore kernel module |
2555 | before running, ensuring that pstore can be used. | |
f77c0840 | 2556 | |
46c41ade LB |
2557 | * systemd-logind gained a new StopIdleSessionSec= option to stop an idle |
2558 | session after a preconfigure timeout. | |
f77c0840 LB |
2559 | |
2560 | * systemd-homed will now wait up to 30 seconds for workers to terminate, | |
2561 | rather than indefinitely. | |
2562 | ||
7dbbb393 LB |
2563 | * homectl gained a new '--luks-sector-size=' flag that allows users to |
2564 | select the preferred LUKS sector size. Must be a power of 2 between 512 | |
2565 | and 4096. systemd-userdbd records gained a corresponding field. | |
2566 | ||
f77c0840 LB |
2567 | * systemd-sysusers will now respect the 'SOURCE_DATE_EPOCH' environment |
2568 | variable when generating the 'sp_lstchg' field, to ensure an image | |
2569 | build can be reproducible. | |
2570 | ||
02380e19 | 2571 | * 'udevadm wait' will now listen to kernel uevents too when called with |
a0769ee4 LP |
2572 | --initialized=no. |
2573 | ||
2574 | * When naming network devices udev will now consult the Devicetree | |
2575 | "alias" fields for the device. | |
f77c0840 | 2576 | |
9ca1efbc ZJS |
2577 | * systemd-udev will now create infiniband/by-path and |
2578 | infiniband/by-ibdev links for Infiniband verbs devices. | |
2579 | ||
ea3e581d LB |
2580 | * systemd-udev-trigger.service will now also prioritize input devices. |
2581 | ||
a0769ee4 LP |
2582 | * ConditionACPower= and systemd-ac-power will now assume the system is |
2583 | running on AC power if no battery can be found. | |
46c41ade | 2584 | |
f77c0840 LB |
2585 | * All features and tools using the TPM2 will now communicate with it |
2586 | using a bind key. Beforehand, the tpm2 support used encrypted sessions | |
2587 | by creating a primary key that was used to encrypt traffic. This | |
2588 | creates a problem as the key created for encrypting the traffic could | |
2589 | be faked by an active interposer on the bus. In cases when a pin is | |
2590 | used, a bind key will be used. The pin is used as the auth value for | |
2591 | the seal key, aka the disk encryption key, and that auth value will be | |
2592 | used in the session establishment. An attacker would need the pin | |
2593 | value to create the secure session and thus an active interposer | |
02380e19 | 2594 | without the pin cannot interpose on TPM2 traffic. |
f77c0840 | 2595 | |
043ba6a1 | 2596 | * systemd-growfs no longer requires udev to run. |
f77c0840 | 2597 | |
46c41ade LB |
2598 | * systemd-backlight now will better support systems with multiple |
2599 | graphic cards. | |
2600 | ||
2601 | * systemd-cryptsetup's keyfile-timeout= option now also works when a | |
2602 | device is used as a keyfile. | |
2603 | ||
a0769ee4 LP |
2604 | * systemd-cryptenroll gained a new --unlock-key-file= option to get the |
2605 | unlocking key from a key file (instead of prompting the user). Note | |
2606 | that this is the key for unlocking the volume in order to be able to | |
2607 | enroll a new key, but it is not the key that is enrolled. | |
46c41ade | 2608 | |
a0769ee4 LP |
2609 | * systemd-dissect gained a new --umount switch that will safely and |
2610 | synchronously unmount all partitions of an image previously mounted | |
02380e19 | 2611 | with 'systemd-dissect --mount'. |
46c41ade LB |
2612 | |
2613 | * When using gcrypt, all systemd tools and services will now configure | |
02380e19 | 2614 | it to prefer the OS random number generator if present. |
46c41ade | 2615 | |
ea3e581d LB |
2616 | * All example code shipped with documentation has been relicensed from CC0 |
2617 | to MIT-0. | |
2618 | ||
2619 | * Unit tests will no longer fail when running on a system without | |
2620 | /etc/machine-id. | |
2621 | ||
e49d111b ZJS |
2622 | Experimental features: |
2623 | ||
f77c0840 LB |
2624 | * BPF programs can now be compiled with bpf-gcc (requires libbpf >= 1.0 |
2625 | and bpftool >= 7.0). | |
e49d111b ZJS |
2626 | |
2627 | * sd-boot can automatically enroll SecureBoot keys from files found on | |
2628 | the ESP. This enrollment can be either automatic ('force' mode) or | |
a0769ee4 LP |
2629 | controlled by the user ('manual' mode). It is sufficient to place the |
2630 | SecureBoot keys in the right place in the ESP and they will be picked | |
2631 | up by sd-boot and shown in the boot menu. | |
2632 | ||
9ca1efbc ZJS |
2633 | * The mkosi config in systemd gained support for automatically |
2634 | compiling a kernel with the configuration appropriate for testing | |
2635 | systemd. This may be useful when developing or testing systemd in | |
2636 | tandem with the kernel. | |
2637 | ||
9ca1efbc ZJS |
2638 | Contributions from: 김인수, Adam Williamson, adrian5, Aidan Dang, |
2639 | Akihiko Odaki, Alban Bedel, Albert Mikaelyan, Aleksey Vasenev, | |
fa9b3a5f | 2640 | Alexander Graf, Alexander Shopov, Alexander Wilson, |
f3e70c8a LB |
2641 | Alper Nebi Yasak, anarcat, Anders Jonsson, Andre Kalb, |
2642 | Andrew Stone, Andrey Albershteyn, Anita Zhang, Ansgar Burchardt, | |
fa9b3a5f LB |
2643 | Antonio Alvarez Feijoo, Arnaud Ferraris, Aryan singh, asavah, |
2644 | Avamander, Avram Lubkin, Balázs Meskó, Bastien Nocera, | |
2645 | Benjamin Franzke, BerndAdameit, bin456789, Celeste Liu, | |
2646 | Chih-Hsuan Yen, Christian Brauner, Christian Göttsche, | |
2647 | Christian Hesse, Clyde Byrd III, codefiles, Colin Walters, | |
2648 | Cristian Rodríguez, Daan De Meyer, Daniel Braunwarth, | |
2649 | Daniel Rusek, Dan Streetman, Darsey Litzenberger, David Edmundson, | |
9ca1efbc ZJS |
2650 | David Jaša, David Rheinsberg, David Seifert, David Tardon, |
2651 | dependabot[bot], Devendra Tewari, Dominique Martinet, drosdeck, | |
2652 | Edson Juliano Drosdeck, Eduard Tolosa, eggfly, Einsler Lee, | |
2653 | Elias Probst, Eli Schwartz, Evgeny Vereshchagin, exploide, Fei Li, | |
2654 | Foster Snowhill, Franck Bui, Frank Dana, Frantisek Sumsal, | |
fa9b3a5f LB |
2655 | Gerd Hoffmann, Gio, Goffredo Baroncelli, gtwang01, |
2656 | Guillaume W. Bres, H A, Hans de Goede, Heinrich Schuchardt, | |
2657 | Hugo Carvalho, i-do-cpp, igo95862, j00512545, Jacek Migacz, | |
2658 | Jade Bilkey, James Hilliard, Jan B, Janis Goldschmidt, | |
2659 | Jan Janssen, Jan Kuparinen, Jan Luebbe, Jan Macku, | |
2660 | Jason A. Donenfeld, Javkhlanbayar Khongorzul, Jeremy Soller, | |
2661 | JeroenHD, jiangchuangang, João Loureiro, | |
f3e70c8a LB |
2662 | Joaquín Ignacio Aramendía, Jochen Sprickerhof, |
2663 | Johannes Schauer Marin Rodrigues, Jonas Kümmerlin, | |
2664 | Jonas Witschel, Jonathan Kang, Jonathan Lebon, Joost Heitbrink, | |
2665 | Jörg Thalheim, josh-gordon-fb, Joyce, Kai Lueke, lastkrick, | |
2666 | Lennart Poettering, Leon M. George, licunlong, Li kunyu, | |
9ca1efbc ZJS |
2667 | LockBlock-dev, Loïc Collignon, Lubomir Rintel, Luca Boccassi, |
2668 | Luca BRUNO, Ludwig Nussel, Łukasz Stelmach, Maccraft123, | |
2669 | Marc Kleine-Budde, Marius Vollmer, Martin Wilck, matoro, | |
2670 | Matthias Lisin, Max Gautier, Maxim Mikityanskiy, Michael Biebl, | |
2671 | Michal Koutný, Michal Sekletár, Michal Stanke, Mike Gilbert, | |
fa9b3a5f | 2672 | Mitchell Freiderich, msizanoen1, Nick Rosbrook, nl6720, Oğuz Ersen, |
f3e70c8a LB |
2673 | Oleg Solovyov, Olga Smirnova, Pablo Ceballos, Pavel Zhukov, |
2674 | Phaedrus Leeds, Philipp Gortan, Piotr Drąg, Pyfisch, | |
2675 | Quentin Deslandes, Rahil Bhimjiani, Rene Hollander, Richard Huang, | |
2676 | Richard Phibel, Rudi Heitbaum, Sam James, Sarah Brofeldt, | |
2677 | Sean Anderson, Sebastian Scheibner, Shreenidhi Shedi, | |
2678 | Sonali Srivastava, Steve Ramage, Suraj Krishnan, Swapnil Devesh, | |
2679 | Takashi Sakamoto, Ted X. Toth, Temuri Doghonadze, Thomas Blume, | |
2680 | Thomas Haller, Thomas Hebb, Tomáš Hnyk, Tomasz Paweł Gajc, | |
2681 | Topi Miettinen, Ulrich Ölmann, undef, Uriel Corfa, | |
2682 | Victor Westerhuis, Vincent Dagonneau, Vishal Chillara Srinivas, | |
2683 | Vito Caputo, Weblate, Wenchao Hao, William Roberts, williamsumendap, | |
2684 | wineway, xiaoyang, Yuri Chornoivan, Yu Watanabe, | |
2685 | Zbigniew Jędrzejewski-Szmek, Zhaofeng Li, наб | |
23992ce1 | 2686 | |
e8dc5276 | 2687 | – The Great Beyond, 2022-10-31 👻 |
e49d111b | 2688 | |
73849408 | 2689 | CHANGES WITH 251: |
00b29ca1 ZJS |
2690 | |
2691 | Backwards-incompatible changes: | |
2692 | ||
61ade257 | 2693 | * The minimum kernel version required has been bumped from 3.13 to 4.15, |
b586cbde LB |
2694 | and CLOCK_BOOTTIME is now assumed to always exist. |
2695 | ||
7503fbd4 | 2696 | * C11 with GNU extensions (aka "gnu11") is now used to build our |
8c70e802 | 2697 | components. Public API headers are still restricted to ISO C89. |
7503fbd4 | 2698 | |
00b29ca1 ZJS |
2699 | * In v250, a systemd-networkd feature that automatically configures |
2700 | routes to addresses specified in AllowedIPs= was added and enabled by | |
2701 | default. However, this causes network connectivity issues in many | |
2702 | existing setups. Hence, it has been disabled by default since | |
2703 | systemd-stable 250.3. The feature can still be used by explicitly | |
2704 | configuring RouteTable= setting in .netdev files. | |
1d2842d1 | 2705 | |
169bb1de LB |
2706 | * Jobs started via StartUnitWithFlags() will no longer return 'skipped' |
2707 | when a Condition*= check does not succeed, restoring the JobRemoved | |
2708 | signal to the behaviour it had before v250. | |
2709 | ||
e1f0c136 LP |
2710 | * The org.freedesktop.portable1 methods GetMetadataWithExtensions() and |
2711 | GetImageMetadataWithExtensions() have been fixed to provide an extra | |
00b29ca1 ZJS |
2712 | return parameter, containing the actual extension release metadata. |
2713 | The current implementation was judged to be broken and unusable, and | |
2714 | thus the usual procedure of adding a new set of methods was skipped, | |
2715 | and backward compatibility broken instead on the assumption that | |
2716 | nobody can be affected given the current state of this interface. | |
ce3ca32c | 2717 | |
65df0ce3 ZJS |
2718 | * All kernels supported by systemd mix bytes returned by RDRAND (or |
2719 | similar) into the entropy pool at early boot. This means that on | |
2720 | those systems, even if /dev/urandom is not yet initialized, it still | |
2721 | returns bytes that are of at least RDRAND quality. For that reason, | |
2722 | we no longer have reason to invoke RDRAND from systemd itself, which | |
2723 | has historically been a source of bugs. Furthermore, kernels ≥5.6 | |
2724 | provide the getrandom(GRND_INSECURE) interface for returning random | |
2725 | bytes before the entropy pool is initialized without warning into | |
2726 | kmsg, which is what we attempt to use if available. systemd's direct | |
2727 | usage of RDRAND has been removed. x86 systems ≥Broadwell that are | |
2728 | running an older kernel may experience kmsg warnings that were not | |
2729 | seen with 250. For newer kernels, non-x86 systems, or older x86 | |
2730 | systems, there should be no visible changes. | |
00b29ca1 | 2731 | |
7eff3e2c ZJS |
2732 | * sd-boot will now measure the kernel command line into TPM PCR 12 |
2733 | rather than PCR 8. This improves usefulness of the measurements on | |
60f53dd5 | 2734 | systems where sd-boot is chainloaded from Grub. Grub measures all |
7eff3e2c ZJS |
2735 | commands its executes into PCR 8, which makes it very hard to use |
2736 | reasonably, hence separate ourselves from that and use PCR 12 | |
00b29ca1 | 2737 | instead, which is what certain Ubuntu editions already do. To retain |
0f6f9dc6 | 2738 | compatibility with systems running older systemd systems a new meson |
00b29ca1 | 2739 | option 'efi-tpm-pcr-compat' has been added (which defaults to false). |
7eff3e2c ZJS |
2740 | If enabled, the measurement is done twice: into the new-style PCR 12 |
2741 | *and* the old-style PCR 8. It's strongly advised to migrate all users | |
2742 | to PCR 12 for this purpose in the long run, as we intend to remove | |
ec5e113f | 2743 | this compatibility feature in two years' time. |
00b29ca1 ZJS |
2744 | |
2745 | * busctl capture now writes output in the newer pcapng format instead | |
2746 | of pcap. | |
2747 | ||
65df0ce3 ZJS |
2748 | * A udev rule that imported hwdb matches for USB devices with lowercase |
2749 | hexadecimal vendor/product ID digits was added in systemd 250. This | |
2750 | has been reverted, since uppercase hexadecimal digits are supposed to | |
2751 | be used, and we already had a rule with the appropriate match. | |
00b29ca1 ZJS |
2752 | |
2753 | Users might need to adjust their local hwdb entries. | |
2754 | ||
e1f0c136 LP |
2755 | * arch_prctl(2) has been moved to the @default set in the syscall filters |
2756 | (as exposed via the SystemCallFilter= setting in service unit files). | |
00b29ca1 ZJS |
2757 | It is apparently used by the linker now. |
2758 | ||
eb1446f8 DDM |
2759 | * The tmpfiles entries that create the /run/systemd/netif directory and |
2760 | its subdirectories were moved from tmpfiles.d/systemd.conf to | |
2761 | tmpfiles.d/systemd-network.conf. | |
2762 | ||
2763 | Users might need to adjust their files that override tmpfiles.d/systemd.conf | |
2764 | to account for this change. | |
2765 | ||
7f2ec323 LB |
2766 | * The requirement for Portable Services images to contain a well-formed |
2767 | os-release file (i.e.: contain at least an ID field) is now enforced. | |
2768 | This applies to base images and extensions, and also to systemd-sysext. | |
2769 | ||
942473dc | 2770 | Changes in the Boot Loader Specification, kernel-install and sd-boot: |
00b29ca1 | 2771 | |
5e9c57d2 LP |
2772 | * kernel-install's and bootctl's Boot Loader Specification Type #1 |
2773 | entry generation logic has been reworked. The user may now pick | |
2774 | explicitly by which "token" string to name the installation's boot | |
00b29ca1 | 2775 | entries, via the new /etc/kernel/entry-token file or the new |
5e9c57d2 LP |
2776 | --entry-token= switch to bootctl. By default — as before — the |
2777 | entries are named after the local machine ID. However, in "golden | |
2778 | image" environments, where the machine ID shall be initialized on | |
2779 | first boot (as opposed to at installation time before first boot) the | |
e1f0c136 | 2780 | machine ID will not be available at build time. In this case the |
00b29ca1 ZJS |
2781 | --entry-token= switch to bootctl (or the /etc/kernel/entry-token |
2782 | file) may be used to override the "token" for the entries, for | |
2783 | example the IMAGE_ID= or ID= fields from /etc/os-release. This will | |
2784 | make the OS images independent of any machine ID, and ensure that the | |
2785 | images will not carry any identifiable information before first boot, | |
2786 | but on the other hand means that multiple parallel installations of | |
2787 | the very same image on the same disk cannot be supported. | |
5e9c57d2 | 2788 | |
00b29ca1 ZJS |
2789 | Summary: if you are building golden images that shall acquire |
2790 | identity information exclusively on first boot, make sure to both | |
2791 | remove /etc/machine-id *and* to write /etc/kernel/entry-token to the | |
e1f0c136 | 2792 | value of the IMAGE_ID= or ID= field of /etc/os-release or another |
00b29ca1 | 2793 | suitable identifier before deploying the image. |
deb5c820 | 2794 | |
00b29ca1 | 2795 | * The Boot Loader Specification has been extended with |
e1f0c136 LP |
2796 | /loader/entries.srel file located in the EFI System Partition (ESP) |
2797 | that disambiguates the format of the entries in the /loader/entries/ | |
2798 | directory (in order to discern them from incompatible uses of this | |
2799 | directory by other projects). For entries that follow the | |
2800 | Specification, the string "type1" is stored in this file. | |
3fbd5f20 | 2801 | |
e1f0c136 LP |
2802 | bootctl will now write this file automatically when installing the |
2803 | systemd-boot boot loader. | |
00b29ca1 ZJS |
2804 | |
2805 | * kernel-install supports a new initrd_generator= setting in | |
2806 | /etc/kernel/install.conf, that is exported as | |
2807 | $KERNEL_INSTALL_INITRD_GENERATOR to kernel-install plugins. This | |
e1f0c136 | 2808 | allows choosing different initrd generators. |
00b29ca1 ZJS |
2809 | |
2810 | * kernel-install will now create a "staging area" (an initially-empty | |
2811 | directory to gather files for a Boot Loader Specification Type #1 | |
2812 | entry). The path to this directory is exported as | |
2813 | $KERNEL_INSTALL_STAGING_AREA to kernel-install plugins, which should | |
2814 | drop files there instead of writing them directly to the final | |
2815 | location. kernel-install will move them when all files have been | |
2816 | prepared successfully. | |
429cddba | 2817 | |
0c6e746b YW |
2818 | * New option sort-key= has been added to the Boot Loader Specification |
2819 | to override the sorting order of the entries in the boot menu. It is | |
2820 | read by sd-boot and bootctl, and will be written by kernel-install, | |
2821 | with the default value of IMAGE_ID= or ID= fields from | |
2822 | os-release. Together, this means that on multiboot installations, | |
2823 | entries should be grouped and sorted in a predictable way. | |
2824 | ||
942473dc ZJS |
2825 | * The sort order of boot entries has been updated: entries which have |
2826 | the new field sort-key= are sorted by it first, and all entries | |
2827 | without it are ordered later. After that, entries are sorted by | |
2828 | version so that newest entries are towards the beginning of the list. | |
2829 | ||
0c6e746b YW |
2830 | * The kernel-install tool gained a new 'inspect' verb which shows the |
2831 | paths and other settings used. | |
2832 | ||
2833 | * sd-boot can now optionally beep when the menu is shown and menu | |
2834 | entries are selected, which can be useful on machines without a | |
2835 | working display. (Controllable via a loader.conf setting.) | |
2836 | ||
2837 | * The --make-machine-id-directory= switch to bootctl has been replaced | |
2838 | by --make-entry-directory=, given that the entry directory is not | |
2839 | necessarily named after the machine ID, but after some other suitable | |
2840 | ID as selected via --entry-token= described above. The old name of | |
2841 | the option is still understood to maximize compatibility. | |
2842 | ||
2843 | * 'bootctl list' gained support for a new --json= switch to output boot | |
2844 | menu entries in JSON format. | |
2845 | ||
211b564a ZJS |
2846 | * 'bootctl is-installed' now supports the --graceful, and various verbs |
2847 | omit output with the new option --quiet. | |
2848 | ||
942473dc | 2849 | Changes in systemd-homed: |
0c6e746b | 2850 | |
dfdaf9f2 LP |
2851 | * Starting with v250 systemd-homed uses UID/GID mapping on the mounts |
2852 | of activated home directories it manages (if the kernel and selected | |
2853 | file systems support it). So far it mapped three UID ranges: the | |
2854 | range from 0…60000, the user's own UID, and the range 60514…65534, | |
da890466 | 2855 | leaving everything else unmapped (in other words, the 16-bit UID range |
dfdaf9f2 | 2856 | is mapped almost fully, with the exception of the UID subrange used |
00b29ca1 ZJS |
2857 | for systemd-homed users, with one exception: the user's own UID). |
2858 | Unmapped UIDs may not be used for file ownership in the home | |
5cf84d25 | 2859 | directory — any chown() attempts with them will fail. With this |
dfdaf9f2 LP |
2860 | release a fourth range is added to these mappings: |
2861 | 524288…1879048191. This range is the UID range intended for container | |
2862 | uses, see: | |
2863 | ||
2864 | https://systemd.io/UIDS-GIDS | |
2865 | ||
2866 | This range may be used for container managers that place container OS | |
2867 | trees in the home directory (which is a questionable approach, for | |
2868 | quota, permission, SUID handling and network file system | |
2869 | compatibility reasons, but nonetheless apparently commonplace). Note | |
2870 | that this mapping is mapped 1:1 in a pass-through fashion, i.e. the | |
2871 | UID assignments from the range are not managed or mapped by | |
2872 | `systemd-homed`, and must be managed with other mechanisms, in the | |
2873 | context of the local system. | |
2874 | ||
2875 | Typically, a better approach to user namespacing in relevant | |
2876 | container managers would be to leave container OS trees on disk at | |
2877 | UID offset 0, but then map them to a dynamically allocated runtime | |
2878 | UID range via another UID mount map at container invocation | |
2879 | time. That way user namespace UID ranges become strictly a runtime | |
2880 | concept, and do not leak into persistent file systems, persistent | |
2881 | user databases or persistent configuration, thus greatly simplifying | |
2882 | handling, and improving compatibility with home directories intended | |
2883 | to be portable like the ones managed by systemd-homed. | |
7e7a9f9c | 2884 | |
942473dc | 2885 | Changes in shared libraries: |
00b29ca1 ZJS |
2886 | |
2887 | * A new libsystemd-core-<version>.so private shared library is | |
2888 | installed under /usr/lib/systemd/system, mirroring the existing | |
2889 | libsystemd-shared-<version>.so library. This allows the total | |
e1f0c136 | 2890 | installation size to be reduced by binary code reuse. |
00b29ca1 | 2891 | |
e1f0c136 | 2892 | * The <version> tag used in the name of libsystemd-shared.so and |
0f6f9dc6 ZJS |
2893 | libsystemd-core.so can be configured via the meson option |
2894 | 'shared-lib-tag'. Distributions may build subsequent versions of the | |
2895 | systemd package with unique tags (e.g. the full package version), | |
2896 | thus allowing multiple installations of those shared libraries to be | |
2897 | available at the same time. This is intended to fix an issue where | |
2898 | programs that link to those libraries would fail to execute because | |
2899 | they were installed earlier or later than the appropriate version of | |
2900 | the library. | |
00b29ca1 | 2901 | |
0c6e746b YW |
2902 | * The sd-id128 API gained a new call sd_id128_to_uuid_string() that is |
2903 | similar to sd_id128_to_string() but formats the ID in RFC 4122 UUID | |
ec5e113f | 2904 | format instead of as a simple series of hex characters. |
0c6e746b | 2905 | |
61ade257 LP |
2906 | * The sd-device API gained two new calls sd_device_new_from_devname() |
2907 | and sd_device_new_from_path() which permit allocating an sd_device | |
2908 | object from a device node name or file system path. | |
2909 | ||
2910 | * sd-device also gained a new call sd_device_open() which will open the | |
2911 | device node associated with a device for which an sd_device object | |
2912 | has been allocated. The call is supposed to address races around | |
2913 | device nodes being removed/recycled due to hotplug events, or media | |
2914 | change events: the call checks internally whether the major/minor of | |
2915 | the device node and the "diskseq" (in case of block devices) match | |
2916 | with the metadata loaded in the sd_device object, thus ensuring that | |
2917 | the device once opened really matches the provided sd_device object. | |
2918 | ||
942473dc | 2919 | Changes in PID1, systemctl, and systemd-oomd: |
0c6e746b | 2920 | |
a8c122c4 LB |
2921 | * A new set of service monitor environment variables will be passed to |
2922 | OnFailure=/OnSuccess= handlers, but only if exactly one unit lists the | |
2923 | handler unit as OnFailure=/OnSuccess=. The variables are: | |
2924 | $MONITOR_SERVICE_RESULT, $MONITOR_EXIT_CODE, $MONITOR_EXIT_STATUS, | |
2925 | $MONITOR_INVOCATION_ID and $MONITOR_UNIT. For cases when a single | |
2926 | handler needs to watch multiple units, use a templated handler. | |
2927 | ||
e1f0c136 LP |
2928 | * A new ExtensionDirectories= setting in service unit files allows |
2929 | system extensions to be loaded from a directory. (It is similar to | |
2930 | ExtensionImages=, but takes paths to directories, instead of | |
2931 | disk image files.) | |
00b29ca1 | 2932 | |
e1f0c136 | 2933 | 'portablectl attach --extension=' now also accepts directory paths. |
00b29ca1 | 2934 | |
e1f0c136 LP |
2935 | * The user.delegate and user.invocation_id extended attributes on |
2936 | cgroups are used in addition to trusted.delegate and | |
2937 | trusted.invocation_id. The latter pair requires privileges to set, | |
2938 | but the former doesn't and can be also set by the unprivileged user | |
2939 | manager. | |
00b29ca1 ZJS |
2940 | |
2941 | (Only supported on kernels ≥5.6.) | |
2942 | ||
942473dc ZJS |
2943 | * Units that were killed by systemd-oomd will now have a service result |
2944 | of 'oom-kill'. The number of times a service was killed is tallied | |
2945 | in the 'user.oomd_ooms' extended attribute. | |
2946 | ||
2947 | The OOMPolicy= unit file setting is now also honoured by | |
2948 | systemd-oomd. | |
2949 | ||
e1f0c136 LP |
2950 | * In unit files the new %y/%Y specifiers can be used to refer to |
2951 | normalized unit file path, which is particularly useful for symlinked | |
2952 | unit files. | |
00b29ca1 | 2953 | |
d0aba07f | 2954 | The new %q specifier resolves to the pretty hostname |
e1f0c136 | 2955 | (i.e. PRETTY_HOSTNAME= from /etc/machine-info). |
00b29ca1 | 2956 | |
e1f0c136 LP |
2957 | The new %d specifier resolves to the credentials directory of a |
2958 | service (same as $CREDENTIALS_DIRECTORY). | |
00b29ca1 | 2959 | |
f72f8021 LB |
2960 | * The RootDirectory=, MountAPIVFS=, ExtensionDirectories=, |
2961 | *Capabilities*=, ProtectHome=, *Directory=, TemporaryFileSystem=, | |
2962 | PrivateTmp=, PrivateDevices=, PrivateNetwork=, NetworkNamespacePath=, | |
2963 | PrivateIPC=, IPCNamespacePath=, PrivateUsers=, ProtectClock=, | |
2964 | ProtectKernelTunables=, ProtectKernelModules=, ProtectKernelLogs=, | |
2965 | MountFlags= service settings now also work in unprivileged user | |
2966 | services, i.e. those run by the user's --user service manager, as long | |
2967 | as user namespaces are enabled on the system. | |
60a777b5 | 2968 | |
e1f0c136 LP |
2969 | * Services with Restart=always and a failing ExecCondition= will no |
2970 | longer be restarted, to bring ExecCondition= behaviour in line with | |
2971 | Condition*= settings. | |
00b29ca1 ZJS |
2972 | |
2973 | * LoadCredential= now accepts a directory as the argument; all files | |
e1f0c136 | 2974 | from the directory will be loaded as credentials. |
00b29ca1 | 2975 | |
60a777b5 LP |
2976 | * A new D-Bus property ControlGroupId is now exposed on service units, |
2977 | that encapsulates the service's numeric cgroup ID that newer kernels | |
d6297626 | 2978 | assign to each cgroup. |
60a777b5 | 2979 | |
0c6e746b YW |
2980 | * PID 1 gained support for configuring the "pre-timeout" of watchdog |
2981 | devices and the associated governor, via the new | |
2982 | RuntimeWatchdogPreSec= and RuntimeWatchdogPreGovernor= configuration | |
2983 | options in /etc/systemd/system.conf. | |
00b29ca1 | 2984 | |
0c6e746b YW |
2985 | * systemctl's --timestamp= option gained a new choice "unix", to show |
2986 | timestamp as unix times, i.e. seconds since 1970, Jan 1st. | |
2987 | ||
61ade257 LP |
2988 | * A new "taint" flag named "old-kernel" is introduced which is set when |
2989 | the kernel systemd runs on is older then the current baseline version | |
2990 | (see above). The flag is shown in "systemctl status" output. | |
2991 | ||
2992 | * Two additional taint flags "short-uid-range" and "short-gid-range" | |
2993 | have been added as well, which are set when systemd notices it is run | |
2994 | within a userns namespace that does not define the full 0…65535 UID | |
2995 | range | |
2996 | ||
2997 | * A new "unmerged-usr" taint flag has been added that is set whenever | |
2998 | running on systems where /bin/ + /sbin/ are *not* symlinks to their | |
ec5e113f PL |
2999 | counterparts in /usr/, i.e. on systems where the /usr/-merge has not |
3000 | been completed. | |
61ade257 LP |
3001 | |
3002 | * Generators invoked by PID 1 will now have a couple of useful | |
3003 | environment variables set describing the execution context a | |
3004 | bit. $SYSTEMD_SCOPE encodes whether the generator is called from the | |
3005 | system service manager, or from the per-user service | |
3006 | manager. $SYSTEMD_IN_INITRD encodes whether the generator is invoked | |
3007 | in initrd context or on the host. $SYSTEMD_FIRST_BOOT encodes whether | |
3008 | systemd considers the current boot to be a "first" | |
3009 | boot. $SYSTEMD_VIRTUALIZATION encode whether virtualization is | |
3010 | detected and which type of hypervisor/container | |
3011 | manager. $SYSTEMD_ARCHITECTURE indicates which architecture the | |
3012 | kernel is built for. | |
3013 | ||
be1e6592 LP |
3014 | * PID 1 will now automatically pick up system credentials from qemu's |
3015 | fw_cfg interface, thus allowing passing arbitrary data into VM | |
3016 | systems similar to how this is already supported for passing them | |
3017 | into `systemd-nspawn` containers. Credentials may now also be passed | |
3018 | in via the new kernel command line option `systemd.set_credential=` | |
3019 | (note that kernel command line options are world-readable during | |
3020 | runtime, and only useful for credentials that require no | |
3021 | confidentiality). The credentials that can be passed to unified | |
3022 | kernels that use the `systemd-stub` UEFI stub are now similarly | |
3023 | picked up automatically. Automatic importing of system credentials | |
3024 | this way can be turned off via the new | |
3025 | `systemd.import_credentials=no` kernel command line option. | |
3026 | ||
50db8d97 ZJS |
3027 | * LoadCredential= will now automatically look for credentials in the |
3028 | /etc/credstore/, /run/credstore/, /usr/lib/credstore/ directories if | |
3029 | the argument is not an absolute path. Similarly, | |
3030 | LoadCredentialEncrypted= will check the same directories plus | |
be1e6592 | 3031 | /etc/credstore.encrypted/, /run/credstore.encrypted/ and |
50db8d97 ZJS |
3032 | /usr/lib/credstore.encrypted/. The idea is to use those directories |
3033 | as the system-wide location for credentials that services should pick | |
3034 | up automatically. | |
be1e6592 LP |
3035 | |
3036 | * System and service credentials are described in great detail in a new | |
3037 | document: | |
3038 | ||
3039 | https://systemd.io/CREDENTIALS | |
3040 | ||
942473dc | 3041 | Changes in systemd-journald: |
0c6e746b YW |
3042 | |
3043 | * The journal JSON export format has been added to listed of stable | |
3044 | interfaces (https://systemd.io/PORTABILITY_AND_STABILITY/). | |
3045 | ||
3046 | * journalctl --list-boots now supports JSON output and the --reverse option. | |
3047 | ||
3048 | * Under docs/: JOURNAL_EXPORT_FORMATS was imported from the wiki and | |
3049 | updated, BUILDING_IMAGES is new: | |
3050 | ||
3051 | https://systemd.io/JOURNAL_EXPORT_FORMATS | |
3052 | https://systemd.io/BUILDING_IMAGES | |
3053 | ||
942473dc | 3054 | Changes in udev: |
0c6e746b YW |
3055 | |
3056 | * Two new hwdb files have been added. One lists "handhelds" (PDAs, | |
3057 | calculators, etc.), the other AV production devices (DJ tables, | |
3058 | keypads, etc.) that should accessible to the seat owner user by | |
3059 | default. | |
3060 | ||
3061 | * udevadm trigger gained a new --prioritized-subsystem= option to | |
3062 | process certain subsystems (and all their parent devices) earlier. | |
3063 | ||
3064 | systemd-udev-trigger.service now uses this new option to trigger | |
3065 | block and TPM devices first, hopefully making the boot a bit faster. | |
3066 | ||
3067 | * udevadm trigger now implements --type=all, --initialized-match, | |
3068 | --initialized-nomatch to trigger both subsystems and devices, only | |
3069 | already-initialized devices, and only devices which haven't been | |
3070 | initialized yet, respectively. | |
3071 | ||
61ade257 LP |
3072 | * udevadm gained a new "wait" command for safely waiting for a specific |
3073 | device to show up in the udev device database. This is useful in | |
3074 | scripts that asynchronously allocate a block device (e.g. through | |
3075 | repartitioning, or allocating a loopback device or similar) and need | |
3076 | to synchronize on the creation to complete. | |
3077 | ||
3078 | * udevadm gained a new "lock" command for locking one or more block | |
3079 | devices while formatting it or writing a partition table to it. It is | |
3080 | an implementation of https://systemd.io/BLOCK_DEVICE_LOCKING and | |
3081 | usable in scripts dealing with block devices. | |
3082 | ||
3083 | * udevadm info will show a couple of additional device fields in its | |
3084 | output, and will not apply a limited set of coloring to line types. | |
3085 | ||
3086 | * udevadm info --tree will now show a tree of objects (i.e. devices and | |
3087 | suchlike) in the /sys/ hierarchy. | |
3088 | ||
3089 | * Block devices will now get a new set of device symlinks in | |
3090 | /dev/disk/by-diskseq/<nr>, which may be used to reference block | |
3091 | device nodes via the kernel's "diskseq" value. Note that this does | |
3092 | not guarantee that opening a device by a symlink like this will | |
3093 | guarantee that the opened device actually matches the specified | |
3094 | diskseq value. To be safe against races, the actual diskseq value of | |
3095 | the opened device (BLKGETDISKSEQ ioctl()) must still be compred with | |
3096 | the one in the symlink path. | |
3097 | ||
0c6e746b | 3098 | * .link files gained support for setting MDI/MID-X on a link. |
60a777b5 | 3099 | |
00b29ca1 ZJS |
3100 | * .link files gained support for [Match] Firmware= setting to match on |
3101 | the device firmware description string. By mistake, it was previously | |
3102 | only supported in .network files. | |
3103 | ||
0c6e746b YW |
3104 | * .link files gained support for [Link] SR-IOVVirtualFunctions= setting |
3105 | and [SR-IOV] section to configure SR-IOV virtual functions. | |
3106 | ||
942473dc | 3107 | Changes in systemd-networkd: |
0c6e746b YW |
3108 | |
3109 | * The default scope for unicast routes configured through [Route] | |
3110 | section is changed to "link", to make the behavior consistent with | |
3111 | "ip route" command. The manual configuration of [Route] Scope= is | |
3112 | still honored. | |
3113 | ||
3114 | * A new unit systemd-networkd-wait-online@<interface>.service has been | |
3115 | added that can be used to wait for a specific network interface to be | |
3116 | up. | |
3117 | ||
3118 | * systemd-networkd gained a new [Bridge] Isolated=true|false setting | |
3119 | that configures the eponymous kernel attribute on the bridge. | |
3120 | ||
3121 | * .netdev files now can be used to create virtual WLAN devices, and | |
3122 | configure various settings on them, via the [WLAN] section. | |
3123 | ||
00b29ca1 ZJS |
3124 | * .link/.network files gained support for [Match] Kind= setting to match |
3125 | on device kind ("bond", "bridge", "gre", "tun", "veth", etc.) | |
3126 | ||
3127 | This value is also shown by 'networkctl status'. | |
3128 | ||
0c6e746b YW |
3129 | * The Local= setting in .netdev files for various virtual network |
3130 | devices gained support for specifying, in addition to the network | |
3131 | address, the name of a local interface which must have the specified | |
3132 | address. | |
60a777b5 | 3133 | |
0c6e746b YW |
3134 | * systemd-networkd gained a new [Tunnel] External= setting in .netdev |
3135 | files, to configure tunnels in external mode (a.k.a. collect metadata | |
3136 | mode). | |
3137 | ||
3138 | * [Network] L2TP= setting was removed. Please use interface specifier in | |
3139 | Local= setting in .netdev files of corresponding L2TP interface. | |
00b29ca1 ZJS |
3140 | |
3141 | * New [DHCPServer] BootServerName=, BootServerAddress=, and | |
3142 | BootFilename= settings can be used to configure the server address, | |
3143 | server name, and file name sent in the DHCP packet (e.g. to configure | |
3144 | PXE boot). | |
3145 | ||
942473dc | 3146 | Changes in systemd-resolved: |
00b29ca1 | 3147 | |
0c6e746b YW |
3148 | * systemd-resolved is started earlier (in sysinit.target), so it |
3149 | available earlier and will also be started in the initrd if installed | |
3150 | there. | |
e1f0c136 | 3151 | |
942473dc | 3152 | Changes in disk encryption: |
00b29ca1 | 3153 | |
0c6e746b YW |
3154 | * systemd-cryptenroll can now control whether to require the user to |
3155 | enter a PIN when using TPM-based unlocking of a volume via the new | |
3156 | --tpm2-with-pin= option. | |
60a777b5 | 3157 | |
0c6e746b | 3158 | Option tpm2-pin= can be used in /etc/crypttab. |
60a777b5 | 3159 | |
0c6e746b YW |
3160 | * When unlocking devices via TPM, TPM2 parameter encryption is now |
3161 | used, to ensure that communication between CPU and discrete TPM chips | |
3162 | cannot be eavesdropped to acquire disk encryption keys. | |
60a777b5 | 3163 | |
61ade257 LP |
3164 | * A new switch --fido2-credential-algorithm= has been added to |
3165 | systemd-cryptenroll allowing selection of the credential algorithm to | |
3166 | use when binding encryption to FIDO2 tokens. | |
3167 | ||
942473dc | 3168 | Changes in systemd-hostnamed: |
60a777b5 | 3169 | |
0c6e746b YW |
3170 | * HARDWARE_VENDOR= and HARDWARE_MODEL= can be set in /etc/machine-info |
3171 | to override the values gleaned from the hwdb. | |
3172 | ||
3173 | * A ID_CHASSIS property can be set in the hwdb (for the DMI device | |
3174 | /sys/class/dmi/id) to override the chassis that is reported by | |
3175 | hostnamed. | |
3176 | ||
3177 | * hostnamed's D-Bus interface gained a new method GetHardwareSerial() | |
61ade257 LP |
3178 | for reading the hardware serial number, as reportd by DMI. It also |
3179 | exposes a new method D-Bus property FirmwareVersion that encode the | |
3180 | firmware version of the system. | |
0c6e746b | 3181 | |
942473dc | 3182 | Changes in other components: |
0c6e746b YW |
3183 | |
3184 | * /etc/locale.conf is now populated through tmpfiles.d factory /etc/ | |
3185 | handling with the values that were configured during systemd build | |
3186 | (if /etc/locale.conf has not been created through some other | |
3187 | mechanism). This means that /etc/locale.conf should always have | |
3188 | reasonable contents and we avoid a potential mismatch in defaults. | |
3189 | ||
3190 | * The userdbctl tool will now show UID range information as part of the | |
3191 | list of known users. | |
60a777b5 | 3192 | |
942473dc ZJS |
3193 | * A new build-time configuration setting default-user-shell= can be |
3194 | used to set the default shell for user records and nspawn shell | |
30fd9a2d | 3195 | invocations (instead of the default /bin/bash). |
942473dc | 3196 | |
61ade257 LP |
3197 | * systemd-timesyncd now provides a D-Bus API for receiving NTP server |
3198 | information dynamically at runtime via IPC. | |
3199 | ||
3200 | * The systemd-creds tool gained a new "has-tpm2" verb, which reports | |
3201 | whether a functioning TPM2 infrastructure is available, i.e. if | |
3202 | firmware, kernel driver and systemd all have TPM2 support enabled and | |
3203 | a device found. | |
3204 | ||
3205 | * The systemd-creds tool gained support for generating encrypted | |
3206 | credentials that are using an empty encryption key. While this | |
3207 | provides no integrity nor confidentiality it's useful to implement | |
3208 | codeflows that work the same on TPM-ful and TPM2-less systems. The | |
3209 | service manager will only accept credentials "encrypted" that way if | |
3210 | a TPM2 device cannot be detected, to ensure that credentials | |
3211 | "encrypted" like that cannot be used to trick TPM2 systems. | |
3212 | ||
211b564a ZJS |
3213 | * When deciding whether to colorize output, all systemd programs now |
3214 | also check $COLORTERM (in addition to $NO_COLOR, $SYSTEMD_COLORS, and | |
3215 | $TERM). | |
3216 | ||
7f2ec323 LB |
3217 | * Meson's new install_tag feature is now in use for several components, |
3218 | allowing to build and install select binaries only: pam, nss, devel | |
3219 | (pkg-config files), systemd-boot, libsystemd, libudev. Example: | |
3220 | $ meson build systemd-boot | |
3221 | $ meson install --tags systemd-boot --no-rebuild | |
3222 | https://mesonbuild.com/Installing.html#installation-tags | |
3223 | ||
3224 | * A new build configuration option has been added, to allow selecting the | |
3225 | default compression algorithm used by systemd-journald and systemd-coredump. | |
3226 | This allows to build-in support for decompressing all supported formats, | |
3227 | but choose a specific one for compression. E.g.: | |
3228 | $ meson -Ddefault-compression=xz | |
3229 | ||
00b29ca1 ZJS |
3230 | Experimental features: |
3231 | ||
3232 | * sd-boot gained a new *experimental* setting "reboot-for-bitlocker" in | |
3233 | loader.conf that implements booting Microsoft Windows from the | |
3234 | sd-boot in a way that first reboots the system, to reset the TPM | |
3235 | PCRs. This improves compatibility with BitLocker's TPM use, as the | |
3236 | PCRs will only record the Windows boot process, and not sd-boot | |
3237 | itself, thus retaining the PCR measurements not involving sd-boot. | |
3238 | Note that this feature is experimental for now, and is likely going | |
3239 | to be generalized and renamed in a future release, without retaining | |
3240 | compatibility with the current implementation. | |
3241 | ||
3242 | * A new systemd-sysupdate component has been added that automatically | |
3243 | discovers, downloads, and installs A/B-style updates for the host | |
3244 | installation itself, or container images, portable service images, | |
3245 | and other assets. See the new systemd-sysupdate man page for updates. | |
ffa047a0 | 3246 | |
73849408 | 3247 | Contributions from: 4piu, Adam Williamson, adrian5, Albert Brox, |
0950eee5 LB |
3248 | AlexCatze, Alex Henrie, Alfonso Sánchez-Beato, Alice S, |
3249 | Alvin Šipraga, amarjargal, Amarjargal, Andrea Pappacoda, | |
3250 | Andreas Rammhold, Andy Chi, Anita Zhang, Antonio Alvarez Feijoo, | |
3251 | Arfrever Frehtes Taifersar Arahesis, ash, Bastien Nocera, Be, | |
410b996a LB |
3252 | bearhoney, Ben Efros, Benjamin Berg, Benjamin Franzke, |
3253 | Brett Holman, Christian Brauner, Clyde Byrd III, Curtis Klein, | |
3254 | Daan De Meyer, Daniele Medri, Daniel Mack, Danilo Krummrich, | |
3255 | David, David Bond, Davide Cavalca, David Tardon, davijosw, | |
3256 | dependabot[bot], Donald Chan, Dorian Clay, Eduard Tolosa, | |
3257 | Elias Probst, Eli Schwartz, Erik Sjölund, Evgeny Vereshchagin, | |
3258 | Federico Ceratto, Franck Bui, Frantisek Sumsal, Gaël PORTAY, | |
3259 | Georges Basile Stavracas Neto, Gibeom Gwon, Goffredo Baroncelli, | |
3260 | Grigori Goronzy, Hans de Goede, Heiko Becker, Hugo Carvalho, | |
3261 | Jakob Lell, James Hilliard, Jan Janssen, Jason A. Donenfeld, | |
3262 | Joan Bruguera, Joerie de Gram, Josh Triplett, Julia Kartseva, | |
3263 | Kazuo Moriwaka, Khem Raj, ksa678491784, Lance, Lan Tian, | |
3264 | Laura Barcziova, Lennart Poettering, Leviticoh, licunlong, | |
3265 | Lidong Zhong, lincoln auster, Lubomir Rintel, Luca Boccassi, | |
3266 | Luca BRUNO, lucagoc, Ludwig Nussel, Marcel Hellwig, march1993, | |
3267 | Marco Scardovi, Mario Limonciello, Mariusz Tkaczyk, | |
3268 | Markus Weippert, Martin, Martin Liska, Martin Wilck, Matija Skala, | |
0950eee5 LB |
3269 | Matthew Blythe, Matthias Lisin, Matthijs van Duin, Matt Walton, |
3270 | Max Gautier, Michael Biebl, Michael Olbrich, Michal Koutný, | |
3271 | Michal Sekletár, Mike Gilbert, MkfsSion, Morten Linderud, | |
410b996a LB |
3272 | Nick Rosbrook, Nikolai Grigoriev, Nikolai Kostrigin, |
3273 | Nishal Kulkarni, Noel Kuntze, Pablo Ceballos, Peter Hutterer, | |
3274 | Peter Morrow, Pigmy-penguin, Piotr Drąg, prumian, Richard Neill, | |
211b564a ZJS |
3275 | Rike-Benjamin Schuppner, rodin-ia, Romain Naour, Ruben Kerkhof, |
3276 | Ryan Hendrickson, Santa Wiryaman, Sebastian Pucilowski, Seth Falco, | |
0950eee5 LB |
3277 | Simon Ellmann, Sonali Srivastava, Stefan Seering, |
3278 | Stephen Hemminger, tawefogo, techtino, Temuri Doghonadze, | |
3279 | Thomas Batten, Thomas Haller, Thomas Weißschuh, Tobias Stoeckmann, | |
410b996a LB |
3280 | Tomasz Pala, Tyson Whitehead, Vishal Chillara Srinivas, |
3281 | Vivien Didelot, w30023233, wangyuhang, Weblate, Xiaotian Wu, | |
3282 | yangmingtai, YmrDtnJu, Yonathan Randolph, Yutsuten, Yu Watanabe, | |
73849408 ZJS |
3283 | Zbigniew Jędrzejewski-Szmek, наб |
3284 | ||
7f2ec323 | 3285 | — Edinburgh, 2022-05-21 |
bbfabc44 | 3286 | |
a420d717 | 3287 | CHANGES WITH 250: |
195d181c | 3288 | |
dcdc652f ZJS |
3289 | * Support for encrypted and authenticated credentials has been added. |
3290 | This extends the credential logic introduced with v247 to support | |
3291 | non-interactive symmetric encryption and authentication, based on a | |
3292 | key that is stored on the /var/ file system or in the TPM2 chip (if | |
3293 | available), or the combination of both (by default if a TPM2 chip | |
3294 | exists the combination is used, otherwise the /var/ key only). The | |
3295 | credentials are automatically decrypted at the moment a service is | |
3296 | started, and are made accessible to the service itself in unencrypted | |
3297 | form. A new tool 'systemd-creds' encrypts credentials for this | |
3298 | purpose, and two new service file settings LoadCredentialEncrypted= | |
3299 | and SetCredentialEncrypted= configure such credentials. | |
3300 | ||
3301 | This feature is useful to store sensitive material such as SSL | |
3302 | certificates, passwords and similar securely at rest and only decrypt | |
3303 | them when needed, and in a way that is tied to the local OS | |
3304 | installation or hardware. | |
195d181c LP |
3305 | |
3306 | * systemd-gpt-auto-generator can now automatically set up discoverable | |
3307 | LUKS2 encrypted swap partitions. | |
3308 | ||
dcdc652f ZJS |
3309 | * The GPT Discoverable Partitions Specification has been substantially |
3310 | extended with support for root and /usr/ partitions for the majority | |
3311 | of architectures systemd supports. This includes platforms that do | |
3312 | not natively support UEFI, because even though GPT is specified under | |
3313 | UEFI umbrella, it is useful on other systems too. Specifically, | |
195d181c | 3314 | systemd-nspawn, systemd-sysext, systemd-gpt-auto-generator and |
dcdc652f ZJS |
3315 | Portable Services use the concept without requiring UEFI. |
3316 | ||
3317 | * The GPT Discoverable Partitions Specifications has been extended with | |
3318 | a new set of partitions that may carry PKCS#7 signatures for Verity | |
3319 | partitions, encoded in a simple JSON format. This implements a simple | |
3320 | mechanism for building disk images that are fully authenticated and | |
3321 | can be tested against a set of cryptographic certificates. This is | |
3322 | now implemented for the various systemd tools that can operate with | |
3323 | disk images, such as systemd-nspawn, systemd-sysext, systemd-dissect, | |
3324 | Portable services/RootImage=, systemd-tmpfiles, and systemd-sysusers. | |
3325 | The PKCS#7 signatures are passed to the kernel (where they are | |
195d181c LP |
3326 | checked against certificates from the kernel keyring), or can be |
3327 | verified against certificates provided in userspace (via a simple | |
3328 | drop-in file mechanism). | |
3329 | ||
3330 | * systemd-dissect's inspection logic will now report for which uses a | |
3331 | disk image is intended. Specifically, it will display whether an | |
3332 | image is suitable for booting on UEFI or in a container (using | |
3333 | systemd-nspawn's --image= switch), whether it can be used as portable | |
3334 | service, or attached as system extension. | |
3335 | ||
3336 | * The system-extension.d/ drop-in files now support a new field | |
3337 | SYSEXT_SCOPE= that may encode which purpose a system extension image | |
3338 | is for: one of "initrd", "system" or "portable". This is useful to | |
3339 | make images more self-descriptive, and to ensure system extensions | |
3340 | cannot be attached in the wrong contexts. | |
3341 | ||
3342 | * The os-release file learnt a new PORTABLE_PREFIXES= field which may | |
3343 | be used in portable service images to indicate which unit prefixes | |
3344 | are supported. | |
3345 | ||
3346 | * The GPT image dissection logic in systemd-nspawn/systemd-dissect/… | |
3347 | now is able to decode images for non-native architectures as well. | |
dcdc652f ZJS |
3348 | This allows systemd-nspawn to boot images of non-native architectures |
3349 | if the corresponding user mode emulator is installed and | |
3350 | systemd-binfmtd is running. | |
195d181c | 3351 | |
dcdc652f | 3352 | * systemd-logind gained new settings HandlePowerKeyLongPress=, |
195d181c LP |
3353 | HandleRebootKeyLongPress=, HandleSuspendKeyLongPress= and |
3354 | HandleHibernateKeyLongPress= which may be used to configure actions | |
3355 | when the relevant keys are pressed for more than 5s. This is useful | |
3356 | on devices that only have hardware for a subset of these keys. By | |
3357 | default, if the reboot key is pressed long the poweroff operation is | |
3358 | now triggered, and when the suspend key is pressed long the hibernate | |
3359 | operation is triggered. Long pressing the other two keys currently | |
3360 | does not trigger any operation by default. | |
3361 | ||
3362 | * When showing unit status updates on the console during boot and | |
368910b1 | 3363 | shutdown, and a service is slow to start so that the cylon animation |
195d181c LP |
3364 | is shown, the most recent sd_notify() STATUS= text is now shown as |
3365 | well. Services may use this to make the boot/shutdown output easier | |
3366 | to understand, and to indicate what precisely a service that is slow | |
dcdc652f | 3367 | to start or stop is waiting for. In particular, the per-user service |
195d181c LP |
3368 | manager instance now reports what it is doing and which service it is |
3369 | waiting for this way to the system service manager. | |
3370 | ||
3371 | * The service manager will now re-execute on reception of the | |
3372 | SIGRTMIN+25 signal. It previously already did that on SIGTERM — but | |
3373 | only when running as PID 1. There was no signal to request this when | |
dcdc652f ZJS |
3374 | running as per-user service manager, i.e. as any other PID than 1. |
3375 | SIGRTMIN+25 works for both system and user managers. | |
195d181c LP |
3376 | |
3377 | * The hardware watchdog logic in PID 1 gained support for operating | |
3378 | with the default timeout configured in the hardware, instead of | |
3379 | insisting on re-configuring it. Set RuntimeWatchdogSec=default to | |
3380 | request this behavior. | |
3381 | ||
3382 | * A new kernel command line option systemd.watchdog_sec= is now | |
3383 | understood which may be used to override the hardware watchdog | |
3384 | time-out for the boot. | |
3385 | ||
3386 | * A new setting DefaultOOMScoreAdjust= is now supported in | |
942473dc ZJS |
3387 | /etc/systemd/system.conf and /etc/systemd/user.conf. It may be used |
3388 | to set the default process OOM score adjustment value for processes | |
3389 | started by the service manager. For per-user service managers this | |
195d181c LP |
3390 | now defaults to 100, but for per-system service managers is left as |
3391 | is. This means that by default now services forked off the user | |
3392 | service manager are more likely to be killed by the OOM killer than | |
3393 | system services or the managers themselves. | |
3394 | ||
3395 | * A new per-service setting RestrictFileSystems= as been added that | |
dcdc652f ZJS |
3396 | restricts the file systems a service has access to by their type. |
3397 | This is based on the new BPF LSM of the Linux kernel. It provides an | |
3398 | effective way to make certain API file systems unavailable to | |
195d181c LP |
3399 | services (and thus minimizing attack surface). A new command |
3400 | "systemd-analyze filesystems" has been added that lists all known | |
3401 | file system types (and how they are grouped together under useful | |
3402 | group handles). | |
3403 | ||
3404 | * Services now support a new setting RestrictNetworkInterfaces= for | |
3405 | restricting access to specific network interfaces. | |
3406 | ||
dcdc652f | 3407 | * Service unit files gained new settings StartupAllowedCPUs= and |
195d181c LP |
3408 | StartupAllowedMemoryNodes=. These are similar to their counterparts |
3409 | without the "Startup" prefix and apply during the boot process | |
3410 | only. This is useful to improve boot-time behavior of the system and | |
3411 | assign resources differently during boot than during regular | |
3412 | runtime. This is similar to the preexisting StartupCPUWeight= | |
3413 | vs. CPUWeight. | |
3414 | ||
3415 | * Related to this: the various StartupXYZ= settings | |
3416 | (i.e. StartupCPUWeight=, StartupAllowedCPUs=, …) are now also applied | |
3417 | during shutdown. The settings not prefixed with "Startup" hence apply | |
3418 | during regular runtime, and those that are prefixed like that apply | |
3419 | during boot and shutdown. | |
3420 | ||
bf71ade8 LB |
3421 | * A new per-unit set of conditions/asserts |
3422 | [Condition|Assert][Memory|CPU|IO]Pressure= have been added to make a | |
3423 | unit skip/fail activation if the system's (or a slice's) memory/cpu/io | |
3424 | pressure is above the configured threshold, using the kernel PSI | |
dcdc652f | 3425 | feature. For more details see systemd.unit(5) and |
0e685823 | 3426 | https://docs.kernel.org/accounting/psi.html |
bf71ade8 LB |
3427 | |
3428 | * The combination of ProcSubset=pid and ProtectKernelTunables=yes and/or | |
3429 | ProtectKernelLogs=yes can now be used. | |
3430 | ||
e63fa075 | 3431 | * The default maximum numbers of inodes have been raised from 64k to 1M |
bb7031bc | 3432 | for /dev/, and from 400k to 1M for /tmp/. |
bf71ade8 | 3433 | |
195d181c LP |
3434 | * The per-user service manager learnt support for communicating with |
3435 | systemd-oomd to acquire OOM kill information. | |
3436 | ||
3437 | * A new service setting ExecSearchPath= has been added that allows | |
dcdc652f ZJS |
3438 | changing the search path for executables for services. It affects |
3439 | where we look for the binaries specified in ExecStart= and similar, | |
3440 | and the specified directories are also added the $PATH environment | |
3441 | variable passed to invoked processes. | |
195d181c LP |
3442 | |
3443 | * A new setting RuntimeRandomizedExtraSec= has been added for service | |
3444 | and scope units that allows extending the runtime time-out as | |
3445 | configured by RuntimeMaxSec= with a randomized amount. | |
3446 | ||
3447 | * The syntax of the service unit settings RuntimeDirectory=, | |
3448 | StateDirectory=, CacheDirectory=, LogsDirectory= has been extended: | |
dcdc652f | 3449 | if the specified value is now suffixed with a colon, followed by |
195d181c LP |
3450 | another filename, the latter will be created as symbolic link to the |
3451 | specified directory. This allows creating these service directories | |
3452 | together with alias symlinks to make them available under multiple | |
3453 | names. | |
3454 | ||
3455 | * Service unit files gained two new settings TTYRows=/TTYColumns= for | |
3456 | configuring rows/columns of the TTY device passed to | |
3457 | stdin/stdout/stderr of the service. This is useful to propagate TTY | |
dcdc652f | 3458 | dimensions to a virtual machine. |
195d181c | 3459 | |
dcdc652f ZJS |
3460 | * A new service unit file setting ExitType= has been added that |
3461 | specifies when to assume a service has exited. By default systemd | |
3462 | only watches the main process of a service. By setting | |
3463 | ExitType=cgroup it can be told to wait for the last process in a | |
3464 | cgroup instead. | |
195d181c LP |
3465 | |
3466 | * Automount unit files gained a new setting ExtraOptions= that can be | |
3467 | used to configure additional mount options to pass to the kernel when | |
3468 | mounting the autofs instance. | |
3469 | ||
dcdc652f ZJS |
3470 | * "Urlification" (generation of ESC sequences that generate clickable |
3471 | hyperlinks in modern terminals) may now be turned off altogether | |
3472 | during build-time. | |
195d181c | 3473 | |
616779c3 | 3474 | * Path units gained new TriggerLimitBurst= and TriggerLimitIntervalSec= |
28be02e0 ZJS |
3475 | settings that default to 200 and 2 s respectively. The ratelimit |
3476 | ensures that a path unit cannot cause PID1 to busy-loop when it is | |
3477 | trying to trigger a service that is skipped because of a Condition*= | |
3478 | not being satisfied. This matches the configuration and behaviour of | |
3479 | socket units. | |
616779c3 | 3480 | |
dcdc652f ZJS |
3481 | * The TPM2/FIDO2/PKCS11 support in systemd-cryptsetup is now also built |
3482 | as a plug-in for cryptsetup. This means the plain cryptsetup command | |
3483 | may now be used to unlock volumes set up this way. | |
195d181c LP |
3484 | |
3485 | * The TPM2 logic in cryptsetup will now automatically detect systems | |
dcdc652f | 3486 | where the TPM2 chip advertises SHA256 PCR banks but the firmware only |
195d181c LP |
3487 | updates the SHA1 banks. In such a case PCR policies will be |
3488 | automatically bound to the latter, not the former. This makes the PCR | |
3489 | policies reliable, but of course do not provide the same level of | |
3490 | trust as SHA256 banks. | |
3491 | ||
3492 | * The TPM2 logic in systemd-cryptsetup/systemd-cryptsetup now supports | |
3493 | RSA primary keys in addition to ECC, improving compatibility with | |
3494 | TPM2 chips that do not support ECC. RSA keys are much slower to use | |
3495 | than ECC, and hence are only used if ECC is not available. | |
3496 | ||
3497 | * /etc/crypttab gained support for a new token-timeout= setting for | |
dcdc652f ZJS |
3498 | encrypted volumes that allows configuration of the maximum time to |
3499 | wait for PKCS#11/FIDO2 tokens to be plugged in. If the time elapses | |
3500 | the logic will query the user for a regular passphrase/recovery key | |
195d181c LP |
3501 | instead. |
3502 | ||
3503 | * Support for activating dm-integrity volumes at boot via a new file | |
dcdc652f ZJS |
3504 | /etc/integritytab and the tool systemd-integritysetup have been |
3505 | added. This is similar to /etc/crypttab and /etc/veritytab, but deals | |
3506 | with dm-integrity instead of dm-crypt/dm-verity. | |
195d181c LP |
3507 | |
3508 | * The systemd-veritysetup-generator now understands a new usrhash= | |
3509 | kernel command line option for specifying the Verity root hash for | |
3510 | the partition backing the /usr/ file system. A matching set of | |
3511 | systemd.verity_usr_* kernel command line options has been added as | |
3512 | well. These all work similar to the corresponding options for the | |
3513 | root partition. | |
3514 | ||
3515 | * The sd-device API gained a new API call sd_device_get_diskseq() to | |
3516 | return the DISKSEQ property of a device structure. The "disk | |
3517 | sequence" concept is a new feature recently introduced to the Linux | |
3518 | kernel that allows detecting reuse cycles of block devices, i.e. can | |
3519 | be used to recognize when loopback block devices are reused for a | |
3520 | different purpose or CD-ROM drives get their media changed. | |
3521 | ||
3522 | * A new unit systemd-boot-update.service has been added. If enabled | |
3523 | (the default) and the sd-boot loader is detected to be installed, it | |
dcdc652f ZJS |
3524 | is automatically updated to the newest version when out of date. This |
3525 | is useful to ensure the boot loader remains up-to-date, and updates | |
3526 | automatically propagate from the OS tree in /usr/. | |
195d181c | 3527 | |
bf71ade8 LB |
3528 | * sd-boot will now build with SBAT by default in order to facilitate |
3529 | working with recent versions of Shim that require it to be present. | |
3530 | ||
97b6ed32 ZJS |
3531 | * sd-boot can now parse Microsoft Windows' Boot Configuration Data. |
3532 | This is used to robustly generate boot entry titles for Windows. | |
3533 | ||
195d181c LP |
3534 | * A new generic target unit factory-reset.target has been added. It is |
3535 | hooked into systemd-logind similar in fashion to | |
3536 | reboot/poweroff/suspend/hibernate, and is supposed to be used to | |
3537 | initiate a factory reset operation. What precisely this operation | |
3538 | entails is up for the implementer to decide, the primary goal of the | |
3539 | new unit is provide a framework where to plug in the implementation | |
3540 | and how to trigger it. | |
3541 | ||
3542 | * A new meson build-time option 'clock-valid-range-usec-max' has been | |
3543 | added which takes a time in µs and defaults to 15 years. If the RTC | |
3544 | time is noticed to be more than the specified time ahead of the | |
3545 | built-in epoch of systemd (which by default is the release timestamp | |
3546 | of systemd) it is assumed that the RTC is not working correctly, and | |
3547 | the RTC is reset to the epoch. (It already is reset to the epoch when | |
3548 | noticed to be before it.) This should increase the chance that time | |
3549 | doesn't accidentally jump too far ahead due to faulty hardware or | |
3550 | batteries. | |
3551 | ||
63b7d347 YW |
3552 | * A new setting SaveIntervalSec= has been added to systemd-timesyncd, |
3553 | which may be used to automatically save the current system time to | |
3554 | disk in regular intervals. This is useful to maintain a roughly | |
3555 | monotonic clock even without RTC hardware and with some robustness | |
3556 | against abnormal system shutdown. | |
3557 | ||
195d181c LP |
3558 | * systemd-analyze verify gained support for a pair of new --image= + |
3559 | --root= switches for verifying units below a specific root | |
3560 | directory/image instead of on the host. | |
3561 | ||
3562 | * systemd-analyze verify gained support for verifying unit files under | |
3563 | an explicitly specified unit name, independently of what the filename | |
3564 | actually is. | |
3565 | ||
dcdc652f ZJS |
3566 | * systemd-analyze verify gained a new switch --recursive-errors= which |
3567 | controls whether to only fail on errors found in the specified units | |
3568 | or recursively any dependent units. | |
3569 | ||
3570 | * systemd-analyze security now supports a new --offline mode for | |
3571 | analyzing unit files stored on disk instead of loaded units. It may | |
3572 | be combined with --root=/--image to analyze unit files under a root | |
3573 | directory or disk image. It also learnt a new --threshold= parameter | |
3574 | for specifying an exposure level threshold: if the exposure level | |
3575 | exceeds the specified value the call will fail. It also gained a new | |
3576 | --security-policy= switch for configuring security policies to | |
3577 | enforce on the units. A policy is a JSON file that lists which tests | |
3578 | shall be weighted how much to determine the overall exposure | |
3579 | level. Altogether these new features are useful for fully automatic | |
3580 | analysis and enforcement of security policies on unit files. | |
3581 | ||
3582 | * systemd-analyze security gain a new --json= switch for JSON output. | |
3583 | ||
3584 | * systemd-analyze learnt a new --quiet switch for reducing | |
3585 | non-essential output. It's honored by the "dot", "syscall-filter", | |
3586 | "filesystems" commands. | |
3587 | ||
bb7031bc | 3588 | * systemd-analyze security gained a --profile= option that can be used |
dcdc652f ZJS |
3589 | to take into account a portable profile when analyzing portable |
3590 | services, since a lot of the security-related settings are enabled | |
3591 | through them. | |
3592 | ||
3593 | * systemd-analyze learnt a new inspect-elf verb that parses ELF core | |
3594 | files, binaries and executables and prints metadata information, | |
3595 | including the build-id and other info described on: | |
3596 | https://systemd.io/COREDUMP_PACKAGE_METADATA/ | |
3597 | ||
c0b28d44 YW |
3598 | * .network files gained a new UplinkInterface= in the [IPv6SendRA] |
3599 | section, for automatically propagating DNS settings from other | |
3600 | interfaces. | |
bf71ade8 | 3601 | |
c0b28d44 YW |
3602 | * The static lease DHCP server logic in systemd-networkd may now serve |
3603 | IP addresses outside of the configured IP pool range for the server. | |
97b6ed32 | 3604 | |
c0b28d44 YW |
3605 | * CAN support in systemd-networkd gained four new settings Loopback=, |
3606 | OneShot=, PresumeAck=, ClassicDataLengthCode= for tweaking CAN | |
3607 | control modes. It gained a number of further settings for tweaking | |
3608 | CAN timing quanta. | |
bf71ade8 | 3609 | |
c0b28d44 YW |
3610 | * The [CAN] section in .network file gained new TimeQuantaNSec=, |
3611 | PropagationSegment=, PhaseBufferSegment1=, PhaseBufferSegment2=, | |
3612 | SyncJumpWidth=, DataTimeQuantaNSec=, DataPropagationSegment=, | |
3613 | DataPhaseBufferSegment1=, DataPhaseBufferSegment2=, and | |
3614 | DataSyncJumpWidth= settings to control bit-timing processed by the | |
3615 | CAN interface. | |
bf71ade8 | 3616 | |
c0b28d44 YW |
3617 | * DHCPv4 client support in systemd-networkd learnt a new Label= option |
3618 | for configuring the address label to apply to configure IPv4 | |
3619 | addresses. | |
bf71ade8 | 3620 | |
c0b28d44 YW |
3621 | * The [IPv6AcceptRA] section of .network files gained support for a new |
3622 | UseMTU= setting that may be used to control whether to apply the | |
3623 | announced MTU settings to the local interface. | |
195d181c | 3624 | |
63b7d347 YW |
3625 | * The [DHCPv4] section in .network file gained a new Use6RD= boolean |
3626 | setting to control whether the DHCPv4 client request and process the | |
3627 | DHCP 6RD option. | |
3628 | ||
3629 | * The [DHCPv6PrefixDelegation] section in .network file is renamed to | |
3630 | [DHCPPrefixDelegation], as now the prefix delegation is also supported | |
3631 | with DHCPv4 protocol by enabling the Use6RD= setting. | |
3632 | ||
3633 | * The [DHCPPrefixDelegation] section in .network file gained a new | |
3634 | setting UplinkInterface= to specify the upstream interface. | |
3635 | ||
3636 | * The [DHCPv6] section in .network file gained a new setting | |
3637 | UseDelegatedPrefix= to control whether the delegated prefixes will be | |
3638 | propagated to the downstream interfaces. | |
195d181c LP |
3639 | |
3640 | * The [IPv6AcceptRA] section of .network files now understands two new | |
3641 | settings UseGateway=/UseRoutePrefix= for explicitly configuring | |
3642 | whether to use the relevant fields from the IPv6 Router Advertisement | |
3643 | records. | |
3644 | ||
8adba772 ZJS |
3645 | * The ForceDHCPv6PDOtherInformation= setting in the [DHCPv6] section |
3646 | has been removed. Please use the WithoutRA= and UseDelegatedPrefix= | |
63b7d347 YW |
3647 | settings in the [DHCPv6] section and the DHCPv6Client= setting in the |
3648 | [IPv6AcceptRA] section to control when the DHCPv6 client is started | |
3649 | and how the delegated prefixes are handled by the DHCPv6 client. | |
3650 | ||
63b7d347 YW |
3651 | * The IPv6Token= section in the [Network] section is deprecated, and |
3652 | the [IPv6AcceptRA] section gained the Token= setting for its | |
3653 | replacement. The [IPv6Prefix] section also gained the Token= setting. | |
3654 | The Token= setting gained 'eui64' mode to explicitly configure an | |
3655 | address with the EUI64 algorithm based on the interface MAC address. | |
3656 | The 'prefixstable' mode can now optionally take a secret key. The | |
3657 | Token= setting in the [DHCPPrefixDelegation] section now supports all | |
3658 | algorithms supported by the same settings in the other sections. | |
3659 | ||
3660 | * The [RoutingPolicyRule] section of .network file gained a new | |
3661 | SuppressInterfaceGroup= setting. | |
195d181c LP |
3662 | |
3663 | * The IgnoreCarrierLoss= setting in the [Network] section of .network | |
dcdc652f ZJS |
3664 | files now allows a duration to be specified, controlling how long to |
3665 | wait before reacting to carrier loss. | |
195d181c | 3666 | |
dcdc652f ZJS |
3667 | * The [DHCPServer] section of .network file gained a new Router= |
3668 | setting to specify the router address. | |
11b10922 | 3669 | |
c0b28d44 YW |
3670 | * The [CAKE] section of .network files gained various new settings |
3671 | AutoRateIngress=, CompensationMode=, FlowIsolationMode=, NAT=, | |
3672 | MPUBytes=, PriorityQueueingPreset=, FirewallMark=, Wash=, SplitGSO=, | |
3673 | and UseRawPacketSize= for configuring CAKE. | |
3674 | ||
3675 | * systemd-networkd now ships with new default .network files: | |
3676 | 80-container-vb.network which matches host-side network bridge device | |
3677 | created by systemd-nspawn's --network-bridge or --network-zone | |
3678 | switch, and 80-6rd-tunnel.network which matches automatically created | |
3679 | sit tunnel with 6rd prefix when the DHCP 6RD option is received. | |
3680 | ||
3681 | * systemd-networkd's handling of Endpoint= resolution for WireGuard | |
3682 | interfaces has been improved. | |
3683 | ||
3684 | * systemd-networkd will now automatically configure routes to addresses | |
3685 | specified in AllowedIPs=. This feature can be controlled via | |
3686 | RouteTable= and RouteMetric= settings in [WireGuard] or | |
3687 | [WireGuardPeer] sections. | |
3688 | ||
3689 | * systemd-networkd will now once again automatically generate persistent | |
3690 | MAC addresses for batadv and bridge interfaces. Users can disable this | |
3691 | by using MACAddress=none in .netdev files. | |
3692 | ||
3693 | * systemd-networkd and systemd-udevd now support IP over InfiniBand | |
3694 | interfaces. The Kind= setting in .netdev file accepts "ipoib". And | |
3695 | systemd.netdev files gained the [IPoIB] section. | |
3696 | ||
3697 | * systemd-networkd and systemd-udevd now support net.ifname-policy= | |
3698 | option on the kernel command-line. This is implemented through the | |
3699 | systemd-network-generator service that automatically generates | |
3700 | appropriate .link, .network, and .netdev files. | |
3701 | ||
3702 | * The various systemd-udevd "ethtool" buffer settings now understand | |
3703 | the special value "max" to configure the buffers to the maximum the | |
3704 | hardware supports. | |
3705 | ||
3706 | * systemd-udevd's .link files may now configure a large variety of | |
3707 | NIC coalescing settings, plus more hardware offload settings. | |
3708 | ||
3709 | * .link files gained a new WakeOnLanPassword= setting in the [Link] | |
3710 | section that allows to specify a WoL "SecureOn" password on hardware | |
3711 | that supports this. | |
3712 | ||
195d181c LP |
3713 | * systemd-nspawn's --setenv= switch now supports an additional syntax: |
3714 | if only a variable name is specified (i.e. without being suffixed by | |
3715 | a '=' character and a value) the current value of the environment | |
3716 | variable is propagated to the container. e.g. --setenv=FOO will | |
3717 | lookup the current value of $FOO in the environment, and pass it down | |
3718 | to the container. Similar behavior has been added to homectl's, | |
3719 | machinectl's and systemd-run's --setenv= switch. | |
3720 | ||
3721 | * systemd-nspawn gained a new switch --suppress-sync= which may be used | |
3722 | to optionally suppress the effect of the sync()/fsync()/fdatasync() | |
3723 | system calls for the container payload. This is useful for build | |
3724 | system environments where safety against abnormal system shutdown is | |
3725 | not essential as all build artifacts can be regenerated any time, but | |
3726 | the performance win is beneficial. | |
3727 | ||
dcdc652f ZJS |
3728 | * systemd-nspawn will now raise the RLIMIT_NOFILE hard limit to the |
3729 | same value that PID 1 uses for most forked off processes. | |
195d181c LP |
3730 | |
3731 | * systemd-nspawn's --bind=/--bind-ro= switches now optionally take | |
3732 | uidmap/nouidmap options as last parameter. If "uidmap" is used the | |
3733 | bind mounts are created with UID mapping taking place that ensures | |
3734 | the host's file ownerships are mapped 1:1 to container file | |
3735 | ownerships, even if user namespacing is used. This way | |
3736 | files/directories bound into containers will no longer show up as | |
dcdc652f | 3737 | owned by the nobody user as they typically did if no special care was |
195d181c LP |
3738 | taken to shift them manually. |
3739 | ||
3740 | * When discovering Windows installations sd-boot will now attempt to | |
dcdc652f | 3741 | show the Windows version. |
195d181c LP |
3742 | |
3743 | * The color scheme to use in sd-boot may now be configured at | |
3744 | build-time. | |
3745 | ||
195d181c LP |
3746 | * sd-boot gained the ability to change screen resolution during |
3747 | boot-time, by hitting the "r" key. This will cycle through available | |
dcdc652f | 3748 | resolutions and save the last selection. |
195d181c LP |
3749 | |
3750 | * sd-boot learnt a new hotkey "f". When pressed the system will enter | |
3751 | firmware setup. This is useful in environments where it is difficult | |
3752 | to hit the right keys early enough to enter the firmware, and works | |
3753 | on any firmware regardless which key it natively uses. | |
3754 | ||
3755 | * sd-boot gained support for automatically booting into the menu item | |
3756 | selected on the last boot (using the "@saved" identifier for menu | |
3757 | items). | |
3758 | ||
dcdc652f ZJS |
3759 | * sd-boot gained support for automatically loading all EFI drivers |
3760 | placed in the /EFI/systemd/drivers/ subdirectory of the EFI System | |
3761 | Partition (ESP). These drivers are loaded before the menu entries are | |
3762 | loaded. This is useful e.g. to load additional file system drivers | |
3763 | for the XBOOTLDR partition. | |
3764 | ||
3765 | * systemd-boot will now paint the input cursor on its own instead of | |
3766 | relying on the firmware to do so, increasing compatibility with broken | |
3767 | firmware that doesn't make the cursor reasonably visible. | |
3768 | ||
195d181c LP |
3769 | * sd-boot now embeds a .osrel PE section like we expect from Boot |
3770 | Loader Specification Type #2 Unified Kernels. This means sd-boot | |
3771 | itself may be used in place of a Type #2 Unified Kernel. This is | |
3772 | useful for debugging purposes as it allows chain-loading one a | |
3773 | (development) sd-boot instance from another. | |
3774 | ||
3775 | * sd-boot now supports a new "devicetree" field in Boot Loader | |
3776 | Specification Type #1 entries: if configured the specified device | |
3777 | tree file is installed before the kernel is invoked. This is useful | |
3778 | for installing/applying new devicetree files without updating the | |
3779 | kernel image. | |
3780 | ||
dcdc652f | 3781 | * Similarly, sd-stub now can read devicetree data from a PE section |
195d181c LP |
3782 | ".dtb" and apply it before invoking the kernel. |
3783 | ||
3784 | * sd-stub (the EFI stub that can be glued in front of a Linux kernel) | |
97b6ed32 ZJS |
3785 | gained the ability to pick up credentials and sysext files, wrap them |
3786 | in a cpio archive, and pass as an additional initrd to the invoked | |
3787 | Linux kernel, in effect placing those files in the /.extra/ directory | |
3788 | of the initrd environment. This is useful to implement trusted initrd | |
3789 | environments which are fully authenticated but still can be extended | |
3790 | (via sysexts) and parameterized (via encrypted/authenticated | |
3791 | credentials, see above). | |
3792 | ||
3793 | Credentials can be located next to the kernel image file (credentials | |
3794 | specific to a single boot entry), or in one of the shared directories | |
3795 | (credentials applicable to multiple boot entries). | |
195d181c LP |
3796 | |
3797 | * sd-stub now comes with a full man page, that explains its feature set | |
3798 | and how to combine a kernel image, an initrd and the stub to build a | |
3799 | complete EFI unified kernel image, implementing Boot Loader | |
3800 | Specification Type #2. | |
3801 | ||
dcdc652f | 3802 | * sd-stub may now provide the initrd to the executed kernel via the |
195d181c LP |
3803 | LINUX_EFI_INITRD_MEDIA_GUID EFI protocol, adding compatibility for |
3804 | non-x86 architectures. | |
3805 | ||
dcdc652f ZJS |
3806 | * bootctl learnt new set-timeout and set-timeout-oneshot commands that |
3807 | may be used to set the boot menu time-out of the boot loader (for all | |
3808 | or just the subsequent boot). | |
195d181c | 3809 | |
ffc97a1c LP |
3810 | * bootctl and kernel-install will now read variables |
3811 | KERNEL_INSTALL_LAYOUT= from /etc/machine-info and layout= from | |
3812 | /etc/kernel/install.conf. When set, it specifies the layout to use | |
3813 | for installation directories on the boot partition, so that tools | |
3814 | don't need to guess it based on the already-existing directories. The | |
3815 | only value that is defined natively is "bls", corresponding to the | |
3816 | layout specified in | |
c20ecc94 ZJS |
3817 | https://systemd.io/BOOT_LOADER_SPECIFICATION/. Plugins for |
3818 | kernel-install that implement a different layout can declare other | |
3819 | values for this variable. | |
3820 | ||
3821 | 'bootctl install' will now write KERNEL_INSTALL_LAYOUT=bls, on the | |
3822 | assumption that if the user installed sd-boot to the ESP, they intend | |
3823 | to use the entry layout understood by sd-boot. It'll also write | |
3824 | KERNEL_INSTALL_MACHINE_ID= if it creates any directories using the ID | |
3825 | (and it wasn't specified in the config file yet). Similarly, | |
3826 | kernel-install will now write KERNEL_INSTALL_MACHINE_ID= (if it | |
3827 | wasn't specified in the config file yet). Effectively, those changes | |
3828 | mean that the machine-id used for boot loader entry installation is | |
3829 | "frozen" upon first use and becomes independent of the actual | |
3830 | machine-id. | |
3831 | ||
3832 | Configuring KERNEL_INSTALL_MACHINE_ID fixes the following problem: | |
3833 | images created for distribution ("golden images") are built with no | |
3834 | machine-id, so that a unique machine-id can be created on the first | |
3835 | boot. But those images may contain boot loader entries with the | |
3836 | machine-id used during build included in paths. Using a "frozen" | |
3837 | value allows unambiguously identifying entries that match the | |
3838 | specific installation, while still permitting parallel installations | |
3839 | without conflict. | |
3840 | ||
3841 | Configuring KERNEL_INSTALL_LAYOUT obviates the need for | |
3842 | kernel-install to guess the installation layout. This fixes the | |
3843 | problem where a (possibly empty) directory in the boot partition is | |
3844 | created from a different layout causing kernel-install plugins to | |
3845 | assume the wrong layout. A particular example of how this may happen | |
3846 | is the grub2 package in Fedora which includes directories under /boot | |
3847 | directly in its file list. Various other packages pull in grub2 as a | |
3848 | dependency, so it may be installed even if unused, breaking | |
3849 | installations that use the bls layout. | |
3850 | ||
97b6ed32 ZJS |
3851 | * bootctl and systemd-bless-boot can now be linked statically. |
3852 | ||
195d181c | 3853 | * systemd-sysext now optionally doesn't insist on extension-release.d/ |
dcdc652f | 3854 | files being placed in the image under the image's file name. If the |
195d181c | 3855 | file system xattr user.extension-release.strict is set on the |
dcdc652f | 3856 | extension release file, it is accepted regardless of its name. This |
195d181c LP |
3857 | relaxes security restrictions a bit, as system extension may be |
3858 | attached under a wrong name this way. | |
3859 | ||
3860 | * udevadm's test-builtin command learnt a new --action= switch for | |
3861 | testing the built-in with the specified action (in place of the | |
dcdc652f | 3862 | default 'add'). |
195d181c LP |
3863 | |
3864 | * udevadm info gained new switches --property=/--value for showing only | |
3865 | specific udev properties/values instead of all. | |
3866 | ||
3867 | * A new hwdb database has been added that contains matches for various | |
3868 | types of signal analyzers (protocol analyzers, logic analyzers, | |
3869 | oscilloscopes, multimeters, bench power supplies, etc.) that should | |
3870 | be accessible to regular users. | |
3871 | ||
3872 | * A new hwdb database entry has been added that carries information | |
dcdc652f ZJS |
3873 | about types of cameras (regular or infrared), and in which direction |
3874 | they point (front or back). | |
195d181c | 3875 | |
11b10922 LB |
3876 | * A new rule to allow console users access to rfkill by default has been |
3877 | added to hwdb. | |
3878 | ||
97b6ed32 ZJS |
3879 | * Device nodes for the Software Guard eXtension enclaves (sgx_vepc) are |
3880 | now also owned by the system group "sgx". | |
3881 | ||
195d181c | 3882 | * A new build-time meson option "extra-net-naming-schemes=" has been |
30fd9a2d | 3883 | added to define additional naming schemes for udev's network |
dcdc652f ZJS |
3884 | interface naming logic. This is useful for enterprise distributions |
3885 | and similar which want to pin the schemes of certain distribution | |
3886 | releases under a specific name and previously had to patch the | |
3887 | sources to introduce new named schemes. | |
195d181c LP |
3888 | |
3889 | * The predictable naming logic for network interfaces has been extended | |
3890 | to generate stable names from Xen netfront device information. | |
3891 | ||
3892 | * hostnamed's chassis property can now be sourced from chassis-type | |
dcdc652f | 3893 | field encoded in devicetree (in addition to the existing DMI |
195d181c LP |
3894 | support). |
3895 | ||
dcdc652f | 3896 | * systemd-cgls now optionally displays cgroup IDs and extended |
195d181c LP |
3897 | attributes for each cgroup. (Controllable via the new --xattr= + |
3898 | --cgroup-id= switches.) | |
3899 | ||
3900 | * coredumpctl gained a new --all switch for operating on all | |
3901 | Journal files instead of just the local ones. | |
3902 | ||
bf71ade8 LB |
3903 | * systemd-coredump will now use libdw/libelf via dlopen() rather than |
3904 | directly linking, allowing users to easily opt-out of backtrace/metadata | |
3905 | analysis of core files, and reduce image sizes when this is not needed. | |
3906 | ||
3907 | * systemd-coredump will now analyze core files with libdw/libelf in a | |
3908 | forked, sandboxed process. | |
3909 | ||
195d181c LP |
3910 | * systemd-homed will now try to unmount an activate home area in |
3911 | regular intervals once the user logged out fully. Previously this was | |
3912 | attempted exactly once but if the home directory was busy for some | |
3913 | reason it was not tried again. | |
3914 | ||
dcdc652f | 3915 | * systemd-homed's LUKS2 home area backend will now create a BSD file |
195d181c | 3916 | system lock on the image file while the home area is active |
dcdc652f | 3917 | (i.e. mounted). If a home area is found to be locked, logins are |
195d181c LP |
3918 | politely refused. This should improve behavior when using home areas |
3919 | images that are accessible via the network from multiple clients, and | |
3920 | reduce the chance of accidental file system corruption in that case. | |
3921 | ||
3922 | * Optionally, systemd-homed will now drop the kernel buffer cache once | |
dcdc652f | 3923 | a user has fully logged out, configurable via the new --drop-caches= |
195d181c LP |
3924 | homectl switch. |
3925 | ||
dcdc652f ZJS |
3926 | * systemd-homed now makes use of UID mapped mounts for the home areas. |
3927 | If the kernel and used file system support it, files are now | |
195d181c LP |
3928 | internally owned by the "nobody" user (i.e. the user typically used |
3929 | for indicating "this ownership is not mapped"), and dynamically | |
3930 | mapped to the UID used locally on the system via the UID mapping | |
3931 | mount logic of recent kernels. This makes migrating home areas | |
dcdc652f ZJS |
3932 | between different systems cheaper because recursively chown()ing file |
3933 | system trees is no longer necessary. | |
195d181c LP |
3934 | |
3935 | * systemd-homed's CIFS backend now optionally supports CIFS service | |
3936 | names with a directory suffix, in order to place home directories in | |
3937 | a subdirectory of a CIFS share, instead of the top-level directory. | |
3938 | ||
3939 | * systemd-homed's CIFS backend gained support for specifying additional | |
3940 | mount options in the JSON user record (cifsExtraMountOptions field, | |
3941 | and --cifs-extra-mount-options= homectl switch). This is for example | |
3942 | useful for configuring mount options such as "noserverino" that some | |
dcdc652f ZJS |
3943 | SMB3 services require (use that to run a homed home directory from a |
3944 | FritzBox SMB3 share this way). | |
195d181c LP |
3945 | |
3946 | * systemd-homed will now default to btrfs' zstd compression for home | |
dcdc652f ZJS |
3947 | areas. This is inspired by Fedora's recent decision to switch to zstd |
3948 | by default. | |
195d181c LP |
3949 | |
3950 | * Additional mount options to use when mounting the file system of | |
3951 | LUKS2 volumes in systemd-homed has been added. Via the | |
3952 | $SYSTEMD_HOME_MOUNT_OPTIONS_BTRFS, $SYSTEMD_HOME_MOUNT_OPTIONS_EXT4, | |
3953 | $SYSTEMD_HOME_MOUNT_OPTIONS_XFS environment variables to | |
3954 | systemd-homed or via the luksExtraMountOptions user record JSON | |
3955 | property. (Exposed via homectl --luks-extra-mount-options) | |
3956 | ||
3957 | * homectl's resize command now takes the special size specifications | |
3958 | "min" and "max" to shrink/grow the home area to the minimum/maximum | |
3959 | size possible, taking disk usage/space constraints and file system | |
3960 | limitations into account. Resizing is now generally graceful: the | |
3961 | logic will try to get as close to the specified size as possible, but | |
3962 | not consider it a failure if the request couldn't be fulfilled | |
3963 | precisely. | |
3964 | ||
3965 | * systemd-homed gained the ability to automatically shrink home areas | |
3966 | on logout to their minimal size and grow them again on next | |
dcdc652f ZJS |
3967 | login. This ensures that while inactive, a home area only takes up |
3968 | the minimal space necessary, but once activated, it provides | |
3969 | sufficient space for the user's needs. This behavior is only | |
3970 | supported if btrfs is used as file system inside the home area | |
3971 | (because only for btrfs online growing/shrinking is implemented in | |
3972 | the kernel). This behavior is now enabled by default, but may be | |
3973 | controlled via the new --auto-resize-mode= setting of homectl. | |
195d181c LP |
3974 | |
3975 | * systemd-homed gained support for automatically re-balancing free disk | |
3976 | space among active home areas, in case the LUKS2 backends are used, | |
3977 | and no explicit disk size was requested. This way disk space is | |
3978 | automatically managed and home areas resized in regular intervals and | |
3979 | manual resizing when disk space becomes scarce should not be | |
3980 | necessary anymore. This behavior is only supported if btrfs is used | |
3981 | within the home areas (as only then online shrinking and growing is | |
3982 | supported), and may be configured via the new rebalanceWeight JSON | |
3983 | user record field (as exposed via the new --rebalance-weight= homectl | |
3984 | setting). Re-balancing is mostly automatic, but can also be requested | |
3985 | explicitly via "homectl rebalance", which is synchronous, and thus | |
dcdc652f | 3986 | may be used to wait until the rebalance run is complete. |
195d181c LP |
3987 | |
3988 | * userdbctl gained a --json= switch for configured the JSON formatting | |
3989 | to use when outputting user or group records. | |
3990 | ||
3991 | * userdbctl gained a new --multiplexer= switch for explicitly | |
3992 | configuring whether to use the systemd-userdbd server side user | |
3993 | record resolution logic. | |
3994 | ||
3995 | * userdbctl's ssh-authorized-keys command learnt a new --chain switch, | |
3996 | for chaining up another command to execute after completing the | |
3997 | look-up. Since the OpenSSH's AuthorizedKeysCommand only allows | |
dcdc652f ZJS |
3998 | configuration of a single command to invoke, this maybe used to |
3999 | invoke multiple: first userdbctl's own implementation, and then any | |
4000 | other also configured in the command line. | |
195d181c LP |
4001 | |
4002 | * The sd-event API gained a new function sd_event_add_inotify_fd() that | |
4003 | is similar to sd_event_add_inotify() but accepts a file descriptor | |
4004 | instead of a path in the file system for referencing the inode to | |
4005 | watch. | |
4006 | ||
4007 | * The sd-event API gained a new function | |
4008 | sd_event_source_set_ratelimit_expire_callback() that may be used to | |
4009 | define a callback function that is called whenever an event source | |
4010 | leaves the rate limiting phase. | |
4011 | ||
4012 | * New documentation has been added explaining which steps are necessary | |
4013 | to port systemd to a new architecture: | |
4014 | ||
4015 | https://systemd.io/PORTING_TO_NEW_ARCHITECTURES | |
4016 | ||
4017 | * The x-systemd.makefs option in /etc/fstab now explicitly supports | |
dcdc652f | 4018 | ext2, ext3, and f2fs file systems. |
195d181c | 4019 | |
97b6ed32 ZJS |
4020 | * Mount units and units generated from /etc/fstab entries with 'noauto' |
4021 | are now ordered the same as other units. Effectively, they will be | |
4022 | started earlier (if something actually pulled them in) and stopped | |
4023 | later, similarly to normal mount units that are part of | |
b0b1edc2 | 4024 | fs-local.target. This change should be invisible to users, but |
97b6ed32 ZJS |
4025 | should prevent those units from being stopped too early during |
4026 | shutdown. | |
4027 | ||
195d181c LP |
4028 | * The systemd-getty-generator now honors a new kernel command line |
4029 | argument systemd.getty_auto= and a new environment variable | |
4030 | $SYSTEMD_GETTY_AUTO that allows turning it off at boot. This is for | |
dcdc652f | 4031 | example useful to turn off gettys inside of containers or similar |
195d181c LP |
4032 | environments. |
4033 | ||
4034 | * systemd-resolved now listens on a second DNS stub address: 127.0.0.54 | |
4035 | (in addition to 127.0.0.53, as before). If DNS requests are sent to | |
4036 | this address they are propagated in "bypass" mode only, i.e. are | |
4037 | almost not processed locally, but mostly forwarded as-is to the | |
4038 | current upstream DNS servers. This provides a stable DNS server | |
4039 | address that proxies all requests dynamically to the right upstream | |
4040 | DNS servers even if these dynamically change. This stub does not do | |
4041 | mDNS/LLMNR resolution. However, it will translate look-ups to | |
4042 | DNS-over-TLS if necessary. This new stub is particularly useful in | |
4043 | container/VM environments, or for tethering setups: use DNAT to | |
4044 | redirect traffic to any IP address to this stub. | |
4045 | ||
bf71ade8 LB |
4046 | * systemd-importd now honors new environment variables |
4047 | $SYSTEMD_IMPORT_BTRFS_SUBVOL, $SYSTEMD_IMPORT_BTRFS_QUOTA, | |
4048 | $SYSTEMD_IMPORT_SYNC, which may be used disable btrfs subvolume | |
4049 | generation, btrfs quota setup and disk synchronization. | |
4050 | ||
4051 | * systemd-importd and systemd-resolved can now be optionally built with | |
4052 | OpenSSL instead of libgcrypt. | |
4053 | ||
4054 | * systemd-repart no longer requires OpenSSL. | |
4055 | ||
dcdc652f ZJS |
4056 | * systemd-sysusers will no longer create the redundant 'nobody' group |
4057 | by default, as the 'nobody' user is already created with an | |
4058 | appropriate primary group. | |
bf71ade8 LB |
4059 | |
4060 | * If a unit uses RuntimeMaxSec, systemctl show will now display it. | |
4061 | ||
11b10922 LB |
4062 | * systemctl show-environment gained support for --output=json. |
4063 | ||
bf71ade8 | 4064 | * pam_systemd will now first try to use the X11 abstract socket, and |
dcdc652f ZJS |
4065 | fallback to the socket file in /tmp/.X11-unix/ only if that does not |
4066 | work. | |
11b10922 | 4067 | |
dcdc652f ZJS |
4068 | * systemd-journald will no longer go back to volatile storage |
4069 | regardless of configuration when its unit is restarted. | |
bf71ade8 | 4070 | |
dcdc652f ZJS |
4071 | * Initial support for the LoongArch architecture has been added (system |
4072 | call lists, GPT partition table UUIDs, etc). | |
efeecf40 | 4073 | |
dcdc652f ZJS |
4074 | * systemd-journald's own logging messages are now also logged to the |
4075 | journal itself when systemd-journald logs to /dev/kmsg. | |
6959a051 | 4076 | |
dcdc652f ZJS |
4077 | * systemd-journald now re-enables COW for archived journal files on |
4078 | filesystems that support COW. One benefit of this change is that | |
4079 | archived journal files will now get compressed on btrfs filesystems | |
4080 | that have compression enabled. | |
6959a051 | 4081 | |
dcdc652f ZJS |
4082 | * systemd-journald now deduplicates fields in a single log message |
4083 | before adding it to the journal. In archived journal files, it will | |
4084 | also punch holes for unused parts and truncate the file as | |
4085 | appropriate, leading to reductions in disk usage. | |
6959a051 | 4086 | |
dcdc652f ZJS |
4087 | * journalctl --verify was extended with more informative error |
4088 | messages. | |
6959a051 | 4089 | |
dcdc652f ZJS |
4090 | * More of sd-journal's functions are now resistant against journal file |
4091 | corruption. | |
6959a051 | 4092 | |
e63fa075 ZJS |
4093 | * The shutdown command learnt a new option --show, to display the |
4094 | scheduled shutdown. | |
4095 | ||
dcdc652f ZJS |
4096 | * A LICENSES/ directory is now included in the git tree. It contains a |
4097 | README.md file that explains the licenses used by source files in | |
33db1b90 | 4098 | this repository. It also contains the text of all applicable |
dcdc652f | 4099 | licenses as they appear on spdx.org. |
6959a051 | 4100 | |
97b6ed32 ZJS |
4101 | Contributions from: Aakash Singh, acsfer, Adolfo Jayme Barrientos, |
4102 | Adrian Vovk, Albert Brox, Alberto Mardegan, Alexander Kanavin, | |
4103 | alexlzhu, Alfonso Sánchez-Beato, Alvin Šipraga, Alyssa Ross, | |
4104 | Amir Omidi, Anatol Pomozov, Andika Triwidada, Andreas Rammhold, | |
4105 | Andreas Valder, Andrej Lajovic, Andrew Soutar, Andrew Stone, Andy Chi, | |
4106 | Anita Zhang, Anssi Hannula, Antonio Alvarez Feijoo, | |
4107 | Antony Deepak Thomas, Arnaud Ferraris, Arvid E. Picciani, | |
4108 | Bastien Nocera, Benjamin Berg, Benjamin Herrenschmidt, Ben Stockett, | |
4109 | Bogdan Seniuc, Boqun Feng, Carl Lei, chlorophyll-zz, Chris Packham, | |
4110 | Christian Brauner, Christian Göttsche, Christian Wehrli, | |
4111 | Christoph Anton Mitterer, Cristian Rodríguez, Daan De Meyer, | |
4112 | Daniel Maixner, Dann Frazier, Dan Streetman, Davide Cavalca, | |
4113 | David Seifert, David Tardon, dependabot[bot], Dimitri John Ledkov, | |
4114 | Dimitri Papadopoulos, Dimitry Ishenko, Dmitry Khlebnikov, | |
4115 | Dominique Martinet, duament, Egor, Egor Ignatov, Emil Renner Berthing, | |
4116 | Emily Gonyer, Ettore Atalan, Evgeny Vereshchagin, Florian Klink, | |
4117 | Franck Bui, Frantisek Sumsal, Geass-LL, Gibeom Gwon, GnunuX, | |
4118 | Gogo Gogsi, gregzuro, Greg Zuro, Gustavo Costa, Hans de Goede, | |
484abbe6 ZJS |
4119 | Hela Basa, Henri Chain, hikigaya58, Hugo Carvalho, |
4120 | Hugo Osvaldo Barrera, Iago Lopez Galeiras, Iago López Galeiras, | |
97b6ed32 ZJS |
4121 | I-dont-need-name, igo95862, Jack Dähn, James Hilliard, Jan Janssen, |
4122 | Jan Kuparinen, Jan Macku, Jan Palus, Jarkko Sakkinen, Jayce Fayne, | |
4123 | jiangchuangang, jlempen, John Lindgren, Jonas Dreßler, Jonas Jelten, | |
4124 | Jonas Witschel, Joris Hartog, José Expósito, Julia Kartseva, | |
4125 | Kai-Heng Feng, Kai Wohlfahrt, Kay Siver Bø, KennthStailey, | |
4126 | Kevin Kuehler, Kevin Orr, Khem Raj, Kristian Klausen, Kyle Laker, | |
4127 | lainahai, LaserEyess, Lennart Poettering, Lia Lenckowski, longpanda, | |
4128 | Luca Boccassi, Luca BRUNO, Ludwig Nussel, Lukas Senionis, | |
4129 | Maanya Goenka, Maciek Borzecki, Marcel Menzel, Marco Scardovi, | |
4130 | Marcus Harrison, Mark Boudreau, Matthijs van Duin, Mauricio Vásquez, | |
4131 | Maxime de Roucy, Max Resch, MertsA, Michael Biebl, Michael Catanzaro, | |
4132 | Michal Koutný, Michal Sekletár, Miika Karanki, Mike Gilbert, | |
4133 | Milo Turner, ml, monosans, Nacho Barrientos, nassir90, Nishal Kulkarni, | |
e63fa075 | 4134 | nl6720, Ondrej Kozina, Paulo Neves, Pavel Březina, pedro martelletto, |
484abbe6 | 4135 | Peter Hutterer, Peter Morrow, Piotr Drąg, Rasmus Villemoes, ratijas, |
97b6ed32 | 4136 | Raul Tambre, rene, Riccardo Schirone, Robert-L-Turner, Robert Scheck, |
6e6dc095 YW |
4137 | Ross Jennings, saikat0511, Scott Lamb, Scott Worley, |
4138 | Sergei Trofimovich, Sho Iizuka, Slava Bacherikov, Slimane Selyan Amiri, | |
4139 | StefanBruens, Steven Siloti, svonohr, Taiki Sugawara, Takashi Sakamoto, | |
4140 | Takuro Onoue, Thomas Blume, Thomas Haller, Thomas Mühlbacher, | |
4141 | Tianlu Shao, Toke Høiland-Jørgensen, Tom Yan, Tony Asleson, | |
4142 | Topi Miettinen, Ulrich Ölmann, Urs Ritzmann, Vincent Bernat, | |
4143 | Vito Caputo, Vladimir Panteleev, WANG Xuerui, Wind/owZ, Wu Xiaotian, | |
4144 | xdavidwu, Xiaotian Wu, xujing, yangmingtai, Yao Wei, Yao Wei (魏銘廷), | |
4145 | Yegor Alexeyev, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, | |
4146 | Дамјан Георгиевски, наб | |
4147 | ||
4148 | — Warsaw, 2021-12-23 | |
484abbe6 | 4149 | |
e7fbba56 | 4150 | CHANGES WITH 249: |
f973aea7 LP |
4151 | |
4152 | * When operating on disk images via the --image= switch of various | |
e7fbba56 LB |
4153 | tools (such as systemd-nspawn or systemd-dissect), or when udev finds |
4154 | no 'root=' parameter on the kernel command line, and multiple | |
f973aea7 | 4155 | suitable root or /usr/ partitions exist in the image, then a simple |
28707969 ZJS |
4156 | comparison inspired by strverscmp() is done on the GPT partition |
4157 | label, and the newest partition is picked. This permits a simple and | |
4158 | generic whole-file-system A/B update logic where new operating system | |
f973aea7 LP |
4159 | versions are dropped into partitions whose label is then updated with |
4160 | a matching version identifier. | |
4161 | ||
4162 | * systemd-sysusers now supports querying the passwords to set for the | |
4163 | users it creates via the "credentials" logic introduced in v247: the | |
4164 | passwd.hashed-password.<user> and passwd.plaintext-password.<user> | |
4165 | credentials are consulted for the password to use (either in UNIX | |
4166 | hashed form, or literally). By default these credentials are inherited | |
4167 | down from PID1 (which in turn imports it from a container manager if | |
4168 | there is one). This permits easy configuration of user passwords | |
4169 | during first boot. Example: | |
4170 | ||
4171 | # systemd-nspawn -i foo.raw --volatile=yes --set-credential=passwd.plaintext-password.root:foo | |
4172 | ||
4173 | Note that systemd-sysusers operates in purely additive mode: it | |
4174 | executes no operation if the declared users already exist, and hence | |
4175 | doesn't set any passwords as effect of the command line above if the | |
4176 | specified root user exists already in the image. (Note that | |
4177 | --volatile=yes ensures it doesn't, though.) | |
4178 | ||
4179 | * systemd-firstboot now also supports querying various system | |
4180 | parameters via the credential subsystems. Thus, as above this may be | |
4181 | used to initialize important system parameters on first boot of | |
4182 | previously unprovisioned images (i.e. images with a mostly empty | |
4183 | /etc/). | |
4184 | ||
66e6128f LP |
4185 | * PID 1 may now show both the unit name and the unit description |
4186 | strings in its status output during boot. This may be configured with | |
4187 | StatusUnitFormat=combined in system.conf or | |
4188 | systemd.status-unit-format=combined on the kernel command line. | |
4189 | ||
f973aea7 LP |
4190 | * The systemd-machine-id-setup tool now supports a --image= switch for |
4191 | provisioning a machine ID file into an OS disk image, similar to how | |
4192 | --root= operates on an OS file tree. This matches the existing switch | |
28707969 | 4193 | of the same name for systemd-tmpfiles, systemd-firstboot, and |
f973aea7 LP |
4194 | systemd-sysusers tools. |
4195 | ||
28707969 ZJS |
4196 | * Similarly, systemd-repart gained support for the --image= switch too. |
4197 | In combination with the existing --size= option, this makes the tool | |
4198 | particularly useful for easily growing disk images in a single | |
4199 | invocation, following the declarative rules included in the image | |
4200 | itself. | |
f973aea7 LP |
4201 | |
4202 | * systemd-repart's partition configuration files gained support for a | |
4203 | new switch MakeDirectories= which may be used to create arbitrary | |
4204 | directories inside file systems that are created, before registering | |
28707969 ZJS |
4205 | them in the partition table. This is useful in particular for root |
4206 | partitions to create mount point directories for other partitions | |
4207 | included in the image. For example, a disk image that contains a | |
4208 | root, /home/, and /var/ partitions, may set MakeDirectories=yes to | |
4209 | create /home/ and /var/ as empty directories in the root file system | |
5601400e JB |
4210 | on its creation, so that the resulting image can be mounted |
4211 | immediately, even in read-only mode. | |
28707969 ZJS |
4212 | |
4213 | * systemd-repart's CopyBlocks= setting gained support for the special | |
4214 | value "auto". If used, a suitable matching partition on the booted OS | |
4215 | is found as source to copy blocks from. This is useful when | |
f973aea7 | 4216 | implementing replicating installers, that are booted from one medium |
28707969 | 4217 | and then stream their own root partition onto the target medium. |
f973aea7 LP |
4218 | |
4219 | * systemd-repart's partition configuration files gained support for a | |
66e6128f LP |
4220 | Flags=, a ReadOnly= and a NoAuto= setting, allowing control of these |
4221 | GPT partition flags for the created partitions: this is useful for | |
4222 | marking newly created partitions as read-only, or as not being | |
4223 | subject for automatic mounting from creation on. | |
f973aea7 LP |
4224 | |
4225 | * The /etc/os-release file has been extended with two new (optional) | |
28707969 ZJS |
4226 | variables IMAGE_VERSION= and IMAGE_ID=, carrying identity and version |
4227 | information for OS images that are updated comprehensively and | |
4228 | atomically as one image. Two new specifiers %M, %A now resolve to | |
4229 | these two fields in the various configuration options that resolve | |
4230 | specifiers. | |
f973aea7 LP |
4231 | |
4232 | * portablectl gained a new switch --extension= for enabling portable | |
4233 | service images with extensions that follow the extension image | |
165c23c6 LB |
4234 | concept introduced with v248, and thus allows layering multiple |
4235 | images when setting up the root filesystem of the service. | |
f973aea7 LP |
4236 | |
4237 | * systemd-coredump will now extract ELF build-id information from | |
28707969 ZJS |
4238 | processes dumping core and include it in the coredump report. |
4239 | Moreover, it will look for ELF .note.package sections with | |
4240 | distribution packaging meta-information about the crashing process. | |
4241 | This is useful to directly embed the rpm or deb (or any other) | |
4242 | package name and version in ELF files, making it easy to match | |
4243 | coredump reports with the specific package for which the software was | |
4244 | compiled. This is particularly useful on environments with ELF files | |
4245 | from multiple vendors, different distributions and versions, as is | |
4246 | common today in our containerized and sand-boxed world. For further | |
4247 | information, see: | |
f973aea7 LP |
4248 | |
4249 | https://systemd.io/COREDUMP_PACKAGE_METADATA | |
4250 | ||
28707969 ZJS |
4251 | * A new udev hardware database has been added for FireWire devices |
4252 | (IEEE 1394). | |
f973aea7 | 4253 | |
14e97d24 ZJS |
4254 | * The "net_id" built-in of udev has been updated with three |
4255 | backwards-incompatible changes: | |
4256 | ||
4257 | - PCI hotplug slot names on s390 systems are now parsed as | |
4258 | hexadecimal numbers. They were incorrectly parsed as decimal | |
4259 | previously, or ignored if the name was not a valid decimal | |
4260 | number. | |
4261 | ||
4262 | - PCI onboard indices up to 65535 are allowed. Previously, numbers | |
4263 | above 16383 were rejected. This primarily impacts s390 systems, | |
4264 | where values up to 65535 are used. | |
4265 | ||
4266 | - Invalid characters in interface names are replaced with "_". | |
4267 | ||
4268 | The new version of the net naming scheme is "v249". The previous | |
4269 | scheme can be selected via the "net.naming-scheme=v247" kernel | |
4270 | command line parameter. | |
4271 | ||
f973aea7 LP |
4272 | * sd-bus' sd_bus_is_ready() and sd_bus_is_open() calls now accept a |
4273 | NULL bus object, for which they will return false. Or in other words, | |
4274 | an unallocated bus connection is neither ready nor open. | |
4275 | ||
99c2a955 | 4276 | * The sd-device API acquired a new API function |
28707969 ZJS |
4277 | sd_device_get_usec_initialized() that returns the monotonic time when |
4278 | the udev device first appeared in the database. | |
99c2a955 LP |
4279 | |
4280 | * sd-device gained a new APIs sd_device_trigger_with_uuid() and | |
4281 | sd_device_get_trigger_uuid(). The former is similar to | |
4282 | sd_device_trigger() but returns a randomly generated UUID that is | |
4283 | associated with the synthetic uevent generated by the call. This UUID | |
4284 | may be read from the sd_device object a monitor eventually receives, | |
4285 | via the sd_device_get_trigger_uuid(). This interface requires kernel | |
4286 | 4.13 or above to work, and allows tracking a synthetic uevent through | |
4287 | the entire device management stack. The "udevadm trigger --settle" | |
4288 | logic has been updated to make use of this concept if available to | |
4289 | wait precisely for the uevents it generates. "udevadm trigger" also | |
4290 | gained a new parameter --uuid that prints the UUID for each generated | |
4291 | uevent. | |
4292 | ||
ca1e8584 LP |
4293 | * sd-device also gained new APIs sd_device_new_from_ifname() and |
4294 | sd_device_new_from_ifindex() for allocating an sd-device object for | |
4295 | the specified network interface. The former accepts an interface name | |
4296 | (either a primary or an alternative name), the latter an interface | |
4297 | index. | |
4298 | ||
f973aea7 LP |
4299 | * The native Journal protocol has been documented. Clients may talk |
4300 | this as alternative to the classic BSD syslog protocol for locally | |
4301 | delivering log records to the Journal. The protocol has been stable | |
5601400e | 4302 | for a long time and in fact been implemented already in a variety |
f973aea7 LP |
4303 | of alternative client libraries. This documentation makes the support |
4304 | for that official: | |
4305 | ||
4306 | https://systemd.io/JOURNAL_NATIVE_PROTOCOL | |
4307 | ||
4308 | * A new BPFProgram= setting has been added to service files. It may be | |
4309 | set to a path to a loaded kernel BPF program, i.e. a path to a bpffs | |
4310 | file, or a bind mount or symlink to one. This may be used to upload | |
4311 | and manage BPF programs externally and then hook arbitrary systemd | |
4312 | services into them. | |
4313 | ||
4314 | * The "home.arpa" domain that has been officially declared as the | |
4315 | choice for domain for local home networks per RFC 8375 has been added | |
4316 | to the default NTA list of resolved, since DNSSEC is generally not | |
4317 | available on private domains. | |
4318 | ||
4319 | * The CPUAffinity= setting of unit files now resolves "%" specifiers. | |
4320 | ||
4321 | * A new ManageForeignRoutingPolicyRules= setting has been added to | |
4322 | .network files which may be used to exclude foreign-created routing | |
4323 | policy rules from systemd-networkd management. | |
4324 | ||
4325 | * systemd-network-wait-online gained two new switches -4 and -6 that | |
4326 | may be used to tweak whether to wait for only IPv4 or only IPv6 | |
4327 | connectivity. | |
4328 | ||
4329 | * .network files gained a new RequiredFamilyForOnline= setting to | |
4330 | fine-tune whether to require an IPv4 or IPv6 address in order to | |
4331 | consider an interface "online". | |
4332 | ||
99c2a955 LP |
4333 | * networkctl will now show an over-all "online" state in the per-link |
4334 | information. | |
4335 | ||
4336 | * In .network files a new OutgoingInterface= setting has been added to | |
4337 | specify the output interface in bridge FDB setups. | |
4338 | ||
566c8176 | 4339 | * In .network files the Multipath group ID may now be configured for |
99c2a955 | 4340 | [NextHop] entries, via the new Group= setting. |
f973aea7 LP |
4341 | |
4342 | * The DHCP server logic configured in .network files gained a new | |
28707969 ZJS |
4343 | setting RelayTarget= that turns the server into a DHCP server relay. |
4344 | The RelayAgentCircuitId= and RelayAgentRemoteId= settings may be used | |
4345 | to further tweak the DHCP relay behaviour. | |
f973aea7 | 4346 | |
28707969 ZJS |
4347 | * The DHCP server logic also gained a new ServerAddress= setting in |
4348 | .network files that explicitly specifies the server IP address to | |
4349 | use. If not specified, the address is determined automatically, as | |
4350 | before. | |
f973aea7 | 4351 | |
99c2a955 LP |
4352 | * The DHCP server logic in systemd-networkd gained support for static |
4353 | DHCP leases, configurable via the [DHCPServerStaticLease] | |
4354 | section. This allows explicitly mapping specific MAC addresses to | |
4355 | fixed IP addresses and vice versa. | |
4356 | ||
4357 | * The RestrictAddressFamilies= setting in service files now supports a | |
4358 | new special value "none". If specified sockets of all address | |
4359 | families will be made unavailable to services configured that way. | |
f973aea7 LP |
4360 | |
4361 | * systemd-fstab-generator and systemd-repart have been updated to | |
4362 | support booting from disks that carry only a /usr/ partition but no | |
28707969 ZJS |
4363 | root partition yet, and where systemd-repart can add it in on the |
4364 | first boot. This is useful for implementing systems that ship with a | |
4365 | single /usr/ file system, and whose root file system shall be set up | |
4366 | and formatted on a LUKS-encrypted volume whose key is generated | |
4367 | locally (and possibly enrolled in the TPM) during the first boot. | |
f973aea7 LP |
4368 | |
4369 | * The [Address] section of .network files now accepts a new | |
4370 | RouteMetric= setting that configures the routing metric to use for | |
28707969 ZJS |
4371 | the prefix route created as effect of the address configuration. |
4372 | Similarly, the [DHCPv6PrefixDelegation] and [IPv6Prefix] sections | |
4373 | gained matching settings for their prefix routes. (The option of the | |
4374 | same name in the [DHCPv6] section is moved to [IPv6AcceptRA], since | |
4375 | it conceptually belongs there; the old option is still understood for | |
4376 | compatibility.) | |
f973aea7 | 4377 | |
f973aea7 LP |
4378 | * The DHCPv6 IAID and DUID are now explicitly configurable in .network |
4379 | files. | |
4380 | ||
4381 | * A new udev property ID_NET_DHCP_BROADCAST on network interface | |
28707969 | 4382 | devices is now honoured by systemd-networkd, controlling whether to |
f973aea7 LP |
4383 | issue DHCP offers via broadcasting. This is used to ensure that s390 |
4384 | layer 3 network interfaces work out-of-the-box with systemd-networkd. | |
4385 | ||
4386 | * nss-myhostname and systemd-resolved will now synthesize address | |
4387 | records for a new special hostname "_outbound". The name will always | |
4388 | resolve to the local IP addresses most likely used for outbound | |
28707969 | 4389 | connections towards the default routes. On multi-homed hosts this is |
f973aea7 LP |
4390 | useful to have a stable handle referring to "the" local IP address |
4391 | that matters most, to the point where this is defined. | |
4392 | ||
4393 | * The Discoverable Partition Specification has been updated with a new | |
5601400e JB |
4394 | GPT partition flag "grow-file-system" defined for its partition |
4395 | types. Whenever partitions with this flag set are automatically | |
4396 | mounted (i.e. via systemd-gpt-auto-generator or the --image= switch | |
4397 | of systemd-nspawn or other tools; and as opposed to explicit mounting | |
f973aea7 | 4398 | via /etc/fstab), the file system within the partition is |
28707969 ZJS |
4399 | automatically grown to the full size of the partition. If the file |
4400 | system size already matches the partition size this flag has no | |
4401 | effect. Previously, this functionality has been available via the | |
4402 | explicit x-systemd.growfs mount option, and this new flag extends | |
4403 | this to automatically discovered mounts. A new GrowFileSystem= | |
4404 | setting has been added to systemd-repart drop-in files that allows | |
4405 | configuring this partition flag. This new flag defaults to on for | |
4406 | partitions automatically created by systemd-repart, except if they | |
4407 | are marked read-only. See the specification for further details: | |
f973aea7 LP |
4408 | |
4409 | https://systemd.io/DISCOVERABLE_PARTITIONS | |
4410 | ||
99c2a955 | 4411 | * .network files gained a new setting RoutesToNTP= in the [DHCPv4] |
f973aea7 LP |
4412 | section. If enabled (which is the default), and an NTP server address |
4413 | is acquired through a DHCP lease on this interface an explicit route | |
4414 | to this address is created on this interface to ensure that NTP | |
4415 | traffic to the NTP server acquired on an interface is also routed | |
4076ad9d | 4416 | through that interface. The pre-existing RoutesToDNS= setting that |
f973aea7 LP |
4417 | implements the same for DNS servers is now enabled by default. |
4418 | ||
4419 | * A pair of service settings SocketBindAllow= + SocketBindDeny= have | |
4420 | been added that may be used to restrict the network interfaces | |
4421 | sockets created by the service may be bound to. This is implemented | |
4422 | via BPF. | |
4423 | ||
14e97d24 ZJS |
4424 | * A new ConditionFirmware= setting has been added to unit files to |
4425 | conditionalize on certain firmware features. At the moment it may | |
5bc9ea07 | 4426 | check whether running on a UEFI system, a device.tree system, or if |
14e97d24 ZJS |
4427 | the system is compatible with some specified device-tree feature. |
4428 | ||
4429 | * A new ConditionOSRelease= setting has been added to unit files to | |
4430 | check os-release(5) fields. The "=", "!=", "<", "<=", ">=", ">" | |
4431 | operators may be used to check if some field has some specific value | |
387f6955 YW |
4432 | or do an alphanumerical comparison. Equality comparisons are useful |
4433 | for fields like ID, but relative comparisons for fields like | |
4434 | VERSION_ID or IMAGE_VERSION. | |
f973aea7 LP |
4435 | |
4436 | * hostnamed gained a new Describe() D-Bus method that returns a JSON | |
4437 | serialization of the host data it exposes. This is exposed via | |
28707969 ZJS |
4438 | "hostnamectl --json=" to acquire a host identity description in JSON. |
4439 | It's our intention to add a similar features to most services and | |
4440 | objects systemd manages, in order to simplify integration with | |
4441 | program code that can consume JSON. | |
f973aea7 | 4442 | |
28707969 ZJS |
4443 | * Similarly, networkd gained a Describe() method on its Manager and |
4444 | Link bus objects. This is exposed via "networkctl --json=". | |
5f94ac54 | 4445 | |
28707969 ZJS |
4446 | * hostnamectl's various "get-xyz"/"set-xyz" verb pairs |
4447 | (e.g. "hostnamectl get-hostname", "hostnamectl "set-hostname") have | |
4448 | been replaced by a single "xyz" verb (e.g. "hostnamectl hostname") | |
4449 | that is used both to get the value (when no argument is given), and | |
4450 | to set the value (when an argument is specified). The old names | |
4451 | continue to be supported for compatibility. | |
f973aea7 LP |
4452 | |
4453 | * systemd-detect-virt and ConditionVirtualization= are now able to | |
4454 | correctly identify Amazon EC2 environments. | |
4455 | ||
4456 | * The LogLevelMax= setting of unit files now applies not only to log | |
4457 | messages generated *by* the service, but also to log messages | |
4458 | generated *about* the service by PID 1. To suppress logs concerning a | |
4459 | specific service comprehensively, set this option to a high log | |
4460 | level. | |
4461 | ||
4462 | * bootctl gained support for a new --make-machine-id-directory= switch | |
4463 | that allows precise control on whether to create the top-level | |
28707969 ZJS |
4464 | per-machine directory in the boot partition that typically contains |
4465 | Type 1 boot loader entries. | |
f973aea7 | 4466 | |
5f94ac54 LP |
4467 | * During build SBAT data to include in the systemd-boot EFI PE binaries |
4468 | may be specified now. | |
4469 | ||
4470 | * /etc/crypttab learnt a new option "headless". If specified any | |
4471 | requests to query the user interactively for passwords or PINs will | |
4472 | be skipped. This is useful on systems that are headless, i.e. where | |
4473 | an interactive user is generally not present. | |
4474 | ||
00fb6caf LP |
4475 | * /etc/crypttab also learnt a new option "password-echo=" that allows |
4476 | configuring whether the encryption password prompt shall echo the | |
4477 | typed password and if so, do so literally or via asterisks. (The | |
4478 | default is the same behaviour as before: provide echo feedback via | |
4479 | asterisks.) | |
4480 | ||
5f94ac54 LP |
4481 | * FIDO2 support in systemd-cryptenroll/systemd-cryptsetup and |
4482 | systemd-homed has been updated to allow explicit configuration of the | |
4483 | "user presence" and "user verification" checks, as well as whether a | |
4484 | PIN is required for authentication, via the new switches | |
4485 | --fido2-with-user-presence=, --fido2-with-user-verification=, | |
4486 | --fido2-with-client-pin= to systemd-cryptenroll and homectl. Which | |
4487 | features are available, and may be enabled or disabled depends on the | |
4488 | used FIDO2 token. | |
4489 | ||
4490 | * systemd-nspawn's --private-user= switch now accepts the special value | |
4491 | "identity" which configures a user namespacing environment with an | |
4492 | identity mapping of 65535 UIDs. This means the container UID 0 is | |
4493 | mapped to the host UID 0, and the UID 1 to host UID 1. On first look | |
4494 | this doesn't appear to be useful, however it does reduce the attack | |
4495 | surface a bit, since the resulting container will possess process | |
4496 | capabilities only within its namespace and not on the host. | |
4497 | ||
4498 | * systemd-nspawn's --private-user-chown switch has been replaced by a | |
4499 | more generic --private-user-ownership= switch that accepts one of | |
4500 | three values: "chown" is equivalent to the old --private-user-chown, | |
4501 | and "off" is equivalent to the absence of the old switch. The value | |
4502 | "map" uses the new UID mapping mounts of Linux 5.12 to map ownership | |
4503 | of files and directories of the underlying image to the chosen UID | |
4504 | range for the container. "auto" is equivalent to "map" if UID mapping | |
4505 | mount are supported, otherwise it is equivalent to "chown". The short | |
4506 | -U switch systemd-nspawn now implies --private-user-ownership=auto | |
4507 | instead of the old --private-user-chown. Effectively this means: if | |
4508 | the backing file system supports UID mapping mounts the feature is | |
4509 | now used by default if -U is used. Generally, it's a good idea to use | |
4510 | UID mapping mounts instead of recursive chown()ing, since it allows | |
4511 | running containers off immutable images (since no modifications of | |
4512 | the images need to take place), and share images between multiple | |
4513 | instances. Moreover, the recursive chown()ing operation is slow and | |
4514 | can be avoided. Conceptually it's also a good thing if transient UID | |
4515 | range uses do not leak into persistent file ownership anymore. TLDR: | |
4516 | finally, the last major drawback of user namespacing has been | |
4517 | removed, and -U should always be used (unless you use btrfs, where | |
4518 | UID mapped mounts do not exist; or your container actually needs | |
4519 | privileges on the host). | |
4520 | ||
4521 | * nss-systemd now synthesizes user and group shadow records in addition | |
4522 | to the main user and group records. Thus, hashed passwords managed by | |
4523 | systemd-homed are now accessible via the shadow database. | |
4524 | ||
4525 | * The userdb logic (and thus nss-systemd, and so on) now read | |
4526 | additional user/group definitions in JSON format from the drop-in | |
4527 | directories /etc/userdb/, /run/userdb/, /run/host/userdb/ and | |
4528 | /usr/lib/userdb/. This is a simple and powerful mechanism for making | |
4529 | additional users available to the system, with full integration into | |
4530 | NSS including the shadow databases. Since the full JSON user/group | |
4531 | record format is supported this may also be used to define users with | |
4532 | resource management settings and other runtime settings that | |
4533 | pam_systemd and systemd-logind enforce at login. | |
4534 | ||
4535 | * The userdbctl tool gained two new switches --with-dropin= and | |
4536 | --with-varlink= which can be used to fine-tune the sources used for | |
4537 | user database lookups. | |
4538 | ||
4539 | * systemd-nspawn gained a new switch --bind-user= for binding a host | |
4540 | user account into the container. This does three things: the user's | |
4541 | home directory is bind mounted from the host into the container, | |
4542 | below the /run/userdb/home/ hierarchy. A free UID is picked in the | |
4543 | container, and a user namespacing UID mapping to the host user's UID | |
4544 | installed. And finally, a minimal JSON user and group record (along | |
4545 | with its hashed password) is dropped into /run/host/userdb/. These | |
4546 | records are picked up automatically by the userdb drop-in logic | |
4547 | describe above, and allow the user to login with the same password as | |
4548 | on the host. Effectively this means: if host and container run new | |
4549 | enough systemd versions making a host user available to the container | |
4550 | is trivially simple. | |
4551 | ||
4552 | * systemd-journal-gatewayd now supports the switches --user, --system, | |
4553 | --merge, --file= that are equivalent to the same switches of | |
4554 | journalctl, and permit exposing only the specified subset of the | |
4555 | Journal records. | |
4556 | ||
5f94ac54 LP |
4557 | * The OnFailure= dependency between units is now augmented with a |
4558 | implicit reverse dependency OnFailureOf= (this new dependency cannot | |
4559 | be configured directly it's only created as effect of an OnFailure= | |
4560 | dependency in the reverse order — it's visible in "systemctl show" | |
4561 | however). Similar, Slice= now has an reverse dependency SliceOf=, | |
4562 | that is also not configurable directly, but useful to determine all | |
4563 | units that are members of a slice. | |
4564 | ||
4565 | * A pair of new dependency types between units PropagatesStopTo= + | |
4566 | StopPropagatedFrom= has been added, that allows propagation of unit | |
4567 | stop events between two units. It operates similar to the existing | |
4568 | PropagatesReloadTo= + ReloadPropagatedFrom= dependencies. | |
f973aea7 | 4569 | |
99c2a955 LP |
4570 | * A new dependency type OnSuccess= has been added (plus the reverse |
4571 | dependency OnSuccessOf=, which cannot be configured directly, but | |
28707969 ZJS |
4572 | exists only as effect of the reverse OnSuccess=). It is similar to |
4573 | OnFailure=, but triggers in the opposite case: when a service exits | |
99c2a955 | 4574 | cleanly. This allows "chaining up" of services where one or more |
28707969 | 4575 | services are started once another service has successfully completed. |
99c2a955 LP |
4576 | |
4577 | * A new dependency type Upholds= has been added (plus the reverse | |
28707969 ZJS |
4578 | dependency UpheldBy=, which cannot be configured directly, but exists |
4579 | only as effect of Upholds=). This dependency type is a stronger form | |
4580 | of Wants=: if a unit has an UpHolds= dependency on some other unit | |
4581 | and the former is active then the latter is started whenever it is | |
4582 | found inactive (and no job is queued for it). This is an alternative | |
4583 | to Restart= inside service units, but less configurable, and the | |
4584 | request to uphold a unit is not encoded in the unit itself but in | |
4585 | another unit that intends to uphold it. | |
99c2a955 LP |
4586 | |
4587 | * The systemd-ask-password tool now also supports reading passwords | |
4588 | from the credentials subsystem, via the new --credential= switch. | |
4589 | ||
4590 | * The systemd-ask-password tool learnt a new switch --emoji= which may | |
4591 | be used to explicit control whether the lock and key emoji (🔐) is | |
4592 | shown in the password prompt on suitable TTYs. | |
4593 | ||
4594 | * The --echo switch of systemd-ask-password now optionally takes a | |
28707969 ZJS |
4595 | parameter that controls character echo. It may either show asterisks |
4596 | (default, as before), turn echo off entirely, or echo the typed | |
4597 | characters literally. | |
99c2a955 | 4598 | |
66e6128f LP |
4599 | * The systemd-ask-password tool also gained a new -n switch for |
4600 | suppressing output of a trailing newline character when writing the | |
4601 | acquired password to standard output, similar to /bin/echo's -n | |
4602 | switch. | |
4603 | ||
99c2a955 LP |
4604 | * New documentation has been added that describes the organization of |
4605 | the systemd source code tree: | |
4606 | ||
4607 | https://systemd.io/ARCHITECTURE | |
4608 | ||
165c23c6 LB |
4609 | * Units using ConditionNeedsUpdate= will no longer be activated in |
4610 | the initrd. | |
4611 | ||
00fb6caf LP |
4612 | * It is now possible to list a template unit in the WantedBy= or |
4613 | RequiredBy= settings of the [Install] section of another template | |
4614 | unit, which will be instantiated using the same instance name. | |
165c23c6 LB |
4615 | |
4616 | * A new MemoryAvailable property is available for units. If the unit, | |
0923b425 | 4617 | or the slices it is part of, have a memory limit set via MemoryMax=/ |
165c23c6 | 4618 | MemoryHigh=, MemoryAvailable will indicate how much more memory the |
0923b425 | 4619 | unit can claim before hitting the limits. |
165c23c6 LB |
4620 | |
4621 | * systemd-coredump will now try to stay below the cgroup memory limit | |
4622 | placed on itself or one of the slices it runs under, if the storage | |
4623 | area for core files (/var/lib/systemd/coredump/) is placed on a tmpfs, | |
4624 | since files written on such filesystems count toward the cgroup memory | |
4625 | limit. If there is not enough available memory in such cases to store | |
4626 | the core file uncompressed, systemd-coredump will skip to compressed | |
4627 | storage directly (if enabled) and it will avoid analyzing the core file | |
4628 | to print backtrace and metadata in the journal. | |
4629 | ||
00fb6caf LP |
4630 | * tmpfiles.d/ drop-ins gained a new '=' modifier to check if the type |
4631 | of a path matches the configured expectations, and remove it if not. | |
165c23c6 | 4632 | |
00fb6caf LP |
4633 | * tmpfiles.d/'s 'Age' now accepts an 'age-by' argument, which allows to |
4634 | specify which of the several available filesystem timestamps (access | |
4635 | time, birth time, change time, modification time) to look at when | |
4636 | deciding whether a path has aged enough to be cleaned. | |
165c23c6 | 4637 | |
00fb6caf LP |
4638 | * A new IPv6StableSecretAddress= setting has been added to .network |
4639 | files, which takes an IPv6 address to use as secret for IPv6 address | |
4640 | generation. | |
4641 | ||
4642 | * The [DHCPServer] logic in .network files gained support for a new | |
4643 | UplinkInterface= setting that permits configuration of the uplink | |
4644 | interface name to propagate DHCP lease information from. | |
4645 | ||
4646 | * The WakeOnLan= setting in .link files now accepts a list of flags | |
4647 | instead of a single one, to configure multiple wake-on-LAN policies. | |
4648 | ||
ca1e8584 LP |
4649 | * User-space defined tracepoints (USDT) have been added to udev at |
4650 | strategic locations. This is useful for tracing udev behaviour and | |
4651 | performance with bpftrace and similar tools. | |
4652 | ||
4653 | * systemd-journald-upload gained a new NetworkTimeoutSec= option for | |
4654 | setting a network timeout time. | |
4655 | ||
5b8fdb18 LB |
4656 | * If a system service is running in a new mount namespace (RootDirectory= |
4657 | and friends), all file systems will be mounted with MS_NOSUID by | |
4658 | default, unless the system is running with SELinux enabled. | |
4659 | ||
66e6128f LP |
4660 | * When enumerating time zones the timedatectl tool will now consult the |
4661 | 'tzdata.zi' file shipped by the IANA time zone database package, in | |
4662 | addition to 'zone1970.tab', as before. This makes sure time zone | |
4663 | aliases are now correctly supported. Some distributions so far did | |
4664 | not install this additional file, most do however. If you | |
4665 | distribution does not install it yet, it might make sense to change | |
4666 | that. | |
4667 | ||
3c88ade5 KHF |
4668 | * Intel HID rfkill event is no longer masked, since it's the only |
4669 | source of rfkill event on newer HP laptops. To have both backward and | |
4670 | forward compatibility, userspace daemon needs to debounce duplicated | |
4671 | events in a short time window. | |
4672 | ||
b2f0876b | 4673 | Contributions from: Aakash Singh, adrian5, Albert Brox, |
fc4340b0 LB |
4674 | Alexander Sverdlin, Alexander Tsoy, Alexey Rubtsov, alexlzhu, |
4675 | Allen Webb, Alvin Šipraga, Alyssa Ross, Anders Wenhaug, | |
4676 | Andrea Pappacoda, Anita Zhang, asavah, Balint Reczey, Bertrand Jacquin, | |
4677 | borna-blazevic, caoxia2008cxx, Carlo Teubner, Christian Göttsche, | |
4678 | Christian Hesse, Daniel Schaefer, Dan Streetman, | |
4679 | David Santamaría Rogado, David Tardon, Deepak Rawat, dgcampea, | |
4680 | Dimitri John Ledkov, ei-ke, Emilio Herrera, Emil Renner Berthing, | |
4681 | Eric Cook, Flos Lonicerae, Franck Bui, Francois Gervais, | |
4682 | Frantisek Sumsal, Gibeom Gwon, gitm0, Hamish Moffatt, Hans de Goede, | |
4683 | Harsh Barsaiyan, Henri Chain, Hristo Venev, Icenowy Zheng, Igor Zhbanov, | |
4684 | imayoda, Jakub Warczarek, James Buren, Jan Janssen, Jan Macku, | |
4685 | Jan Synacek, Jason Francis, Jayanth Ananthapadmanaban, Jeremy Szu, | |
4686 | Jérôme Carretero, Jesse Stricker, jiangchuangang, Joerg Behrmann, | |
4687 | Jóhann B. Guðmundsson, Jörg Deckert, Jörg Thalheim, Juergen Hoetzel, | |
4688 | Julia Kartseva, Kai-Heng Feng, Khem Raj, KoyamaSohei, laineantti, | |
4689 | Lennart Poettering, LetzteInstanz, Luca Adrian L, Luca Boccassi, | |
4690 | Lucas Magasweran, Mantas Mikulėnas, Marco Antonio Mauro, Mark Wielaard, | |
4691 | Masahiro Matsuya, Matt Johnston, Michael Catanzaro, Michal Koutný, | |
4692 | Michal Sekletár, Mike Crowe, Mike Kazantsev, Milan, milaq, | |
4693 | Miroslav Suchý, Morten Linderud, nerdopolis, nl6720, Noah Meyerhans, | |
4694 | Oleg Popov, Olle Lundberg, Ondrej Kozina, Paweł Marciniak, Perry.Yuan, | |
4695 | Peter Hutterer, Peter Kjellerstedt, Peter Morrow, Phaedrus Leeds, | |
4696 | plattrap, qhill, Raul Tambre, Roman Beranek, Roshan Shariff, | |
4697 | Ryan Hendrickson, Samuel BF, scootergrisen, Sebastian Blunt, | |
4698 | Seong-ho Cho, Sergey Bugaev, Sevan Janiyan, Sibo Dong, simmon, | |
4699 | Simon Watts, Srinidhi Kaushik, Štěpán Němec, Steve Bonds, Susant Sahani, | |
4700 | sverdlin, syyhao1994, Takashi Sakamoto, Topi Miettinen, tramsay, | |
4701 | Trent Piepho, Uwe Kleine-König, Viktor Mihajlovski, Vincent Dechenaux, | |
4702 | Vito Caputo, William A. Kennington III, Yangyang Shen, Yegor Alexeyev, | |
4703 | Yi Gao, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, zsien, наб | |
6969135f | 4704 | |
f6278558 | 4705 | — Edinburgh, 2021-07-07 |
99c2a955 | 4706 | |
94293d65 LB |
4707 | CHANGES WITH 248: |
4708 | ||
6dd990f3 ZJS |
4709 | * A concept of system extension images is introduced. Such images may |
4710 | be used to extend the /usr/ and /opt/ directory hierarchies at | |
4711 | runtime with additional files (even if the file system is read-only). | |
4712 | When a system extension image is activated, its /usr/ and /opt/ | |
4713 | hierarchies and os-release information are combined via overlayfs | |
4714 | with the file system hierarchy of the host OS. | |
4715 | ||
4716 | A new systemd-sysext tool can be used to merge, unmerge, list, and | |
4717 | refresh system extension hierarchies. See | |
4718 | https://www.freedesktop.org/software/systemd/man/systemd-sysext.html. | |
4719 | ||
4720 | The systemd-sysext.service automatically merges installed system | |
4721 | extensions during boot (before basic.target, but not in very early | |
4722 | boot, since various file systems have to be mounted first). | |
4723 | ||
4724 | The SYSEXT_LEVEL= field in os-release(5) may be used to specify the | |
4725 | supported system extension level. | |
4726 | ||
873b5cbd LB |
4727 | * A new ExtensionImages= unit setting can be used to apply the same |
4728 | system extension image concept from systemd-sysext to the namespaced | |
4729 | file hierarchy of specific services, following the same rules and | |
4730 | constraints. | |
4731 | ||
33e68cce ZJS |
4732 | * Support for a new special "root=tmpfs" kernel command-line option has |
4733 | been added. When specified, a tmpfs is mounted on /, and mount.usr= | |
4734 | should be used to point to the operating system implementation. | |
4735 | ||
6dd990f3 | 4736 | * A new configuration file /etc/veritytab may be used to configure |
2b6a8a4b LP |
4737 | dm-verity integrity protection for block devices. Each line is in the |
4738 | format "volume-name data-device hash-device roothash options", | |
4739 | similar to /etc/crypttab. | |
6dd990f3 | 4740 | |
2b6a8a4b | 4741 | * A new kernel command-line option systemd.verity.root_options= may be |
6dd990f3 ZJS |
4742 | used to configure dm-verity behaviour for the root device. |
4743 | ||
4744 | * The key file specified in /etc/crypttab (the third field) may now | |
1f3315b8 LP |
4745 | refer to an AF_UNIX/SOCK_STREAM socket in the file system. The key is |
4746 | acquired by connecting to that socket and reading from it. This | |
4747 | allows the implementation of a service to provide key information | |
4748 | dynamically, at the moment when it is needed. | |
6dd990f3 | 4749 | |
89b6acfd ZJS |
4750 | * When the hostname is set explicitly to "localhost", systemd-hostnamed |
4751 | will respect this. Previously such a setting would be mostly silently | |
4752 | ignored. The goal is to honour configuration as specified by the | |
4753 | user. | |
4754 | ||
4755 | * The fallback hostname that will be used by the system manager and | |
4756 | systemd-hostnamed can now be configured in two new ways: by setting | |
4757 | DEFAULT_HOSTNAME= in os-release(5), or by setting | |
4758 | $SYSTEMD_DEFAULT_HOSTNAME in the environment block. As before, it can | |
4759 | also be configured during compilation. The environment variable is | |
4760 | intended for testing and local overrides, the os-release(5) field is | |
4761 | intended to allow customization by different variants of a | |
4762 | distribution that share the same compiled packages. | |
4763 | ||
4764 | * The environment block of the manager itself may be configured through | |
4765 | a new ManagerEnvironment= setting in system.conf or user.conf. This | |
4766 | complements existing ways to set the environment block (the kernel | |
4767 | command line for the system manager, the inherited environment and | |
4768 | user@.service unit file settings for the user manager). | |
4769 | ||
4770 | * systemd-hostnamed now exports the default hostname and the source of | |
4771 | the configured hostname ("static", "transient", or "default") as | |
4772 | D-Bus properties. | |
4773 | ||
4774 | * systemd-hostnamed now exports the "HardwareVendor" and | |
4775 | "HardwareModel" D-Bus properties, which are supposed to contain a | |
4776 | pair of cleaned up, human readable strings describing the system's | |
4777 | vendor and model. It's typically sourced from the firmware's DMI | |
4778 | tables, but may be augmented from a new hwdb database. hostnamectl | |
4779 | shows this in the status output. | |
4780 | ||
1f3315b8 LP |
4781 | * Support has been added to systemd-cryptsetup for extracting the |
4782 | PKCS#11 token URI and encrypted key from the LUKS2 JSON embedded | |
4783 | metadata header. This allows the information how to open the | |
4784 | encrypted device to be embedded directly in the device and obviates | |
4785 | the need for configuration in an external file. | |
6dd990f3 | 4786 | |
1f3315b8 | 4787 | * systemd-cryptsetup gained support for unlocking LUKS2 volumes using |
2b6a8a4b LP |
4788 | TPM2 hardware, as well as FIDO2 security tokens (in addition to the |
4789 | pre-existing support for PKCS#11 security tokens). | |
6dd990f3 | 4790 | |
2b6a8a4b LP |
4791 | * systemd-repart may enroll encrypted partitions using TPM2 |
4792 | hardware. This may be useful for example to create an encrypted /var | |
4793 | partition bound to the machine on first boot. | |
6dd990f3 | 4794 | |
1f3315b8 LP |
4795 | * A new systemd-cryptenroll tool has been added to enroll TPM2, FIDO2 |
4796 | and PKCS#11 security tokens to LUKS volumes, list and destroy | |
4797 | them. See: | |
4798 | ||
dc7e580e | 4799 | https://0pointer.net/blog/unlocking-luks2-volumes-with-tpm2-fido2-pkcs11-security-hardware-on-systemd-248.html |
1f3315b8 LP |
4800 | |
4801 | It also supports enrolling "recovery keys" and regular passphrases. | |
4802 | ||
4803 | * The libfido2 dependency is now based on dlopen(), so that the library | |
2b6a8a4b LP |
4804 | is used at runtime when installed, but is not a hard runtime |
4805 | dependency. | |
1f3315b8 LP |
4806 | |
4807 | * systemd-cryptsetup gained support for two new options in | |
2b6a8a4b LP |
4808 | /etc/crypttab: "no-write-workqueue" and "no-read-workqueue" which |
4809 | request synchronous processing of encryption/decryption IO. | |
6dd990f3 | 4810 | |
2b6a8a4b LP |
4811 | * The manager may be configured at compile time to use the fexecve() |
4812 | instead of the execve() system call when spawning processes. Using | |
4813 | fexecve() closes a window between checking the security context of an | |
4814 | executable and spawning it, but unfortunately the kernel displays | |
4815 | stale information in the process' "comm" field, which impacts ps | |
4816 | output and such. | |
6dd990f3 ZJS |
4817 | |
4818 | * The configuration option -Dcompat-gateway-hostname has been dropped. | |
4819 | "_gateway" is now the only supported name. | |
4820 | ||
1f3315b8 LP |
4821 | * The ConditionSecurity=tpm2 unit file setting may be used to check if |
4822 | the system has at least one TPM2 (tpmrm class) device. | |
6dd990f3 | 4823 | |
2b6a8a4b LP |
4824 | * A new ConditionCPUFeature= has been added that may be used to |
4825 | conditionalize units based on CPU features. For example, | |
4826 | ConditionCPUFeature=rdrand will condition a unit so that it is only | |
4827 | run when the system CPU supports the RDRAND opcode. | |
4828 | ||
89b6acfd ZJS |
4829 | * The existing ConditionControlGroupController= setting has been |
4830 | extended with two new values "v1" and "v2". "v2" means that the | |
b49bb286 | 4831 | unified v2 cgroup hierarchy is used, and "v1" means that legacy v1 |
89b6acfd ZJS |
4832 | hierarchy or the hybrid hierarchy are used. |
4833 | ||
33e68cce ZJS |
4834 | * A new PrivateIPC= setting on a unit file allows executed processes to |
4835 | be moved into a private IPC namespace, with separate System V IPC | |
4836 | identifiers and POSIX message queues. | |
4837 | ||
4838 | A new IPCNamespacePath= allows the unit to be joined to an existing | |
4839 | IPC namespace. | |
4840 | ||
64297c86 | 4841 | * The tables of system calls in seccomp filters are now automatically |
6dd990f3 ZJS |
4842 | generated from kernel lists exported on |
4843 | https://fedora.juszkiewicz.com.pl/syscalls.html. | |
4844 | ||
4845 | The following architectures should now have complete lists: | |
4846 | alpha, arc, arm64, arm, i386, ia64, m68k, mips64n32, mips64, mipso32, | |
4847 | powerpc, powerpc64, s390, s390x, tilegx, sparc, x86_64, x32. | |
4848 | ||
94293d65 | 4849 | * The MountAPIVFS= service file setting now additionally mounts a tmpfs |
6dd990f3 ZJS |
4850 | on /run/ if it is not already a mount point. A writable /run/ has |
4851 | always been a requirement for a functioning system, but this was not | |
94293d65 | 4852 | guaranteed when using a read-only image. |
6dd990f3 ZJS |
4853 | |
4854 | Users can always specify BindPaths= or InaccessiblePaths= as | |
4855 | overrides, and they will take precedence. If the host's root mount | |
4856 | point is used, there is no change in behaviour. | |
4857 | ||
4858 | * New bind mounts and file system image mounts may be injected into the | |
4859 | mount namespace of a service (without restarting it). This is exposed | |
9ba008cb LB |
4860 | respectively as 'systemctl bind <unit> <path>…' and |
4861 | 'systemctl mount-image <unit> <image>…'. | |
6dd990f3 | 4862 | |
2b6a8a4b | 4863 | * The StandardOutput= and StandardError= settings can now specify files |
6dd990f3 ZJS |
4864 | to be truncated for output (as "truncate:<path>"). |
4865 | ||
4866 | * The ExecPaths= and NoExecPaths= settings may be used to specify | |
4867 | noexec for parts of the file system. | |
4868 | ||
1f3315b8 | 4869 | * sd-bus has a new function sd_bus_open_user_machine() to open a |
6dd990f3 | 4870 | connection to the session bus of a specific user in a local container |
2b6a8a4b LP |
4871 | or on the local host. This is exposed in the existing -M switch to |
4872 | systemctl and similar tools: | |
1f3315b8 LP |
4873 | |
4874 | systemctl --user -M lennart@foobar start foo | |
4875 | ||
4876 | This will connect to the user bus of a user "lennart" in container | |
4877 | "foobar". If no container name is specified, the specified user on | |
4878 | the host itself is connected to | |
4879 | ||
4880 | systemctl --user -M lennart@ start quux | |
6dd990f3 | 4881 | |
6c41cf44 LP |
4882 | * sd-bus also gained a convenience function sd_bus_message_send() to |
4883 | simplify invocations of sd_bus_send(), taking only a single | |
4884 | parameter: the message to send. | |
1f3315b8 LP |
4885 | |
4886 | * sd-event allows rate limits to be set on event sources, for dealing | |
4887 | with high-priority event sources that might starve out others. See | |
4888 | the new man page sd_event_source_set_ratelimit(3) for details. | |
6dd990f3 ZJS |
4889 | |
4890 | * systemd.link files gained a [Link] Promiscuous= switch, which allows | |
4891 | the device to be raised in promiscuous mode. | |
4892 | ||
4893 | New [Link] TransmitQueues= and ReceiveQueues= settings allow the | |
4894 | number of TX and RX queues to be configured. | |
4895 | ||
4896 | New [Link] TransmitQueueLength= setting allows the size of the TX | |
4897 | queue to be configured. | |
4898 | ||
4899 | New [Link] GenericSegmentOffloadMaxBytes= and | |
4900 | GenericSegmentOffloadMaxSegments= allow capping the packet size and | |
4901 | the number of segments accepted in Generic Segment Offload. | |
4902 | ||
33e68cce ZJS |
4903 | * systemd-networkd gained support for the "B.A.T.M.A.N. advanced" |
4904 | wireless routing protocol that operates on ISO/OSI Layer 2 only and | |
4905 | uses ethernet frames to route/bridge packets. This encompasses a new | |
4906 | "batadv" netdev Type=, a new [BatmanAdvanced] section with a bunch of | |
4907 | new settings in .netdev files, and a new BatmanAdvanced= setting in | |
4908 | .network files. | |
4909 | ||
6dd990f3 ZJS |
4910 | * systemd.network files gained a [Network] RouteTable= configuration |
4911 | switch to select the routing policy table. | |
4912 | ||
4913 | systemd.network files gained a [RoutingPolicyRule] Type= | |
4914 | configuration switch (one of "blackhole, "unreachable", "prohibit"). | |
4915 | ||
4916 | systemd.network files gained a [IPv6AcceptRA] RouteDenyList= and | |
4917 | RouteAllowList= settings to ignore/accept route advertisements from | |
4918 | routers matching specified prefixes. The DenyList= setting has been | |
4919 | renamed to PrefixDenyList= and a new PrefixAllowList= option has been | |
4920 | added. | |
4921 | ||
4922 | systemd.network files gained a [DHCPv6] UseAddress= setting to | |
4923 | optionally ignore the address provided in the lease. | |
4924 | ||
4925 | systemd.network files gained a [DHCPv6PrefixDelegation] | |
4926 | ManageTemporaryAddress= switch. | |
4927 | ||
33e68cce ZJS |
4928 | systemd.network files gained a new ActivationPolicy= setting which |
4929 | allows configuring how the UP state of an interface shall be managed, | |
b480543c | 4930 | i.e. whether the interface is always upped, always downed, or may be |
6d18c13e | 4931 | upped/downed by the user using "ip link set dev". |
1f3315b8 | 4932 | |
897a2561 ZJS |
4933 | * The default for the Broadcast= setting in .network files has slightly |
4934 | changed: the broadcast address will not be configured for wireguard | |
4935 | devices. | |
4936 | ||
6dd990f3 ZJS |
4937 | * systemd.netdev files gained a [VLAN] Protocol=, IngressQOSMaps=, |
4938 | EgressQOSMaps=, and [MACVLAN] BroadcastMulticastQueueLength= | |
4939 | configuration options for VLAN packet handling. | |
4940 | ||
4941 | * udev rules may now set log_level= option. This allows debug logs to | |
4942 | be enabled for select events, e.g. just for a specific subsystem or | |
4943 | even a single device. | |
4944 | ||
4945 | * udev now exports the VOLUME_ID, LOGICAL_VOLUME_ID, VOLUME_SET_ID, and | |
2b6a8a4b LP |
4946 | DATA_PREPARED_ID properties for block devices with ISO9660 file |
4947 | systems. | |
6dd990f3 | 4948 | |
2b6a8a4b LP |
4949 | * udev now exports decoded DMI information about installed memory slots |
4950 | as device properties under the /sys/class/dmi/id/ pseudo device. | |
6dd990f3 | 4951 | |
2b6a8a4b | 4952 | * /dev/ is not mounted noexec anymore. This didn't provide any |
64297c86 TM |
4953 | significant security benefits and would conflict with the executable |
4954 | mappings used with /dev/sgx device nodes. The previous behaviour can | |
4955 | be restored for individual services with NoExecPaths=/dev (or by allow- | |
4956 | listing and excluding /dev from ExecPaths=). | |
6dd990f3 | 4957 | |
de0b8991 | 4958 | * Permissions for /dev/vsock are now set to 0o666, and /dev/vhost-vsock |
6dd990f3 ZJS |
4959 | and /dev/vhost-net are owned by the kvm group. |
4960 | ||
4961 | * The hardware database has been extended with a list of fingerprint | |
2b6a8a4b LP |
4962 | readers that correctly support USB auto-suspend using data from |
4963 | libfprint. | |
6dd990f3 ZJS |
4964 | |
4965 | * systemd-resolved can now answer DNSSEC questions through the stub | |
4966 | resolver interface in a way that allows local clients to do DNSSEC | |
4967 | validation themselves. For a question with DO+CD set, it'll proxy the | |
4968 | DNS query and respond with a mostly unmodified packet received from | |
4969 | the upstream server. | |
4970 | ||
1f3315b8 LP |
4971 | * systemd-resolved learnt a new boolean option CacheFromLocalhost= in |
4972 | resolved.conf. If true the service will provide caching even for DNS | |
4973 | lookups made to an upstream DNS server on the 127.0.0.1/::1 | |
4974 | addresses. By default (and when the option is false) systemd-resolved | |
4975 | will not cache such lookups, in order to avoid duplicate local | |
4976 | caching, under the assumption the local upstream server caches | |
4977 | anyway. | |
4978 | ||
2b6a8a4b LP |
4979 | * systemd-resolved now implements RFC5001 NSID in its local DNS |
4980 | stub. This may be used by local clients to determine whether they are | |
4981 | talking to the DNS resolver stub or a different DNS server. | |
4982 | ||
4983 | * When resolving host names and other records resolvectl will now | |
4984 | report where the data was acquired from (i.e. the local cache, the | |
4985 | network, locally synthesized, …) and whether the network traffic it | |
4986 | effected was encrypted or not. Moreover the tool acquired a number of | |
4987 | new options --cache=, --synthesize=, --network=, --zone=, | |
4988 | --trust-anchor=, --validate= that take booleans and may be used to | |
4989 | tweak a lookup, i.e. whether it may be answered from cached | |
4990 | information, locally synthesized information, information acquired | |
4991 | through the network, the local mDNS/LLMNR zone, the DNSSEC trust | |
4992 | anchor, and whether DNSSEC validation shall be executed for the | |
4993 | lookup. | |
4994 | ||
9ba008cb | 4995 | * systemd-nspawn gained a new --ambient-capability= setting |
6dd990f3 ZJS |
4996 | (AmbientCapability= in .nspawn files) to configure ambient |
4997 | capabilities passed to the container payload. | |
4998 | ||
4999 | * systemd-nspawn gained the ability to configure the firewall using the | |
1f3315b8 | 5000 | nftables subsystem (in addition to the existing iptables |
897a2561 | 5001 | support). Similarly, systemd-networkd's IPMasquerade= option now |
2b6a8a4b LP |
5002 | supports nftables as back-end, too. In both cases NAT on IPv6 is now |
5003 | supported too, in addition to IPv4 (the iptables back-end still is | |
1f3315b8 LP |
5004 | IPv4-only). |
5005 | ||
897a2561 ZJS |
5006 | "IPMasquerade=yes", which was the same as "IPMasquerade=ipv4" before, |
5007 | retains its meaning, but has been deprecated. Please switch to either | |
5008 | "ivp4" or "both" (if covering IPv6 is desired). | |
5009 | ||
1f3315b8 LP |
5010 | * systemd-importd will now download .verity and .roothash.p7s files |
5011 | along with the machine image (as exposed via machinectl pull-raw). | |
6dd990f3 ZJS |
5012 | |
5013 | * systemd-oomd now gained a new DefaultMemoryPressureDurationSec= | |
5014 | setting to configure the time a unit's cgroup needs to exceed memory | |
897a2561 ZJS |
5015 | pressure limits before action will be taken, and a new |
5016 | ManagedOOMPreference=none|avoid|omit setting to avoid killing certain | |
5017 | units. | |
6dd990f3 ZJS |
5018 | |
5019 | systemd-oomd is now considered fully supported (the usual | |
6b5e8240 | 5020 | backwards-compatibility promises apply). Swap is not required for |
6dd990f3 ZJS |
5021 | operation, but it is still recommended. |
5022 | ||
5023 | * systemd-timesyncd gained a new ConnectionRetrySec= setting which | |
5024 | configures the retry delay when trying to contact servers. | |
5025 | ||
5026 | * systemd-stdio-bridge gained --system/--user options to connect to the | |
5027 | system bus (previous default) or the user session bus. | |
5028 | ||
6dd990f3 ZJS |
5029 | * systemd-localed may now call locale-gen to generate missing locales |
5030 | on-demand (UTF-8-only). This improves integration with Debian-based | |
5031 | distributions (Debian/Ubuntu/PureOS/Tanglu/...) and Arch Linux. | |
5032 | ||
1f3315b8 LP |
5033 | * systemctl --check-inhibitors=true may now be used to obey inhibitors |
5034 | even when invoked non-interactively. The old --ignore-inhibitors | |
5035 | switch is now deprecated and replaced by --check-inhibitors=false. | |
6dd990f3 ZJS |
5036 | |
5037 | * systemctl import-environment will now emit a warning when called | |
5038 | without any arguments (i.e. to import the full environment block of | |
5039 | the called program). This command will usually be invoked from a | |
5040 | shell, which means that it'll inherit a bunch of variables which are | |
5041 | specific to that shell, and usually to the TTY the shell is connected | |
5042 | to, and don't have any meaning in the global context of the system or | |
5043 | user service manager. Instead, only specific variables should be | |
5044 | imported into the manager environment block. | |
5045 | ||
5046 | Similarly, programs which update the manager environment block by | |
5047 | directly calling the D-Bus API of the manager, should also push | |
5048 | specific variables, and not the full inherited environment. | |
5049 | ||
1f3315b8 | 5050 | * systemctl's status output now shows unit state with a more careful |
2b6a8a4b LP |
5051 | choice of Unicode characters: units in maintenance show a "○" symbol |
5052 | instead of the usual "●", failed units show "×", and services being | |
5053 | reloaded "↻". | |
1f3315b8 | 5054 | |
6dd990f3 | 5055 | * coredumpctl gained a --debugger-arguments= switch to pass arguments |
2b6a8a4b LP |
5056 | to the debugger. It also gained support for showing coredump info in |
5057 | a simple JSON format. | |
5058 | ||
5059 | * systemctl/loginctl/machinectl's --signal= option now accept a special | |
5060 | value "list", which may be used to show a brief table with known | |
5061 | process signals and their numbers. | |
6dd990f3 ZJS |
5062 | |
5063 | * networkctl now shows the link activation policy in status. | |
5064 | ||
2b6a8a4b | 5065 | * Various tools gained --pager/--no-pager/--json= switches to |
6dd990f3 ZJS |
5066 | enable/disable the pager and provide JSON output. |
5067 | ||
2b6a8a4b LP |
5068 | * Various tools now accept two new values for the SYSTEMD_COLORS |
5069 | environment variable: "16" and "256", to configure how many terminal | |
5070 | colors are used in output. | |
6dd990f3 | 5071 | |
2b6a8a4b LP |
5072 | * less 568 or newer is now required for the auto-paging logic of the |
5073 | various tools. Hyperlink ANSI sequences in terminal output are now | |
5074 | used even if a pager is used, and older versions of less are not able | |
5075 | to display these sequences correctly. SYSTEMD_URLIFY=0 may be used to | |
5076 | disable this output again. | |
6dd990f3 | 5077 | |
2b6a8a4b | 5078 | * Builds with support for separate / and /usr/ hierarchies ("split-usr" |
6dd990f3 ZJS |
5079 | builds, non-merged-usr builds) are now officially deprecated. A |
5080 | warning is emitted during build. Support is slated to be removed in | |
5081 | about a year (when the Debian Bookworm release development starts). | |
5082 | ||
897a2561 ZJS |
5083 | * Systems with the legacy cgroup v1 hierarchy are now marked as |
5084 | "tainted", to make it clearer that using the legacy hierarchy is not | |
5085 | recommended. | |
5086 | ||
33e68cce ZJS |
5087 | * systemd-localed will now refuse to configure a keymap which is not |
5088 | installed in the file system. This is intended as a bug fix, but | |
5089 | could break cases where systemd-localed was used to configure the | |
5090 | keymap in advanced of it being installed. It is necessary to install | |
5091 | the keymap file first. | |
5092 | ||
2b6a8a4b | 5093 | * The main git development branch has been renamed to 'main'. |
94293d65 | 5094 | |
26c59e4e LB |
5095 | * mmcblk[0-9]boot[0-9] devices will no longer be probed automatically |
5096 | for partitions, as in the vast majority of cases they contain none | |
5097 | and are used internally by the bootloader (eg: uboot). | |
5098 | ||
1f3315b8 LP |
5099 | * systemd will now set the $SYSTEMD_EXEC_PID environment variable for |
5100 | spawned processes to the PID of the process itself. This may be used | |
5101 | by programs for detecting whether they were forked off by the service | |
5102 | manager itself or are a process forked off further down the tree. | |
5103 | ||
897a2561 ZJS |
5104 | * The sd-device API gained four new calls: sd_device_get_action() to |
5105 | determine the uevent add/remove/change/… action the device object has | |
5106 | been seen for, sd_device_get_seqno() to determine the uevent sequence | |
5107 | number, sd_device_new_from_stat_rdev() to allocate a new sd_device | |
5108 | object from stat(2) data of a device node, and sd_device_trigger() to | |
5109 | write to the 'uevent' attribute of a device. | |
2b6a8a4b LP |
5110 | |
5111 | * For most tools the --no-legend= switch has been replaced by | |
5112 | --legend=no and --legend=yes, to force whether tables are shown with | |
5113 | headers/legends. | |
5114 | ||
5115 | * Units acquired a new property "Markers" that takes a list of zero, | |
5116 | one or two of the following strings: "needs-reload" and | |
5117 | "needs-restart". These markers may be set via "systemctl | |
5118 | set-property". Once a marker is set, "systemctl reload-or-restart | |
5119 | --marked" may be invoked to execute the operation the units are | |
5120 | marked for. This is useful for package managers that want to mark | |
5121 | units for restart/reload while updating, but effect the actual | |
5122 | operations at a later step at once. | |
5123 | ||
5124 | * The sd_bus_message_read_strv() API call of sd-bus may now also be | |
5125 | used to parse arrays of D-Bus signatures and D-Bus paths, in addition | |
5126 | to regular strings. | |
5127 | ||
5128 | * bootctl will now report whether the UEFI firmware used a TPM2 device | |
5129 | and measured the boot process into it. | |
5130 | ||
5131 | * systemd-tmpfiles learnt support for a new environment variable | |
5132 | $SYSTEMD_TMPFILES_FORCE_SUBVOL which takes a boolean value. If true | |
5133 | the v/q/Q lines in tmpfiles.d/ snippets will create btrfs subvolumes | |
5134 | even if the root fs of the system is not itself a btrfs volume. | |
5135 | ||
5136 | * systemd-detect-virt/ConditionVirtualization= will now explicitly | |
5137 | detect Docker/Podman environments where possible. Moreover, they | |
5138 | should be able to generically detect any container manager as long as | |
5139 | it assigns the container a cgroup. | |
5140 | ||
5141 | * portablectl gained a new "reattach" verb for detaching/reattaching a | |
5142 | portable service image, useful for updating images on-the-fly. | |
5143 | ||
d94d261c LP |
5144 | * Intel SGX enclave device nodes (which expose a security feature of |
5145 | newer Intel CPUs) will now be owned by a new system group "sgx". | |
5146 | ||
8ce22551 ZJS |
5147 | Contributions from: Adam Nielsen, Adrian Vovk, AJ Jordan, Alan Perry, |
5148 | Alastair Pharo, Alexander Batischev, Ali Abdallah, Andrew Balmos, | |
5149 | Anita Zhang, Annika Wickert, Ansgar Burchardt, Antonio Terceiro, | |
5150 | Antonius Frie, Ardy, Arian van Putten, Ariel Fermani, Arnaud T, | |
5151 | A S Alam, Bastien Nocera, Benjamin Berg, Benjamin Robin, Björn Daase, | |
5152 | caoxia, Carlo Wood, Charles Lee, ChopperRob, chri2, Christian Ehrhardt, | |
5153 | Christian Hesse, Christopher Obbard, clayton craft, corvusnix, cprn, | |
5154 | Daan De Meyer, Daniele Medri, Daniel Rusek, Dan Sanders, Dan Streetman, | |
5155 | Darren Ng, David Edmundson, David Tardon, Deepak Rawat, Devon Pringle, | |
5156 | Dmitry Borodaenko, dropsignal, Einsler Lee, Endre Szabo, | |
5157 | Evgeny Vereshchagin, Fabian Affolter, Fangrui Song, Felipe Borges, | |
5158 | feliperodriguesfr, Felix Stupp, Florian Hülsmann, Florian Klink, | |
5159 | Florian Westphal, Franck Bui, Frantisek Sumsal, Gablegritule, | |
5160 | Gaël PORTAY, Gaurav, Giedrius Statkevičius, Greg Depoire-Ferrer, | |
5161 | Gustavo Costa, Hans de Goede, Hela Basa, heretoenhance, hide, | |
5162 | Iago López Galeiras, igo95862, Ilya Dmitrichenko, Jameer Pathan, | |
5163 | Jan Tojnar, Jiehong, Jinyuan Si, Joerg Behrmann, John Slade, | |
5164 | Jonathan G. Underwood, Jonathan McDowell, Josh Triplett, Joshua Watt, | |
5165 | Julia Cartwright, Julien Humbert, Kairui Song, Karel Zak, | |
5166 | Kevin Backhouse, Kevin P. Fleming, Khem Raj, Konomi, krissgjeng, | |
5167 | l4gfcm, Lajos Veres, Lennart Poettering, Lincoln Ramsay, Luca Boccassi, | |
5168 | Luca BRUNO, Lucas Werkmeister, Luka Kudra, Luna Jernberg, | |
5169 | Marc-André Lureau, Martin Wilck, Matthias Klumpp, Matt Turner, | |
5170 | Michael Gisbers, Michael Marley, Michael Trapp, Michal Fabik, | |
5171 | Michał Kopeć, Michal Koutný, Michal Sekletár, Michele Guerini Rocco, | |
5172 | Mike Gilbert, milovlad, moson-mo, Nick, nihilix-melix, Oğuz Ersen, | |
5173 | Ondrej Mosnacek, pali, Pavel Hrdina, Pavel Sapezhko, Perry Yuan, | |
5174 | Peter Hutterer, Pierre Dubouilh, Piotr Drąg, Pjotr Vertaalt, | |
5175 | Richard Laager, RussianNeuroMancer, Sam Lunt, Sebastiaan van Stijn, | |
5176 | Sergey Bugaev, shenyangyang4, simmon, Simonas Kazlauskas, | |
5177 | Slimane Selyan Amiri, Stefan Agner, Steve Ramage, Susant Sahani, | |
5178 | Sven Mueller, Tad Fisher, Takashi Iwai, Thomas Haller, Tom Shield, | |
5179 | Topi Miettinen, Torsten Hilbrich, tpgxyz, Tyler Hicks, ulf-f, | |
5180 | Ulrich Ölmann, Vincent Pelletier, Vinnie Magro, Vito Caputo, Vlad, | |
5181 | walbit-de, Whired Planck, wouter bolsterlee, Xℹ Ruoyao, Yangyang Shen, | |
5182 | Yuri Chornoivan, Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek, | |
5183 | Zmicer Turok, Дамјан Георгиевски | |
d94d261c | 5184 | |
9ca34cf5 | 5185 | — Berlin, 2021-03-30 |
60d31370 | 5186 | |
d0dcf59b | 5187 | CHANGES WITH 247: |
bf6e5c57 | 5188 | |
d90922fb | 5189 | * KERNEL API INCOMPATIBILITY: Linux 4.14 introduced two new uevents |
bf6e5c57 LP |
5190 | "bind" and "unbind" to the Linux device model. When this kernel |
5191 | change was made, systemd-udevd was only minimally updated to handle | |
5192 | and propagate these new event types. The introduction of these new | |
5193 | uevents (which are typically generated for USB devices and devices | |
5194 | needing a firmware upload before being functional) resulted in a | |
b182195a ZJS |
5195 | number of issues which we so far didn't address. We hoped the kernel |
5196 | maintainers would themselves address these issues in some form, but | |
5197 | that did not happen. To handle them properly, many (if not most) udev | |
5198 | rules files shipped in various packages need updating, and so do many | |
5199 | programs that monitor or enumerate devices with libudev or sd-device, | |
5200 | or otherwise process uevents. Please note that this incompatibility | |
5201 | is not fault of systemd or udev, but caused by an incompatible kernel | |
832eedd1 | 5202 | change that happened back in Linux 4.14, but is becoming more and |
dc6a3162 | 5203 | more visible as the new uevents are generated by more kernel drivers. |
bf6e5c57 LP |
5204 | |
5205 | To minimize issues resulting from this kernel change (but not avoid | |
5206 | them entirely) starting with systemd-udevd 247 the udev "tags" | |
5207 | concept (which is a concept for marking and filtering devices during | |
5208 | enumeration and monitoring) has been reworked: udev tags are now | |
5209 | "sticky", meaning that once a tag is assigned to a device it will not | |
5210 | be removed from the device again until the device itself is removed | |
5211 | (i.e. unplugged). This makes sure that any application monitoring | |
5212 | devices that match a specific tag is guaranteed to both see uevents | |
5213 | where the device starts being relevant, and those where it stops | |
5214 | being relevant (the latter now regularly happening due to the new | |
5215 | "unbind" uevent type). The udev tags concept is hence now a concept | |
5216 | tied to a *device* instead of a device *event* — unlike for example | |
5217 | udev properties whose lifecycle (as before) is generally tied to a | |
5218 | device event, meaning that the previously determined properties are | |
5219 | forgotten whenever a new uevent is processed. | |
5220 | ||
5221 | With the newly redefined udev tags concept, sometimes it's necessary | |
5222 | to determine which tags are the ones applied by the most recent | |
5223 | uevent/database update, in order to discern them from those | |
5224 | originating from earlier uevents/database updates of the same | |
5225 | device. To accommodate for this a new automatic property CURRENT_TAGS | |
5226 | has been added that works similar to the existing TAGS property but | |
5227 | only lists tags set by the most recent uevent/database | |
b182195a ZJS |
5228 | update. Similarly, the libudev/sd-device API has been updated with |
5229 | new functions to enumerate these 'current' tags, in addition to the | |
bf6e5c57 LP |
5230 | existing APIs that now enumerate the 'sticky' ones. |
5231 | ||
832eedd1 | 5232 | To properly handle "bind"/"unbind" on Linux 4.14 and newer it is |
bf6e5c57 LP |
5233 | essential that all udev rules files and applications are updated to |
5234 | handle the new events. Specifically: | |
5235 | ||
5236 | • All rule files that currently use a header guard similar to | |
5237 | ACTION!="add|change",GOTO="xyz_end" should be updated to use | |
5238 | ACTION=="remove",GOTO="xyz_end" instead, so that the | |
5239 | properties/tags they add are also applied whenever "bind" (or | |
5240 | "unbind") is seen. (This is most important for all physical device | |
b182195a ZJS |
5241 | types — those for which "bind" and "unbind" are currently |
5242 | generated, for all other device types this change is still | |
bf6e5c57 LP |
5243 | recommended but not as important — but certainly prepares for |
5244 | future kernel uevent type additions). | |
5245 | ||
b182195a | 5246 | • Similarly, all code monitoring devices that contains an 'if' branch |
bf6e5c57 LP |
5247 | discerning the "add" + "change" uevent actions from all other |
5248 | uevents actions (i.e. considering devices only relevant after "add" | |
5249 | or "change", and irrelevant on all other events) should be reworked | |
5250 | to instead negatively check for "remove" only (i.e. considering | |
5251 | devices relevant after all event types, except for "remove", which | |
5252 | invalidates the device). Note that this also means that devices | |
5253 | should be considered relevant on "unbind", even though conceptually | |
5254 | this — in some form — invalidates the device. Since the precise | |
5255 | effect of "unbind" is not generically defined, devices should be | |
5256 | considered relevant even after "unbind", however I/O errors | |
5257 | accessing the device should then be handled gracefully. | |
5258 | ||
5259 | • Any code that uses device tags for deciding whether a device is | |
5260 | relevant or not most likely needs to be updated to use the new | |
5261 | udev_device_has_current_tag() API (or sd_device_has_current_tag() | |
db2db708 LP |
5262 | in case sd-device is used), to check whether the tag is set at the |
5263 | moment an uevent is seen (as opposed to the existing | |
bf6e5c57 LP |
5264 | udev_device_has_tag() API which checks if the tag ever existed on |
5265 | the device, following the API concept redefinition explained | |
5266 | above). | |
5267 | ||
5268 | We are very sorry for this breakage and the requirement to update | |
5269 | packages using these interfaces. We'd again like to underline that | |
5270 | this is not caused by systemd/udev changes, but result of a kernel | |
5271 | behaviour change. | |
5272 | ||
7bfcc0de LP |
5273 | * UPCOMING INCOMPATIBILITY: So far most downstream distribution |
5274 | packages have not retriggered devices once the udev package (or any | |
5275 | auxiliary package installing additional udev rules) is updated. We | |
5276 | intend to work with major distributions to change this, so that | |
75271582 | 5277 | "udevadm trigger -c change" is issued on such upgrades, ensuring that |
7bfcc0de LP |
5278 | the updated ruleset is applied to the devices already discovered, so |
5279 | that (asynchronously) after the upgrade completed the udev database | |
5280 | is consistent with the updated rule set. This means udev rules must | |
5281 | be ready to be retriggered with a "change" action any time, and | |
5282 | result in correct and complete udev database entries. While the | |
5283 | majority of udev rule files known to us currently get this right, | |
5284 | some don't. Specifically, there are udev rules files included in | |
5285 | various packages that only set udev properties on the "add" action, | |
5286 | but do not handle the "change" action. If a device matching those | |
5287 | rules is retriggered with the "change" action (as is intended here) | |
5288 | it would suddenly lose the relevant properties. This always has been | |
48e0abb5 YW |
5289 | problematic, but as soon as all udev devices are triggered on relevant |
5290 | package upgrades this will become particularly so. It is strongly | |
5291 | recommended to fix offending rules so that they can handle a "change" | |
5292 | action at any time, and acquire all necessary udev properties even | |
5293 | then. Or in other words: the header guard mentioned above | |
5294 | (ACTION=="remove",GOTO="xyz_end") is the correct approach to handle | |
5295 | this, as it makes sure rules are rerun on "change" correctly, and | |
5296 | accumulate the correct and complete set of udev properties. udev rule | |
5297 | definitions that cannot handle "change" events being triggered at | |
5298 | arbitrary times should be considered buggy. | |
7bfcc0de | 5299 | |
db2db708 | 5300 | * The MountAPIVFS= service file setting now defaults to on if |
b182195a ZJS |
5301 | RootImage= and RootDirectory= are used, which means that with those |
5302 | two settings /proc/, /sys/ and /dev/ are automatically properly set | |
5303 | up for services. Previous behaviour may be restored by explicitly | |
5304 | setting MountAPIVFS=off. | |
db2db708 | 5305 | |
bae66f4b LP |
5306 | * Since PAM 1.2.0 (2015) configuration snippets may be placed in |
5307 | /usr/lib/pam.d/ in addition to /etc/pam.d/. If a file exists in the | |
5308 | latter it takes precedence over the former, similar to how most of | |
5309 | systemd's own configuration is handled. Given that PAM stack | |
5310 | definitions are primarily put together by OS vendors/distributions | |
69e3234d | 5311 | (though possibly overridden by users), this systemd release moves its |
bae66f4b LP |
5312 | own PAM stack configuration for the "systemd-user" PAM service (i.e. |
5313 | for the PAM session invoked by the per-user user@.service instance) | |
5314 | from /etc/pam.d/ to /usr/lib/pam.d/. We recommend moving all | |
5315 | packages' vendor versions of their PAM stack definitions from | |
5316 | /etc/pam.d/ to /usr/lib/pam.d/, but if such OS-wide migration is not | |
5317 | desired the location to which systemd installs its PAM stack | |
b182195a | 5318 | configuration may be changed via the -Dpamconfdir Meson option. |
bae66f4b | 5319 | |
dc6a3162 LP |
5320 | * The runtime dependencies on libqrencode, libpcre2, libidn/libidn2, |
5321 | libpwquality and libcryptsetup have been changed to be based on | |
5322 | dlopen(): instead of regular dynamic library dependencies declared in | |
5323 | the binary ELF headers, these libraries are now loaded on demand | |
5324 | only, if they are available. If the libraries cannot be found the | |
5325 | relevant operations will fail gracefully, or a suitable fallback | |
5326 | logic is chosen. This is supposed to be useful for general purpose | |
5327 | distributions, as it allows minimizing the list of dependencies the | |
5328 | systemd packages pull in, permitting building of more minimal OS | |
5329 | images, while still making use of these "weak" dependencies should | |
5330 | they be installed. Since many package managers automatically | |
5331 | synthesize package dependencies from ELF shared library dependencies, | |
5332 | some additional manual packaging work has to be done now to replace | |
5333 | those (slightly downgraded from "required" to "recommended" or | |
5334 | whatever is conceptually suitable for the package manager). Note that | |
5335 | this change does not alter build-time behaviour: as before the | |
5336 | build-time dependencies have to be installed during build, even if | |
5337 | they now are optional during runtime. | |
db2db708 LP |
5338 | |
5339 | * sd-event.h gained a new call sd_event_add_time_relative() for | |
5340 | installing timers relative to the current time. This is mostly a | |
5341 | convenience wrapper around the pre-existing sd_event_add_time() call | |
5342 | which installs absolute timers. | |
5343 | ||
dc6a3162 LP |
5344 | * sd-event event sources may now be placed in a new "exit-on-failure" |
5345 | mode, which may be controlled via the new | |
5346 | sd_event_source_get_exit_on_failure() and | |
5347 | sd_event_source_set_exit_on_failure() functions. If enabled, any | |
5348 | failure returned by the event source handler functions will result in | |
5349 | exiting the event loop (unlike the default behaviour of just | |
5350 | disabling the event source but continuing with the event loop). This | |
5351 | feature is useful to set for all event sources that define "primary" | |
5352 | program behaviour (where failure should be fatal) in contrast to | |
5353 | "auxiliary" behaviour (where failure should remain local). | |
5354 | ||
5355 | * Most event source types sd-event supports now accept a NULL handler | |
5356 | function, in which case the event loop is exited once the event | |
5357 | source is to be dispatched, using the userdata pointer — converted to | |
5358 | a signed integer — as exit code of the event loop. Previously this | |
5359 | was supported for IO and signal event sources already. Exit event | |
5360 | sources still do not support this (simply because it makes little | |
5361 | sense there, as the event loop is already exiting when they are | |
5362 | dispatched). | |
5363 | ||
db2db708 LP |
5364 | * A new per-unit setting RootImageOptions= has been added which allows |
5365 | tweaking the mount options for any file system mounted as effect of | |
5366 | the RootImage= setting. | |
5367 | ||
5368 | * Another new per-unit setting MountImages= has been added, that allows | |
5369 | mounting additional disk images into the file system tree accessible | |
5370 | to the service. | |
5371 | ||
6fc5b951 | 5372 | * Timer units gained a new FixedRandomDelay= boolean setting. If |
b5e798de ZJS |
5373 | enabled, the random delay configured with RandomizedDelaySec= is |
5374 | selected in a way that is stable on a given system (though still | |
5375 | different for different units). | |
6fc5b951 LP |
5376 | |
5377 | * Socket units gained a new setting Timestamping= that takes "us", "ns" | |
5378 | or "off". This controls the SO_TIMESTAMP/SO_TIMESTAMPNS socket | |
5379 | options. | |
5380 | ||
b182195a ZJS |
5381 | * systemd-repart now generates JSON output when requested with the new |
5382 | --json= switch. | |
db2db708 LP |
5383 | |
5384 | * systemd-machined's OpenMachineShell() bus call will now pass | |
5385 | additional policy metadata data fields to the PolicyKit | |
5386 | authentication request. | |
5387 | ||
5388 | * systemd-tmpfiles gained a new -E switch, which is equivalent to | |
5389 | --exclude-prefix=/dev --exclude-prefix=/proc --exclude=/run | |
5390 | --exclude=/sys. It's particularly useful in combination with --root=, | |
5391 | when operating on OS trees that do not have any of these four runtime | |
5392 | directories mounted, as this means no files below these subtrees are | |
5393 | created or modified, since those mount points should probably remain | |
5394 | empty. | |
5395 | ||
b182195a ZJS |
5396 | * systemd-tmpfiles gained a new --image= switch which is like --root=, |
5397 | but takes a disk image instead of a directory as argument. The | |
5398 | specified disk image is mounted inside a temporary mount namespace | |
5399 | and the tmpfiles.d/ drop-ins stored in the image are executed and | |
5400 | applied to the image. systemd-sysusers similarly gained a new | |
5401 | --image= switch, that allows the sysusers.d/ drop-ins stored in the | |
5402 | image to be applied onto the image. | |
db2db708 | 5403 | |
b182195a ZJS |
5404 | * Similarly, the journalctl command also gained an --image= switch, |
5405 | which is a quick one-step solution to look at the log data included | |
5406 | in OS disk images. | |
db2db708 LP |
5407 | |
5408 | * journalctl's --output=cat option (which outputs the log content | |
5409 | without any metadata, just the pure text messages) will now make use | |
b182195a | 5410 | of terminal colors when run on a suitable terminal, similarly to the |
db2db708 LP |
5411 | other output modes. |
5412 | ||
5413 | * JSON group records now support a "description" string that may be | |
5414 | used to add a human-readable textual description to such groups. This | |
5415 | is supposed to match the user's GECOS field which traditionally | |
5416 | didn't have a counterpart for group records. | |
5417 | ||
5418 | * The "systemd-dissect" tool that may be used to inspect OS disk images | |
5419 | and that was previously installed to /usr/lib/systemd/ has now been | |
b182195a | 5420 | moved to /usr/bin/, reflecting its updated status of an officially |
db2db708 LP |
5421 | supported tool with a stable interface. It gained support for a new |
5422 | --mkdir switch which when combined with --mount has the effect of | |
5423 | creating the directory to mount the image to if it is missing | |
5424 | first. It also gained two new commands --copy-from and --copy-to for | |
5425 | copying files and directories in and out of an OS image without the | |
5426 | need to manually mount it. It also acquired support for a new option | |
b182195a | 5427 | --json= to generate JSON output when inspecting an OS image. |
db2db708 LP |
5428 | |
5429 | * The cgroup2 file system is now mounted with the | |
5430 | "memory_recursiveprot" mount option, supported since kernel 5.7. This | |
5431 | means that the MemoryLow= and MemoryMin= unit file settings now apply | |
5432 | recursively to whole subtrees. | |
5433 | ||
5434 | * systemd-homed now defaults to using the btrfs file system — if | |
5435 | available — when creating home directories in LUKS volumes. This may | |
b182195a ZJS |
5436 | be changed with the DefaultFileSystemType= setting in homed.conf. |
5437 | It's now the default file system in various major distributions and | |
5438 | has the major benefit for homed that it can be grown and shrunk while | |
5439 | mounted, unlike the other contenders ext4 and xfs, which can both be | |
5440 | grown online, but not shrunk (in fact xfs is the technically most | |
5441 | limited option here, as it cannot be shrunk at all). | |
db2db708 LP |
5442 | |
5443 | * JSON user records managed by systemd-homed gained support for | |
5444 | "recovery keys". These are basically secondary passphrases that can | |
b182195a ZJS |
5445 | unlock user accounts/home directories. They are computer-generated |
5446 | rather than user-chosen, and typically have greater entropy. | |
5447 | homectl's --recovery-key= option may be used to add a recovery key to | |
5448 | a user account. The generated recovery key is displayed as a QR code, | |
5449 | so that it can be scanned to be kept in a safe place. This feature is | |
5450 | particularly useful in combination with systemd-homed's support for | |
5451 | FIDO2 or PKCS#11 authentication, as a secure fallback in case the | |
5452 | security tokens are lost. Recovery keys may be entered wherever the | |
5453 | system asks for a password. | |
db2db708 LP |
5454 | |
5455 | * systemd-homed now maintains a "dirty" flag for each LUKS encrypted | |
b182195a | 5456 | home directory which indicates that a home directory has not been |
db2db708 LP |
5457 | deactivated cleanly when offline. This flag is useful to identify |
5458 | home directories for which the offline discard logic did not run when | |
5459 | offlining, and where it would be a good idea to log in again to catch | |
5460 | up. | |
5461 | ||
5462 | * systemctl gained a new parameter --timestamp= which may be used to | |
5463 | change the style in which timestamps are output, i.e. whether to show | |
5464 | them in local timezone or UTC, or whether to show µs granularity. | |
5465 | ||
5466 | * Alibaba's "pouch" container manager is now detected by | |
dc6a3162 LP |
5467 | systemd-detect-virt, ConditionVirtualization= and similar |
5468 | constructs. Similar, they now also recognize IBM PowerVM machine | |
5469 | virtualization. | |
db2db708 LP |
5470 | |
5471 | * systemd-nspawn has been reworked to use the /run/host/incoming/ as | |
5472 | place to use for propagating external mounts into the | |
b182195a ZJS |
5473 | container. Similarly /run/host/notify is now used as the socket path |
5474 | for container payloads to communicate with the container manager | |
5475 | using sd_notify(). The container manager now uses the | |
5476 | /run/host/inaccessible/ directory to place "inaccessible" file nodes | |
5477 | of all relevant types which may be used by the container payload as | |
5478 | bind mount source to over-mount inodes to make them inaccessible. | |
5479 | /run/host/container-manager will now be initialized with the same | |
5480 | string as the $container environment variable passed to the | |
5481 | container's PID 1. /run/host/container-uuid will be initialized with | |
5482 | the same string as $container_uuid. This means the /run/host/ | |
5483 | hierarchy is now the primary way to make host resources available to | |
5484 | the container. The Container Interface documents these new files and | |
5485 | directories: | |
db2db708 LP |
5486 | |
5487 | https://systemd.io/CONTAINER_INTERFACE | |
5488 | ||
5489 | * Support for the "ConditionNull=" unit file condition has been | |
b182195a ZJS |
5490 | deprecated and undocumented for 6 years. systemd started to warn |
5491 | about its use 1.5 years ago. It has now been removed entirely. | |
db2db708 | 5492 | |
db2db708 LP |
5493 | * sd-bus.h gained a new API call sd_bus_error_has_names(), which takes |
5494 | a sd_bus_error struct and a list of error names, and checks if the | |
5495 | error matches one of these names. It's a convenience wrapper that is | |
5496 | useful in cases where multiple errors shall be handled the same way. | |
5497 | ||
5498 | * A new system call filter list "@known" has been added, that contains | |
b182195a | 5499 | all system calls known at the time systemd was built. |
db2db708 LP |
5500 | |
5501 | * Behaviour of system call filter allow lists has changed slightly: | |
387f6955 | 5502 | system calls that are contained in @known will result in EPERM by |
db2db708 | 5503 | default, while those not contained in it result in ENOSYS. This |
dc6a3162 | 5504 | should improve compatibility because known system calls will thus be |
db2db708 LP |
5505 | communicated as prohibited, while unknown (and thus newer ones) will |
5506 | be communicated as not implemented, which hopefully has the greatest | |
5507 | chance of triggering the right fallback code paths in client | |
5508 | applications. | |
5509 | ||
dc6a3162 LP |
5510 | * "systemd-analyze syscall-filter" will now show two separate sections |
5511 | at the bottom of the output: system calls known during systemd build | |
5512 | time but not included in any of the filter groups shown above, and | |
5513 | system calls defined on the local kernel but known during systemd | |
5514 | build time. | |
5515 | ||
5516 | * If the $SYSTEMD_LOG_SECCOMP=1 environment variable is set for | |
5517 | systemd-nspawn all system call filter violations will be logged by | |
5518 | the kernel (audit). This is useful for tracking down system calls | |
5519 | invoked by container payloads that are prohibited by the container's | |
5520 | system call filter policy. | |
5521 | ||
6fc5b951 LP |
5522 | * If the $SYSTEMD_SECCOMP=0 environment variable is set for |
5523 | systemd-nspawn (and other programs that use seccomp) all seccomp | |
5524 | filtering is turned off. | |
5525 | ||
db2db708 | 5526 | * Two new unit file settings ProtectProc= and ProcSubset= have been |
b182195a ZJS |
5527 | added that expose the hidepid= and subset= mount options of procfs. |
5528 | All processes of the unit will only see processes in /proc that are | |
5529 | are owned by the unit's user. This is an important new sandboxing | |
5530 | option that is recommended to be set on all system services. All | |
5531 | long-running system services that are included in systemd itself set | |
5532 | this option now. This option is only supported on kernel 5.8 and | |
5533 | above, since the hidepid= option supported on older kernels was not a | |
5534 | per-mount option but actually applied to the whole PID namespace. | |
db2db708 LP |
5535 | |
5536 | * Socket units gained a new boolean setting FlushPending=. If enabled | |
5537 | all pending socket data/connections are flushed whenever the socket | |
5538 | unit enters the "listening" state, i.e. after the associated service | |
5539 | exited. | |
5540 | ||
b182195a ZJS |
5541 | * The unit file setting NUMAMask= gained a new "all" value: when used, |
5542 | all existing NUMA nodes are added to the NUMA mask. | |
db2db708 LP |
5543 | |
5544 | * A new "credentials" logic has been added to system services. This is | |
5545 | a simple mechanism to pass privileged data to services in a safe and | |
5546 | secure way. It's supposed to be used to pass per-service secret data | |
5547 | such as passwords or cryptographic keys but also associated less | |
5548 | private information such as user names, certificates, and similar to | |
5549 | system services. Each credential is identified by a short user-chosen | |
5550 | name and may contain arbitrary binary data. Two new unit file | |
b182195a ZJS |
5551 | settings have been added: SetCredential= and LoadCredential=. The |
5552 | former allows setting a credential to a literal string, the latter | |
5553 | sets a credential to the contents of a file (or data read from a | |
5554 | user-chosen AF_UNIX stream socket). Credentials are passed to the | |
5555 | service via a special credentials directory, one file for each | |
5556 | credential. The path to the credentials directory is passed in a new | |
5557 | $CREDENTIALS_DIRECTORY environment variable. Since the credentials | |
db2db708 | 5558 | are passed in the file system they may be easily referenced in |
b182195a ZJS |
5559 | ExecStart= command lines too, thus no explicit support for the |
5560 | credentials logic in daemons is required (though ideally daemons | |
5561 | would look for the bits they need in $CREDENTIALS_DIRECTORY | |
5562 | themselves automatically, if set). The $CREDENTIALS_DIRECTORY is | |
5563 | backed by unswappable memory if privileges allow it, immutable if | |
5564 | privileges allow it, is accessible only to the service's UID, and is | |
5565 | automatically destroyed when the service stops. | |
db2db708 LP |
5566 | |
5567 | * systemd-nspawn supports the same credentials logic. It can both | |
5568 | consume credentials passed to it via the aforementioned | |
5569 | $CREDENTIALS_DIRECTORY protocol as well as pass these credentials on | |
5570 | to its payload. The service manager/PID 1 has been updated to match | |
5571 | this: it can also accept credentials from the container manager that | |
b182195a ZJS |
5572 | invokes it (in fact: any process that invokes it), and passes them on |
5573 | to its services. Thus, credentials can be propagated recursively down | |
5574 | the tree: from a system's service manager to a systemd-nspawn | |
5575 | service, to the service manager that runs as container payload and to | |
5576 | the service it runs below. Credentials may also be added on the | |
5577 | systemd-nspawn command line, using new --set-credential= and | |
5578 | --load-credential= command line switches that match the | |
5579 | aforementioned service settings. | |
db2db708 LP |
5580 | |
5581 | * systemd-repart gained new settings Format=, Encrypt=, CopyFiles= in | |
5582 | the partition drop-ins which may be used to format/LUKS | |
5583 | encrypt/populate any created partitions. The partitions are | |
5584 | encrypted/formatted/populated before they are registered in the | |
b182195a ZJS |
5585 | partition table, so that they appear atomically: either the |
5586 | partitions do not exist yet or they exist fully encrypted, formatted, | |
5587 | and populated — there is no time window where they are | |
5588 | "half-initialized". Thus the system is robust to abrupt shutdown: if | |
5589 | the tool is terminated half-way during its operations on next boot it | |
5590 | will start from the beginning. | |
db2db708 LP |
5591 | |
5592 | * systemd-repart's --size= operation gained a new "auto" value. If | |
5593 | specified, and operating on a loopback file it is automatically sized | |
5594 | to the minimal size the size constraints permit. This is useful to | |
5595 | use "systemd-repart" as an image builder for minimally sized images. | |
5596 | ||
b182195a ZJS |
5597 | * systemd-resolved now gained a third IPC interface for requesting name |
5598 | resolution: besides D-Bus and local DNS to 127.0.0.53 a Varlink | |
5599 | interface is now supported. The nss-resolve NSS module has been | |
5600 | modified to use this new interface instead of D-Bus. Using Varlink | |
5601 | has a major benefit over D-Bus: it works without a broker service, | |
5602 | and thus already during earliest boot, before the dbus daemon has | |
5603 | been started. This means name resolution via systemd-resolved now | |
5604 | works at the same time systemd-networkd operates: from earliest boot | |
5605 | on, including in the initrd. | |
db2db708 LP |
5606 | |
5607 | * systemd-resolved gained support for a new DNSStubListenerExtra= | |
5608 | configuration file setting which may be used to specify additional IP | |
5609 | addresses the built-in DNS stub shall listen on, in addition to the | |
5610 | main one on 127.0.0.53:53. | |
5611 | ||
5612 | * Name lookups issued via systemd-resolved's D-Bus and Varlink | |
5613 | interfaces (and thus also via glibc NSS if nss-resolve is used) will | |
5614 | now honour a trailing dot in the hostname: if specified the search | |
b182195a ZJS |
5615 | path logic is turned off. Thus "resolvectl query foo." is now |
5616 | equivalent to "resolvectl query --search=off foo.". | |
db2db708 LP |
5617 | |
5618 | * systemd-resolved gained a new D-Bus property "ResolvConfMode" that | |
5619 | exposes how /etc/resolv.conf is currently managed: by resolved (and | |
5620 | in which mode if so) or another subsystem. "resolvctl" will display | |
5621 | this property in its status output. | |
5622 | ||
5623 | * The resolv.conf snippets systemd-resolved provides will now set "." | |
b182195a ZJS |
5624 | as the search domain if no other search domain is known. This turns |
5625 | off the derivation of an implicit search domain by nss-dns for the | |
5626 | hostname, when the hostname is set to an FQDN. This change is done to | |
5627 | make nss-dns using resolv.conf provided by systemd-resolved behave | |
5628 | more similarly to nss-resolve. | |
db2db708 LP |
5629 | |
5630 | * systemd-tmpfiles' file "aging" logic (i.e. the automatic clean-up of | |
5631 | /tmp/ and /var/tmp/ based on file timestamps) now looks at the | |
b182195a ZJS |
5632 | "birth" time (btime) of a file in addition to the atime, mtime, and |
5633 | ctime. | |
db2db708 LP |
5634 | |
5635 | * systemd-analyze gained a new verb "capability" that lists all known | |
5636 | capabilities by the systemd build and by the kernel. | |
5637 | ||
5638 | * If a file /usr/lib/clock-epoch exists, PID 1 will read its mtime and | |
5639 | advance the system clock to it at boot if it is noticed to be before | |
5640 | that time. Previously, PID 1 would only advance the time to an epoch | |
5641 | time that is set during build-time. With this new file OS builders | |
5642 | can change this epoch timestamp on individual OS images without | |
5643 | having to rebuild systemd. | |
5644 | ||
5645 | * systemd-logind will now listen to the KEY_RESTART key from the Linux | |
b182195a ZJS |
5646 | input layer and reboot the system if it is pressed, similarly to how |
5647 | it already handles KEY_POWER, KEY_SUSPEND or KEY_SLEEP. KEY_RESTART | |
5648 | was originally defined in the Multimedia context (to restart playback | |
5649 | of a song or film), but is now primarily used in various embedded | |
5650 | devices for "Reboot" buttons. Accordingly, systemd-logind will now | |
5651 | honour it as such. This may configured in more detail via the new | |
db2db708 LP |
5652 | HandleRebootKey= and RebootKeyIgnoreInhibited=. |
5653 | ||
5654 | * systemd-nspawn/systemd-machined will now reconstruct hardlinks when | |
5655 | copying OS trees, for example in "systemd-nspawn --ephemeral", | |
5656 | "systemd-nspawn --template=", "machinectl clone" and similar. This is | |
5657 | useful when operating with OSTree images, which use hardlinks heavily | |
5658 | throughout, and where such copies previously resulting in "exploding" | |
5659 | hardlinks. | |
5660 | ||
5661 | * systemd-nspawn's --console= setting gained support for a new | |
5662 | "autopipe" value, which is identical to "interactive" when invoked on | |
5663 | a TTY, and "pipe" otherwise. | |
5664 | ||
5665 | * systemd-networkd's .network files gained support for explicitly | |
5666 | configuring the multicast membership entries of bridge devices in the | |
5667 | [BridgeMDB] section. It also gained support for the PIE queuing | |
5668 | discipline in the [FlowQueuePIE] sections. | |
5669 | ||
5670 | * systemd-networkd's .netdev files may now be used to create "BareUDP" | |
905a03e7 | 5671 | tunnels, configured in the new [BareUDP] setting. |
db2db708 | 5672 | |
dc6a3162 | 5673 | * systemd-networkd's Gateway= setting in .network files now accepts the |
70a51d7e YW |
5674 | special values "_dhcp4" and "_ipv6ra" to configure additional, |
5675 | locally defined, explicit routes to the gateway acquired via DHCP or | |
5676 | IPv6 Router Advertisements. The old setting "_dhcp" is deprecated, | |
5677 | but still accepted for backwards compatibility. | |
dc6a3162 | 5678 | |
cb9a48a9 ZJS |
5679 | * systemd-networkd's [IPv6PrefixDelegation] section and |
5680 | IPv6PrefixDelegation= options have been renamed as [IPv6SendRA] and | |
5681 | IPv6SendRA= (the old names are still accepted for backwards | |
5682 | compatibility). | |
5683 | ||
70a51d7e YW |
5684 | * systemd-networkd's .network files gained the DHCPv6PrefixDelegation= |
5685 | boolean setting in [Network] section. If enabled, the delegated prefix | |
5686 | gained by another link will be configured, and an address within the | |
5687 | prefix will be assigned. | |
5688 | ||
5689 | * systemd-networkd's .network files gained the Announce= boolean setting | |
5690 | in [DHCPv6PrefixDelegation] section. When enabled, the delegated | |
5691 | prefix will be announced through IPv6 router advertisement (IPv6 RA). | |
5692 | The setting is enabled by default. | |
5693 | ||
905a03e7 ZJS |
5694 | * VXLAN tunnels may now be marked as independent of any underlying |
5695 | network interface via the new Independent= boolean setting. | |
70a51d7e | 5696 | |
b182195a ZJS |
5697 | * systemctl gained support for two new verbs: "service-log-level" and |
5698 | "service-log-target" may be used on services that implement the | |
5699 | generic org.freedesktop.LogControl1 D-Bus interface to dynamically | |
5700 | adjust the log level and target. All of systemd's long-running | |
5701 | services support this now, but ideally all system services would | |
5702 | implement this interface to make the system more uniformly | |
db2db708 LP |
5703 | debuggable. |
5704 | ||
5705 | * The SystemCallErrorNumber= unit file setting now accepts the new | |
5706 | "kill" and "log" actions, in addition to arbitrary error number | |
10f34849 YW |
5707 | specifications as before. If "kill" the processes are killed on the |
5708 | event, if "log" the offending system call is audit logged. | |
db2db708 LP |
5709 | |
5710 | * A new SystemCallLog= unit file setting has been added that accepts a | |
dc6a3162 | 5711 | list of system calls that shall be logged about (audit). |
db2db708 LP |
5712 | |
5713 | * The OS image dissection logic (as used by RootImage= in unit files or | |
b182195a | 5714 | systemd-nspawn's --image= switch) has gained support for identifying |
db2db708 LP |
5715 | and mounting explicit /usr/ partitions, which are now defined in the |
5716 | discoverable partition specification. This should be useful for | |
5717 | environments where the root file system is | |
5718 | generated/formatted/populated dynamically on first boot and combined | |
5719 | with an immutable /usr/ tree that is supplied by the vendor. | |
5720 | ||
5721 | * In the final phase of shutdown, within the systemd-shutdown binary | |
5722 | we'll now try to detach MD devices (i.e software RAID) in addition to | |
5723 | loopback block devices and DM devices as before. This is supposed to | |
5724 | be a safety net only, in order to increase robustness if things go | |
5725 | wrong. Storage subsystems are expected to properly detach their | |
5726 | storage volumes during regular shutdown already (or in case of | |
5727 | storage backing the root file system: in the initrd hook we return to | |
5728 | later). | |
5729 | ||
5730 | * If the SYSTEMD_LOG_TID environment variable is set all systemd tools | |
5731 | will now log the thread ID in their log output. This is useful when | |
5732 | working with heavily threaded programs. | |
5733 | ||
dc6a3162 | 5734 | * If the SYSTEMD_RDRAND environment variable is set to "0", systemd will |
b182195a ZJS |
5735 | not use the RDRAND CPU instruction. This is useful in environments |
5736 | such as replay debuggers where non-deterministic behaviour is not | |
5737 | desirable. | |
db2db708 | 5738 | |
dc6a3162 LP |
5739 | * The autopaging logic in systemd's various tools (such as systemctl) |
5740 | has been updated to turn on "secure" mode in "less" | |
5741 | (i.e. $LESSECURE=1) if execution in a "sudo" environment is | |
5742 | detected. This disables invoking external programs from the pager, | |
5743 | via the pipe logic. This behaviour may be overridden via the new | |
5744 | $SYSTEMD_PAGERSECURE environment variable. | |
db2db708 | 5745 | |
6e88f9e7 ZJS |
5746 | * Units which have resource limits (.service, .mount, .swap, .slice, |
5747 | .socket, and .slice) gained new configuration settings | |
5748 | ManagedOOMSwap=, ManagedOOMMemoryPressure=, and | |
90616bb9 | 5749 | ManagedOOMMemoryPressureLimitPercent= that specify resource pressure |
6e88f9e7 ZJS |
5750 | limits and optional action taken by systemd-oomd. |
5751 | ||
5752 | * A new service systemd-oomd has been added. It monitors resource | |
5753 | contention for selected parts of the unit hierarchy using the PSI | |
5754 | information reported by the kernel, and kills processes when memory | |
5755 | or swap pressure is above configured limits. This service is only | |
08c1864f ZJS |
5756 | enabled by default in developer mode (see below) and should be |
5757 | considered a preview in this release. Behaviour details and option | |
5758 | names are subject to change without the usual backwards-compatibility | |
5759 | promises. | |
6e88f9e7 ZJS |
5760 | |
5761 | * A new helper oomctl has been added to introspect systemd-oomd state. | |
042b028a | 5762 | It is only enabled by default in developer mode and should be |
08c1864f ZJS |
5763 | considered a preview without the usual backwards-compatibility |
5764 | promises. | |
6e88f9e7 ZJS |
5765 | |
5766 | * New meson option -Dcompat-mutable-uid-boundaries= has been added. If | |
5767 | enabled, systemd reads the system UID boundaries from /etc/login.defs | |
5768 | at runtime, instead of using the built-in values selected during | |
5769 | build. This is an option to improve compatibility for upgrades from | |
5770 | old systems. It's strongly recommended not to make use of this | |
5771 | functionality on new systems (or even enable it during build), as it | |
5772 | makes something runtime-configurable that is mostly an implementation | |
5773 | detail of the OS, and permits avoidable differences in deployments | |
5774 | that create all kinds of problems in the long run. | |
5775 | ||
5776 | * New meson option '-Dmode=developer|release' has been added. When | |
5777 | 'developer', additional checks and features are enabled that are | |
5778 | relevant during upstream development, e.g. verification that | |
5779 | semi-automatically-generated documentation has been properly updated | |
5780 | following API changes. Those checks are considered hints for | |
5781 | developers and are not actionable in downstream builds. In addition, | |
5782 | extra features that are not ready for general consumption may be | |
5783 | enabled in developer mode. It is thus recommended to set | |
5784 | '-Dmode=release' in end-user and distro builds. | |
5785 | ||
da7511d5 LP |
5786 | * systemd-cryptsetup gained support for processing detached LUKS |
5787 | headers specified on the kernel command line via the header= | |
5788 | parameter of the luks.options= kernel command line option. The same | |
5789 | device/path syntax as for key files is supported for header files | |
5790 | like this. | |
5791 | ||
5792 | * The "net_id" built-in of udev has been updated to ignore ACPI _SUN | |
5793 | slot index data for devices that are connected through a PCI bridge | |
5794 | where the _SUN index is associated with the bridge instead of the | |
5795 | network device itself. Previously this would create ambiguous device | |
5796 | naming if multiple network interfaces were connected to the same PCI | |
5797 | bridge. Since this is a naming scheme incompatibility on systems that | |
5798 | possess hardware like this it has been introduced as new naming | |
5799 | scheme "v247". The previous scheme can be selected via the | |
5800 | "net.naming-scheme=v245" kernel command line parameter. | |
5801 | ||
5802 | * ConditionFirstBoot= semantics have been modified to be safe towards | |
5803 | abnormal system power-off during first boot. Specifically, the | |
5804 | "systemd-machine-id-commit.service" service now acts as boot | |
5805 | milestone indicating when the first boot process is sufficiently | |
5806 | complete in order to not consider the next following boot also a | |
5807 | first boot. If the system is reset before this unit is reached the | |
5808 | first time, the next boot will still be considered a first boot; once | |
5809 | it has been reached, no further boots will be considered a first | |
5810 | boot. The "first-boot-complete.target" unit now acts as official hook | |
5811 | point to order against this. If a service shall be run on every boot | |
5812 | until the first boot fully succeeds it may thus be ordered before | |
5813 | this target unit (and pull it in) and carry ConditionFirstBoot= | |
5814 | appropriately. | |
5815 | ||
5816 | * bootctl's set-default and set-oneshot commands now accept the three | |
5817 | special strings "@default", "@oneshot", "@current" in place of a boot | |
5818 | entry id. These strings are resolved to the current default and | |
5819 | oneshot boot loader entry, as well as the currently booted one. Thus | |
5820 | a command "bootctl set-default @current" may be used to make the | |
5821 | currently boot menu item the new default for all subsequent boots. | |
5822 | ||
da7511d5 LP |
5823 | * "systemctl edit" has been updated to show the original effective unit |
5824 | contents in commented form in the text editor. | |
5825 | ||
9101cd1a ZJS |
5826 | * Units in user mode are now segregated into three new slices: |
5827 | session.slice (units that form the core of graphical session), | |
5828 | app.slice ("normal" user applications), and background.slice | |
5829 | (low-priority tasks). Unless otherwise configured, user units are | |
5830 | placed in app.slice. The plan is to add resource limits and | |
5831 | protections for the different slices in the future. | |
5832 | ||
7bfcc0de LP |
5833 | * New GPT partition types for RISCV32/64 for the root and /usr |
5834 | partitions, and their associated Verity partitions have been defined, | |
5835 | and are now understood by systemd-gpt-auto-generator, and the OS | |
5836 | image dissection logic. | |
5837 | ||
a5322567 | 5838 | Contributions from: Adolfo Jayme Barrientos, afg, Alec Moskvin, Alyssa |
6fc5b951 | 5839 | Ross, Amitanand Chikorde, Andrew Hangsleben, Anita Zhang, Ansgar |
a5322567 LP |
5840 | Burchardt, Arian van Putten, Aurelien Jarno, Axel Rasmussen, bauen1, |
5841 | Beniamino Galvani, Benjamin Berg, Bjørn Mork, brainrom, Chandradeep | |
d0dcf59b ZJS |
5842 | Dey, Charles Lee, Chris Down, Christian Göttsche, Christof Efkemann, |
5843 | Christoph Ruegge, Clemens Gruber, Daan De Meyer, Daniele Medri, Daniel | |
5844 | Mack, Daniel Rusek, Dan Streetman, David Tardon, Dimitri John Ledkov, | |
5845 | Dmitry Borodaenko, Elias Probst, Elisei Roca, ErrantSpore, Etienne | |
5846 | Doms, Fabrice Fontaine, fangxiuning, Felix Riemann, Florian Klink, | |
5847 | Franck Bui, Frantisek Sumsal, fwSmit, George Rawlinson, germanztz, | |
5848 | Gibeom Gwon, Glen Whitney, Gogo Gogsi, Göran Uddeborg, Grant Mathews, | |
5849 | Hans de Goede, Hans Ulrich Niedermann, Haochen Tong, Harald Seiler, | |
5850 | huangyong, Hubert Kario, igo95862, Ikey Doherty, Insun Pyo, Jan Chren, | |
5851 | Jan Schlüter, Jérémy Nouhaud, Jian-Hong Pan, Joerg Behrmann, Jonathan | |
5852 | Lebon, Jörg Thalheim, Josh Brobst, Juergen Hoetzel, Julien Humbert, | |
5853 | Kai-Chuan Hsieh, Kairui Song, Kamil Dudka, Kir Kolyshkin, Kristijan | |
5854 | Gjoshev, Kyle Huey, Kyle Russell, Lee Whalen, Lennart Poettering, | |
5855 | lichangze, Luca Boccassi, Lucas Werkmeister, Luca Weiss, Marc | |
5856 | Kleine-Budde, Marco Wang, Martin Wilck, Marti Raudsepp, masmullin2000, | |
5857 | Máté Pozsgay, Matt Fenwick, Michael Biebl, Michael Scherer, Michal | |
5858 | Koutný, Michal Sekletár, Michal Suchanek, Mikael Szreder, Milo | |
5859 | Casagrande, mirabilos, Mitsuha_QuQ, mog422, Muhammet Kara, Nazar | |
5860 | Vinnichuk, Nicholas Narsing, Nicolas Fella, Njibhu, nl6720, Oğuz Ersen, | |
5861 | Olivier Le Moal, Ondrej Kozina, onlybugreports, Pass Automated Testing | |
5862 | Suite, Pat Coulthard, Pavel Sapezhko, Pedro Ruiz, perry_yuan, Peter | |
5863 | Hutterer, Phaedrus Leeds, PhoenixDiscord, Piotr Drąg, Plan C, | |
5864 | Purushottam choudhary, Rasmus Villemoes, Renaud Métrich, Robert Marko, | |
5865 | Roman Beranek, Ronan Pigott, Roy Chen (陳彥廷), RussianNeuroMancer, | |
5866 | Samanta Navarro, Samuel BF, scootergrisen, Sorin Ionescu, Steve Dodd, | |
5867 | Susant Sahani, Timo Rothenpieler, Tobias Hunger, Tobias Kaufmann, Topi | |
5868 | Miettinen, vanou, Vito Caputo, Weblate, Wen Yang, Whired Planck, | |
5869 | williamvds, Yu, Li-Yu, Yuri Chornoivan, Yu Watanabe, Zbigniew | |
5870 | Jędrzejewski-Szmek, Zmicer Turok, Дамјан Георгиевски | |
5871 | ||
5872 | – Warsaw, 2020-11-26 | |
a5322567 | 5873 | |
72e51908 | 5874 | CHANGES WITH 246: |
b0d0e0ef | 5875 | |
7f56c26d ZJS |
5876 | * The service manager gained basic support for cgroup v2 freezer. Units |
5877 | can now be suspended or resumed either using new systemctl verbs, | |
5878 | freeze and thaw respectively, or via D-Bus. | |
5879 | ||
5880 | * PID 1 may now automatically load pre-compiled AppArmor policies from | |
5881 | /etc/apparmor/earlypolicy during early boot. | |
5882 | ||
5883 | * The CPUAffinity= setting in service unit files now supports a new | |
5884 | special value "numa" that causes the CPU affinity masked to be set | |
5885 | based on the NUMA mask. | |
5886 | ||
5887 | * systemd will now log about all left-over processes remaining in a | |
5888 | unit when the unit is stopped. It will now warn about services using | |
5889 | KillMode=none, as this is generally an unsafe thing to make use of. | |
5890 | ||
5891 | * Two new unit file settings | |
5892 | ConditionPathIsEncrypted=/AssertPathIsEncrypted= have been | |
5893 | added. They may be used to check whether a specific file system path | |
5894 | resides on a block device that is encrypted on the block level | |
5895 | (i.e. using dm-crypt/LUKS). | |
5896 | ||
5897 | * Another pair of new settings ConditionEnvironment=/AssertEnvironment= | |
5898 | has been added that may be used for simple environment checks. This | |
5899 | is particularly useful when passing in environment variables from a | |
5900 | container manager (or from PAM in case of the systemd --user | |
5901 | instance). | |
5902 | ||
5903 | * .service unit files now accept a new setting CoredumpFilter= which | |
5904 | allows configuration of the memory sections coredumps of the | |
5905 | service's processes shall include. | |
5906 | ||
5907 | * .mount units gained a new ReadWriteOnly= boolean option. If set | |
5908 | it will not be attempted to mount a file system read-only if mounting | |
5909 | in read-write mode doesn't succeed. An option x-systemd.rw-only is | |
5910 | available in /etc/fstab to control the same. | |
5911 | ||
5912 | * .socket units gained a new boolean setting PassPacketInfo=. If | |
5913 | enabled, the kernel will attach additional per-packet metadata to all | |
5149990e CR |
5914 | packets read from the socket, as an ancillary message. This controls |
5915 | the IP_PKTINFO, IPV6_RECVPKTINFO, NETLINK_PKTINFO socket options, | |
7f56c26d ZJS |
5916 | depending on socket type. |
5917 | ||
5918 | * .service units gained a new setting RootHash= which may be used to | |
5919 | specify the root hash for verity enabled disk images which are | |
5920 | specified in RootImage=. RootVerity= may be used to specify a path to | |
5921 | the Verity data matching a RootImage= file system. (The latter is | |
5922 | only useful for images that do not contain the Verity data embedded | |
5923 | into the same image that carries a GPT partition table following the | |
5924 | Discoverable Partition Specification). Similarly, systemd-nspawn | |
5925 | gained a new switch --verity-data= that takes a path to a file with | |
5926 | the verity data of the disk image supplied in --image=, if the image | |
5927 | doesn't contain the verity data itself. | |
5928 | ||
5929 | * .service units gained a new setting RootHashSignature= which takes | |
5930 | either a base64 encoded PKCS#7 signature of the root hash specified | |
5931 | with RootHash=, or a path to a file to read the signature from. This | |
5932 | allows validation of the root hash against public keys available in | |
5933 | the kernel keyring, and is only supported on recent kernels | |
5934 | (>= 5.4)/libcryptsetup (>= 2.30). A similar switch has been added to | |
5935 | systemd-nspawn and systemd-dissect (--root-hash-sig=). Support for | |
5936 | this mechanism has also been added to systemd-veritysetup. | |
5937 | ||
5938 | * .service unit files gained two new options | |
5939 | TimeoutStartFailureMode=/TimeoutStopFailureMode= that may be used to | |
5940 | tune behaviour if a start or stop timeout is hit, i.e. whether to | |
5941 | terminate the service with SIGTERM, SIGABRT or SIGKILL. | |
5942 | ||
5943 | * Most options in systemd that accept hexadecimal values prefixed with | |
5944 | 0x in additional to the usual decimal notation now also support octal | |
cb713f16 | 5945 | notation when the 0o prefix is used and binary notation if the 0b |
7f56c26d ZJS |
5946 | prefix is used. |
5947 | ||
337f0b00 ZJS |
5948 | * Various command line parameters and configuration file settings that |
5949 | configure key or certificate files now optionally take paths to | |
5950 | AF_UNIX sockets in the file system. If configured that way a stream | |
5951 | connection is made to the socket and the required data read from | |
5952 | it. This is a simple and natural extension to the existing regular | |
5953 | file logic, and permits other software to provide keys or | |
5954 | certificates via simple IPC services, for example when unencrypted | |
5955 | storage on disk is not desired. Specifically, systemd-networkd's | |
5956 | Wireguard and MACSEC key file settings as well as | |
5957 | systemd-journal-gatewayd's and systemd-journal-remote's PEM | |
5958 | key/certificate parameters support this now. | |
5959 | ||
7f56c26d ZJS |
5960 | * Unit files, tmpfiles.d/ snippets, sysusers.d/ snippets and other |
5961 | configuration files that support specifier expansion learnt six new | |
5962 | specifiers: %a resolves to the current architecture, %o/%w/%B/%W | |
5963 | resolve to the various ID fields from /etc/os-release, %l resolves to | |
5964 | the "short" hostname of the system, i.e. the hostname configured in | |
5965 | the kernel truncated at the first dot. | |
5966 | ||
5967 | * Support for the .include syntax in unit files has been removed. The | |
5968 | concept has been obsolete for 6 years and we started warning about | |
5969 | its pending removal 2 years ago (also see NEWS file below). It's | |
5970 | finally gone now. | |
5971 | ||
5972 | * StandardError= and StandardOutput= in unit files no longer support | |
5973 | the "syslog" and "syslog-console" switches. They were long removed | |
5974 | from the documentation, but will now result in warnings when used, | |
5975 | and be converted to "journal" and "journal+console" automatically. | |
5976 | ||
5d043c9f LP |
5977 | * If the service setting User= is set to the "nobody" user, a warning |
5978 | message is now written to the logs (but the value is nonetheless | |
5979 | accepted). Setting User=nobody is unsafe, since the primary purpose | |
5980 | of the "nobody" user is to own all files whose owner cannot be mapped | |
5981 | locally. It's in particular used by the NFS subsystem and in user | |
5982 | namespacing. By running a service under this user's UID it might get | |
5983 | read and even write access to all these otherwise unmappable files, | |
5984 | which is quite likely a major security problem. | |
5985 | ||
02e1eb6d ZJS |
5986 | * tmpfs mounts automatically created by systemd (/tmp, /run, /dev/shm, |
5987 | and others) now have a size and inode limits applied (50% of RAM for | |
c2cfb126 ZJS |
5988 | /tmp and /dev/shm, 10% of RAM for other mounts, etc.). Please note |
5989 | that the implicit kernel default is 50% too, so there is no change | |
5990 | in the size limit for /tmp and /dev/shm. | |
02e1eb6d | 5991 | |
337f0b00 ZJS |
5992 | * nss-mymachines lost support for resolution of users and groups, and |
5993 | now only does resolution of hostnames. This functionality is now | |
5994 | provided by nss-systemd. Thus, the 'mymachines' entry should be | |
5995 | removed from the 'passwd:' and 'group:' lines in /etc/nsswitch.conf | |
5996 | (and 'systemd' added if it is not already there). | |
5997 | ||
7f56c26d ZJS |
5998 | * A new kernel command line option systemd.hostname= has been added |
5999 | that allows controlling the hostname that is initialized early during | |
6000 | boot. | |
6001 | ||
6002 | * A kernel command line option "udev.blockdev_read_only" has been | |
6003 | added. If specified all hardware block devices that show up are | |
6004 | immediately marked as read-only by udev. This option is useful for | |
6005 | making sure that a specific boot under no circumstances modifies data | |
6006 | on disk. Use "blockdev --setrw" to undo the effect of this, per | |
6007 | device. | |
b0d0e0ef LP |
6008 | |
6009 | * A new boolean kernel command line option systemd.swap= has been | |
6010 | added, which may be used to turn off automatic activation of swap | |
7f56c26d | 6011 | devices listed in /etc/fstab. |
b0d0e0ef | 6012 | |
7f56c26d ZJS |
6013 | * New kernel command line options systemd.condition-needs-update= and |
6014 | systemd.condition-first-boot= have been added, which override the | |
6015 | result of the ConditionNeedsUpdate= and ConditionFirstBoot= | |
6016 | conditions. | |
b0d0e0ef | 6017 | |
7f56c26d ZJS |
6018 | * A new kernel command line option systemd.clock-usec= has been added |
6019 | that allows setting the system clock to the specified time in µs | |
6020 | since Jan 1st, 1970 early during boot. This is in particular useful | |
6021 | in order to make test cases more reliable. | |
b0d0e0ef | 6022 | |
7f56c26d ZJS |
6023 | * The fs.suid_dumpable sysctl is set to 2 / "suidsafe". This allows |
6024 | systemd-coredump to save core files for suid processes. When saving | |
6025 | the core file, systemd-coredump will use the effective uid and gid of | |
6026 | the process that faulted. | |
6027 | ||
6028 | * The /sys/module/kernel/parameters/crash_kexec_post_notifiers file is | |
6029 | now automatically set to "Y" at boot, in order to enable pstore | |
6030 | generation for collection with systemd-pstore. | |
6031 | ||
c2cfb126 | 6032 | * We provide a set of udev rules to enable auto-suspend on PCI and USB |
69e3234d | 6033 | devices that were tested to correctly support it. Previously, this |
c2cfb126 ZJS |
6034 | was distributed as a set of udev rules, but has now been replaced by |
6035 | by a set of hwdb entries (and a much shorter udev rule to take action | |
6036 | if the device modalias matches one of the new hwdb entries). | |
6037 | ||
6038 | As before, entries are periodically imported from the database | |
6039 | maintained by the ChromiumOS project. If you have a device that | |
6040 | supports auto-suspend correctly and where it should be enabled by | |
6041 | default, please submit a patch that adds it to the database (see | |
6042 | /usr/lib/udev/hwdb.d/60-autosuspend.hwdb). | |
7f56c26d | 6043 | |
5149990e CR |
6044 | * systemd-udevd gained the new configuration option timeout_signal= as well |
6045 | as a corresponding kernel command line option udev.timeout_signal=. | |
7f56c26d ZJS |
6046 | The option can be used to configure the UNIX signal that the main |
6047 | daemon sends to the worker processes on timeout. Setting the signal | |
6048 | to SIGABRT is useful for debugging. | |
6049 | ||
6050 | * .link files managed by systemd-udevd gained options RxFlowControl=, | |
6051 | TxFlowControl=, AutoNegotiationFlowControl= in the [Link] section, in | |
6052 | order to configure various flow control parameters. They also gained | |
6053 | RxMiniBufferSize= and RxJumboBufferSize= in order to configure jumbo | |
6054 | frame ring buffer sizes. | |
b0d0e0ef | 6055 | |
3ea58e01 | 6056 | * networkd.conf gained a new boolean setting ManageForeignRoutes=. If |
aa0b850b | 6057 | enabled systemd-networkd manages all routes configured by other tools. |
3ea58e01 | 6058 | |
71af436c YW |
6059 | * .network files managed by systemd-networkd gained a new section |
6060 | [SR-IOV], in order to configure SR-IOV capable network devices. | |
6061 | ||
b0d0e0ef LP |
6062 | * systemd-networkd's [IPv6Prefix] section in .network files gained a |
6063 | new boolean setting Assign=. If enabled an address from the prefix is | |
6064 | automatically assigned to the interface. | |
6065 | ||
c05b8edd YW |
6066 | * systemd-networkd gained a new section [DHCPv6PrefixDelegation] which |
6067 | controls delegated prefixes assigned by DHCPv6 client. The section | |
6068 | has three settings: SubnetID=, Assign=, and Token=. The setting | |
6069 | SubnetID= allows explicit configuration of the preferred subnet that | |
6070 | systemd-networkd's Prefix Delegation logic assigns to interfaces. If | |
6071 | Assign= is enabled (which is the default) an address from any acquired | |
6072 | delegated prefix is automatically chosen and assigned to the | |
6073 | interface. The setting Token= specifies an optional address generation | |
6074 | mode for Assign=. | |
b0d0e0ef | 6075 | |
3ea58e01 YW |
6076 | * systemd-networkd's [Network] section gained a new setting |
6077 | IPv4AcceptLocal=. If enabled the interface accepts packets with local | |
6078 | source addresses. | |
6079 | ||
b0d0e0ef LP |
6080 | * systemd-networkd gained support for configuring the HTB queuing |
6081 | discipline in the [HierarchyTokenBucket] and | |
6082 | [HierarchyTokenBucketClass] sections. Similar the "pfifo" qdisc may | |
6083 | be configured in the [PFIFO] section, "GRED" in | |
6084 | [GenericRandomEarlyDetection], "SFB" in [StochasticFairBlue], "cake" | |
6085 | in [CAKE], "PIE" in [PIE], "DRR" in [DeficitRoundRobinScheduler] and | |
6086 | [DeficitRoundRobinSchedulerClass], "BFIFO" in [BFIFO], | |
1f19ae0f LP |
6087 | "PFIFOHeadDrop" in [PFIFOHeadDrop], "PFIFOFast" in [PFIFOFast], "HHF" |
6088 | in [HeavyHitterFilter], "ETS" in [EnhancedTransmissionSelection] and | |
3ea58e01 | 6089 | "QFQ" in [QuickFairQueueing] and [QuickFairQueueingClass]. |
b0d0e0ef LP |
6090 | |
6091 | * systemd-networkd gained support for a new Termination= setting in the | |
6092 | [CAN] section for configuring the termination resistor. It also | |
6093 | gained a new ListenOnly= setting for controlling whether to only | |
6094 | listen on CAN interfaces, without interfering with traffic otherwise | |
6095 | (which is useful for debugging/monitoring CAN network | |
6096 | traffic). DataBitRate=, DataSamplePoint=, FDMode=, FDNonISO= have | |
6097 | been added to configure various CAN-FD aspects. | |
6098 | ||
554a97ba YW |
6099 | * systemd-networkd's [DHCPv6] section gained a new option WithoutRA=. |
6100 | When enabled, DHCPv6 will be attempted right-away without requiring an | |
6101 | Router Advertisement packet suggesting it first (i.e. without the 'M' | |
6102 | or 'O' flags set). The [IPv6AcceptRA] section gained a boolean option | |
6103 | DHCPv6Client= that may be used to turn off the DHCPv6 client even if | |
6104 | the RA packets suggest it. | |
b0d0e0ef LP |
6105 | |
6106 | * systemd-networkd's [DHCPv4] section gained a new setting UseGateway= | |
6107 | which may be used to turn off use of the gateway information provided | |
6108 | by the DHCP lease. A new FallbackLeaseLifetimeSec= setting may be | |
6109 | used to configure how to process leases that lack a lifetime option. | |
6110 | ||
6111 | * systemd-networkd's [DHCPv4] and [DHCPServer] sections gained a new | |
6112 | setting SendVendorOption= allowing configuration of additional vendor | |
6113 | options to send in the DHCP requests/responses. The [DHCPv6] section | |
6114 | gained a new SendOption= setting for sending arbitrary DHCP | |
6115 | options. RequestOptions= has been added to request arbitrary options | |
6116 | from the server. UserClass= has been added to set the DHCP user class | |
6117 | field. | |
6118 | ||
6119 | * systemd-networkd's [DHCPServer] section gained a new set of options | |
2a71d57f | 6120 | EmitPOP3=/POP3=, EmitSMTP=/SMTP=, EmitLPR=/LPR= for including server |
b0d0e0ef LP |
6121 | information about these three protocols in the DHCP lease. It also |
6122 | gained support for including "MUD" URLs ("Manufacturer Usage | |
6123 | Description"). Support for "MUD" URLs was also added to the LLDP | |
6124 | stack, configurable in the [LLDP] section in .network files. | |
6125 | ||
6b949d26 YW |
6126 | * The Mode= settings in [MACVLAN] and [MACVTAP] now support 'source' |
6127 | mode. Also, the sections now support a new setting SourceMACAddress=. | |
6128 | ||
7f56c26d ZJS |
6129 | * systemd-networkd's .netdev files now support a new setting |
6130 | VLANProtocol= in the [Bridge] section that allows configuration of | |
6131 | the VLAN protocol to use. | |
6132 | ||
6133 | * systemd-networkd supports a new Group= setting in the [Link] section | |
6134 | of the .network files, to control the link group. | |
6135 | ||
6f6296b9 | 6136 | * systemd-networkd's [Network] section gained a new |
3ea58e01 YW |
6137 | IPv6LinkLocalAddressGenerationMode= setting, which specifies how IPv6 |
6138 | link local address is generated. | |
6139 | ||
7f56c26d ZJS |
6140 | * A new default .network file is now shipped that matches TUN/TAP |
6141 | devices that begin with "vt-" in their name. Such interfaces will | |
6142 | have IP routing onto the host links set up automatically. This is | |
6143 | supposed to be used by VM managers to trivially acquire a network | |
6144 | interface which is fully set up for host communication, simply by | |
6145 | carefully picking an interface name to use. | |
6146 | ||
3ea58e01 | 6147 | * systemd-networkd's [DHCPv6] section gained a new setting RouteMetric= |
aa0b850b | 6148 | which sets the route priority for routes specified by the DHCP server. |
3ea58e01 YW |
6149 | |
6150 | * systemd-networkd's [DHCPv6] section gained a new setting VendorClass= | |
6151 | which configures the vendor class information sent to DHCP server. | |
6152 | ||
7f56c26d ZJS |
6153 | * The BlackList= settings in .network files' [DHCPv4] and |
6154 | [IPv6AcceptRA] sections have been renamed DenyList=. The old names | |
6155 | are still understood to provide compatibility. | |
6156 | ||
6157 | * networkctl gained the new "forcerenew" command for forcing all DHCP | |
6158 | server clients to renew their lease. The interface "status" output | |
6159 | will now show numerous additional fields of information about an | |
6160 | interface. There are new "up" and "down" commands to bring specific | |
6161 | interfaces up or down. | |
6162 | ||
337f0b00 ZJS |
6163 | * systemd-resolved's DNS= configuration option now optionally accepts a |
6164 | port number (after ":") and a host name (after "#"). When the host | |
6165 | name is specified, the DNS-over-TLS certificate is validated to match | |
6166 | the specified hostname. Additionally, in case of IPv6 addresses, an | |
6167 | interface may be specified (after "%"). | |
b0d0e0ef | 6168 | |
2bd5e1b2 ZJS |
6169 | * systemd-resolved may be configured to forward single-label DNS names. |
6170 | This is not standard-conformant, but may make sense in setups where | |
6171 | public DNS servers are not used. | |
6172 | ||
6173 | * systemd-resolved's DNS-over-TLS support gained SNI validation. | |
6174 | ||
7f56c26d ZJS |
6175 | * systemd-nspawn's --resolv-conf= switch gained a number of new |
6176 | supported values. Specifically, options starting with "replace-" are | |
6177 | like those prefixed "copy-" but replace any existing resolv.conf | |
6178 | file. And options ending in "-uplink" and "-stub" can now be used to | |
6179 | propagate other flavours of resolv.conf into the container (as | |
6180 | defined by systemd-resolved). | |
b0d0e0ef | 6181 | |
7f56c26d ZJS |
6182 | * The various programs included in systemd can now optionally output |
6183 | their log messages on stderr prefixed with a timestamp, controlled by | |
6184 | the $SYSTEMD_LOG_TIME environment variable. | |
366d0772 | 6185 | |
b0d0e0ef LP |
6186 | * systemctl gained a new "-P" switch that is a shortcut for "--value |
6187 | --property=…". | |
6188 | ||
7f56c26d ZJS |
6189 | * "systemctl list-units" and "systemctl list-machines" no longer hide |
6190 | their first output column with --no-legend. To hide the first column, | |
6191 | use --plain. | |
b0d0e0ef | 6192 | |
7f56c26d ZJS |
6193 | * "systemctl reboot" takes the option "--reboot-argument=". |
6194 | The optional positional argument to "systemctl reboot" is now | |
6195 | being deprecated in favor of this option. | |
b0d0e0ef LP |
6196 | |
6197 | * systemd-run gained a new switch --slice-inherit. If specified the | |
6198 | unit it generates is placed in the same slice as the systemd-run | |
6199 | process itself. | |
6200 | ||
7f56c26d ZJS |
6201 | * systemd-journald gained support for zstd compression of large fields |
6202 | in journal files. The hash tables in journal files have been hardened | |
6203 | against hash collisions. This is an incompatible change and means | |
6204 | that journal files created with new systemd versions are not readable | |
6205 | with old versions. If the $SYSTEMD_JOURNAL_KEYED_HASH boolean | |
6206 | environment variable for systemd-journald.service is set to 0 this | |
6207 | new hardening functionality may be turned off, so that generated | |
6208 | journal files remain compatible with older journalctl | |
6209 | implementations. | |
b0d0e0ef | 6210 | |
7f56c26d | 6211 | * journalctl will now include a clickable link in the default output for |
5bc9ea07 | 6212 | each log message for which a URL with further documentation is |
7f56c26d ZJS |
6213 | known. This is only supported on terminal emulators that support |
6214 | clickable hyperlinks, and is turned off if a pager is used (since | |
6215 | "less" still doesn't support hyperlinks, | |
6216 | unfortunately). Documentation URLs may be included in log messages | |
6217 | either by including a DOCUMENTATION= journal field in it, or by | |
6218 | associating a journal message catalog entry with the log message's | |
6219 | MESSAGE_ID, which then carries a "Documentation:" tag. | |
b0d0e0ef LP |
6220 | |
6221 | * journald.conf gained a new boolean setting Audit= that may be used to | |
6222 | control whether systemd-journald will enable audit during | |
6223 | initialization. | |
6224 | ||
7f56c26d ZJS |
6225 | * when systemd-journald's log stream is broken up into multiple lines |
6226 | because the PID of the sender changed this is indicated in the | |
6227 | generated log records via the _LINE_BREAK=pid-change field. | |
b0d0e0ef | 6228 | |
7f56c26d ZJS |
6229 | * journalctl's "-o cat" output mode will now show one or more journal |
6230 | fields specified with --output-fields= instead of unconditionally | |
6231 | MESSAGE=. This is useful to retrieve a very specific set of fields | |
6232 | without any decoration. | |
b0d0e0ef | 6233 | |
495454f4 LP |
6234 | * The sd-journal.h API gained two new functions: |
6235 | sd_journal_enumerate_available_unique() and | |
6236 | sd_journal_enumerate_available_data() that operate like their | |
6237 | counterparts that lack the _available_ in the name, but skip items | |
6238 | that cannot be read and processed by the local implementation | |
6239 | (i.e. are compressed in an unsupported format or such), | |
6240 | ||
7f56c26d ZJS |
6241 | * coredumpctl gained a new --file= switch, matching the same one in |
6242 | journalctl: a specific journal file may be specified to read the | |
6243 | coredump data from. | |
6244 | ||
6245 | * coredumps collected by systemd-coredump may now be compressed using | |
6246 | the zstd algorithm. | |
b0d0e0ef LP |
6247 | |
6248 | * systemd-binfmt gained a new switch --unregister for unregistering all | |
6249 | registered entries at once. This is now invoked automatically at | |
6250 | shutdown, so that binary formats registered with the "F" flag will | |
6251 | not block clean file system unmounting. | |
6252 | ||
b0d0e0ef | 6253 | * systemd-notify's --pid= switch gained new values: "parent", "self", |
1d16f661 | 6254 | "auto" for controlling which PID to send to the service manager: the |
b0d0e0ef LP |
6255 | systemd-notify process' PID, or the one of the process invoking it. |
6256 | ||
b0d0e0ef LP |
6257 | * systemd-logind's Session bus object learnt a new method call |
6258 | SetType() for temporarily updating the session type of an already | |
6259 | allocated session. This is useful for upgrading tty sessions to | |
6260 | graphical ones once a compositor is invoked. | |
6261 | ||
b0d0e0ef LP |
6262 | * systemd-socket-proxy gained a new switch --exit-idle-time= for |
6263 | configuring an exit-on-idle time. | |
6264 | ||
b0d0e0ef LP |
6265 | * systemd-repart's --empty= setting gained a new value "create". If |
6266 | specified a new empty regular disk image file is created under the | |
1d16f661 | 6267 | specified name. Its size may be specified with the new --size= |
b0d0e0ef LP |
6268 | option. The latter is also supported without the "create" mode, in |
6269 | order to grow existing disk image files to the specified size. These | |
1d16f661 FS |
6270 | two new options are useful when creating or manipulating disk images |
6271 | instead of operating on actual block devices. | |
b0d0e0ef LP |
6272 | |
6273 | * systemd-repart drop-ins now support a new UUID= setting to control | |
6274 | the UUID to assign to a newly created partition. | |
6275 | ||
1f19ae0f LP |
6276 | * systemd-repart's SizeMin= per-partition parameter now defaults to 10M |
6277 | instead of 0. | |
6278 | ||
6279 | * systemd-repart's Label= setting now support the usual, simple | |
6280 | specifier expansion. | |
6281 | ||
7f56c26d ZJS |
6282 | * systemd-homed's LUKS backend gained the ability to discard empty file |
6283 | system blocks automatically when the user logs out. This is enabled | |
6284 | by default to ensure that home directories take minimal space when | |
6285 | logged out but get full size guarantees when logged in. This may be | |
6286 | controlled with the new --luks-offline-discard= switch to homectl. | |
b0d0e0ef | 6287 | |
7f56c26d ZJS |
6288 | * If systemd-homed detects that /home/ is encrypted as a whole it will |
6289 | now default to the directory or subvolume backends instead of the | |
6290 | LUKS backend, in order to avoid double encryption. The default | |
6291 | storage and file system may now be configured explicitly, too, via | |
6292 | the new /etc/systemd/homed.conf configuration file. | |
b0d0e0ef | 6293 | |
5d043c9f LP |
6294 | * systemd-homed now supports unlocking home directories with FIDO2 |
6295 | security tokens that support the 'hmac-secret' extension, in addition | |
6296 | to the existing support for PKCS#11 security token unlocking | |
6297 | support. Note that many recent hardware security tokens support both | |
6298 | interfaces. The FIDO2 support is accessible via homectl's | |
6299 | --fido2-device= option. | |
6300 | ||
6301 | * homectl's --pkcs11-uri= setting now accepts two special parameters: | |
6302 | if "auto" is specified and only one suitable PKCS#11 security token | |
6303 | is plugged in, its URL is automatically determined and enrolled for | |
6304 | unlocking the home directory. If "list" is specified a brief table of | |
6305 | suitable PKCS#11 security tokens is shown. Similar, the new | |
6306 | --fido2-device= option also supports these two special values, for | |
6307 | automatically selecting and listing suitable FIDO2 devices. | |
6308 | ||
b0d0e0ef LP |
6309 | * The /etc/crypttab tmp option now optionally takes an argument |
6310 | selecting the file system to use. Moreover, the default is now | |
6311 | changed from ext2 to ext4. | |
6312 | ||
6313 | * There's a new /etc/crypttab option "keyfile-erase". If specified the | |
6314 | key file listed in the same line is removed after use, regardless if | |
6315 | volume activation was successful or not. This is useful if the key | |
6316 | file is only acquired transiently at runtime and shall be erased | |
6317 | before the system continues to boot. | |
6318 | ||
6319 | * There's also a new /etc/crypttab option "try-empty-password". If | |
6320 | specified, before asking the user for a password it is attempted to | |
6321 | unlock the volume with an empty password. This is useful for | |
6322 | installing encrypted images whose password shall be set on first boot | |
6323 | instead of at installation time. | |
6324 | ||
6325 | * systemd-cryptsetup will now attempt to load the keys to unlock | |
6326 | volumes with automatically from files in | |
6327 | /etc/cryptsetup-keys.d/<volume>.key and | |
6328 | /run/cryptsetup-keys.d/<volume>.key, if any of these files exist. | |
6329 | ||
7f56c26d ZJS |
6330 | * systemd-cryptsetup may now activate Microsoft BitLocker volumes via |
6331 | /etc/crypttab, during boot. | |
6332 | ||
b0d0e0ef LP |
6333 | * logind.conf gained a new RuntimeDirectoryInodesMax= setting to |
6334 | control the inode limit for the per-user $XDG_RUNTIME_DIR tmpfs | |
6335 | instance. | |
6336 | ||
b0d0e0ef | 6337 | * A new generator systemd-xdg-autostart-generator has been added. It |
7f56c26d ZJS |
6338 | generates systemd unit files from XDG autostart .desktop files, and |
6339 | may be used to let the systemd user instance manage services that are | |
6340 | started automatically as part of the desktop session. | |
47373451 | 6341 | |
cb713f16 | 6342 | * "bootctl" gained a new verb "reboot-to-firmware" that may be used |
6ad3dc40 ZJS |
6343 | to query and change the firmware's 'Reboot Into Firmware Interface' |
6344 | setup flag. | |
1f19ae0f | 6345 | |
7f56c26d ZJS |
6346 | * systemd-firstboot gained a new switch --kernel-command-line= that may |
6347 | be used to initialize the /etc/kernel/cmdline file of the image. It | |
6348 | also gained a new switch --root-password-hashed= which is like | |
6349 | --root-password= but accepts a pre-hashed UNIX password as | |
6350 | argument. The new option --delete-root-password may be used to unset | |
495454f4 LP |
6351 | any password for the root user (dangerous!). The --root-shell= switch |
6352 | may be used to control the shell to use for the root account. A new | |
6353 | --force option may be used to override any already set settings with | |
6354 | the parameters specified on the command line (by default, the tool | |
6355 | will not override what has already been set before, i.e. is purely | |
6356 | incremental). | |
1f19ae0f | 6357 | |
a38c68a1 LP |
6358 | * systemd-firstboot gained support for a new --image= switch, which is |
6359 | similar to --root= but accepts the path to a disk image file, on | |
6360 | which it then operates. | |
6361 | ||
7f56c26d ZJS |
6362 | * A new sd-path.h API has been added to libsystemd. It provides a |
6363 | simple API for retrieving various search paths and primary | |
6364 | directories for various resources. | |
1f19ae0f | 6365 | |
7f56c26d ZJS |
6366 | * A new call sd_notify_barrier() has been added to the sd-daemon.h |
6367 | API. The call will block until all previously sent sd_notify() | |
6368 | messages have been processed by the service manager. This is useful | |
6369 | to remove races caused by a process already having disappeared at the | |
6370 | time a notification message is processed by the service manager, | |
6371 | making correct attribution impossible. The systemd-notify tool will | |
6372 | now make use of this call implicitly, but this can be turned off again | |
6373 | via the new --no-block switch. | |
1f19ae0f | 6374 | |
7f56c26d ZJS |
6375 | * When sending a file descriptor (fd) to the service manager to keep |
6376 | track of, using the sd_notify() mechanism, a new parameter FDPOLL=0 | |
6377 | may be specified. If passed the service manager will refrain from | |
6378 | poll()ing on the file descriptor. Traditionally (and when the | |
6379 | parameter is not specified), the service manager will poll it for | |
6380 | POLLHUP or POLLERR events, and immediately close the fds in that | |
6381 | case. | |
1f19ae0f LP |
6382 | |
6383 | * The service manager (PID1) gained a new D-Bus method call | |
6384 | SetShowStatus() which may be used to control whether it shall show | |
6385 | boot-time status output on the console. This method has a similar | |
6386 | effect to sending SIGRTMIN+20/SIGRTMIN+21 to PID 1. | |
6387 | ||
7f56c26d ZJS |
6388 | * The sd-bus API gained a number of convenience functions that take |
6389 | va_list arguments rather than "...". For example, there's now | |
6390 | sd_bus_call_methodv() to match sd_bus_call_method(). Those calls make | |
6391 | it easier to build wrappers that accept variadic arguments and want | |
6392 | to pass a ready va_list structure to sd-bus. | |
1f19ae0f | 6393 | |
7f56c26d ZJS |
6394 | * sd-bus vtable entries can have a new SD_BUS_VTABLE_ABSOLUTE_OFFSET |
6395 | flag which alters how the userdata pointer to pass to the callbacks | |
6396 | is determined. When the flag is set, the offset field is converted | |
6397 | as-is into a pointer, without adding it to the object pointer the | |
6398 | vtable is associated with. | |
1f19ae0f | 6399 | |
7f56c26d ZJS |
6400 | * sd-bus now exposes four new functions: |
6401 | sd_bus_interface_name_is_valid() + sd_bus_service_name_is_valid() + | |
6402 | sd_bus_member_name_is_valid() + sd_bus_object_path_is_valid() will | |
6403 | validate strings to check if they qualify as various D-Bus concepts. | |
1f19ae0f | 6404 | |
7f56c26d ZJS |
6405 | * The sd-bus API gained the SD_BUS_METHOD_WITH_ARGS(), |
6406 | SD_BUS_METHOD_WITH_ARGS_OFFSET() and SD_BUS_SIGNAL_WITH_ARGS() macros | |
6407 | that simplify adding argument names to D-Bus methods and signals. | |
1f19ae0f | 6408 | |
7f56c26d | 6409 | * The man pages for the sd-bus and sd-hwdb APIs have been completed. |
1f19ae0f | 6410 | |
7f56c26d ZJS |
6411 | * Various D-Bus APIs of systemd daemons now have man pages that |
6412 | document the methods, signals and properties. | |
1f19ae0f | 6413 | |
7f56c26d | 6414 | * The expectations on user/group name syntax are now documented in |
5149990e CR |
6415 | detail; documentation on how classic home directories may be |
6416 | converted into home directories managed by homed has been added; | |
6417 | documentation regarding integration of homed/userdb functionality in | |
6418 | desktops has been added: | |
7f56c26d ZJS |
6419 | |
6420 | https://systemd.io/USER_NAMES | |
6421 | https://systemd.io/CONVERTING_TO_HOMED | |
6422 | https://systemd.io/USERDB_AND_DESKTOPS | |
6423 | ||
6424 | * Documentation for the on-disk Journal file format has been updated | |
6425 | and has now moved to: | |
6426 | ||
6427 | https://systemd.io/JOURNAL_FILE_FORMAT | |
6428 | ||
5961d35a ZJS |
6429 | * The interface for containers (https://systemd.io/CONTAINER_INTERFACE) |
6430 | has been extended by a set of environment variables that expose | |
6431 | select fields from the host's os-release file to the container | |
6432 | payload. Similarly, host's os-release files can be mounted into the | |
54971969 | 6433 | container underneath /run/host. Together, those mechanisms provide a |
5961d35a ZJS |
6434 | standardized way to expose information about the host to the |
6435 | container payload. Both interfaces are implemented in systemd-nspawn. | |
6436 | ||
7f56c26d ZJS |
6437 | * All D-Bus services shipped in systemd now implement the generic |
6438 | LogControl1 D-Bus API which allows clients to change log level + | |
6439 | target of the service during runtime. | |
1f19ae0f | 6440 | |
337f0b00 ZJS |
6441 | * Only relevant for developers: the mkosi.default symlink has been |
6442 | dropped from version control. Please create a symlink to one of the | |
6443 | distribution-specific defaults in .mkosi/ based on your preference. | |
4c967576 | 6444 | |
72e51908 | 6445 | Contributions from: 24bisquitz, Adam Nielsen, Alan Perry, Alexander |
ae366f3a ZJS |
6446 | Malafeev, Amitanand.Chikorde, Alin Popa, Alvin Šipraga, Amos Bird, |
6447 | Andreas Rammhold, AndreRH, Andrew Doran, Anita Zhang, Ankit Jain, | |
6448 | antznin, Arnaud Ferraris, Arthur Moraes do Lago, Arusekk, Balaji | |
6449 | Punnuru, Balint Reczey, Bastien Nocera, bemarek, Benjamin Berg, | |
6450 | Benjamin Dahlhoff, Benjamin Robin, Chris Down, Chris Kerr, Christian | |
6451 | Göttsche, Christian Hesse, Christian Oder, Ciprian Hacman, Clinton Roy, | |
6452 | codicodi, Corey Hinshaw, Daan De Meyer, Dana Olson, Dan Callaghan, | |
6453 | Daniel Fullmer, Daniel Rusek, Dan Streetman, Dave Reisner, David | |
6454 | Edmundson, David Wood, Denis Pronin, Diego Escalante Urrelo, Dimitri | |
6455 | John Ledkov, dolphrundgren, duguxy, Einsler Lee, Elisei Roca, Emmanuel | |
6456 | Garette, Eric Anderson, Eric DeVolder, Evgeny Vereshchagin, | |
6457 | ExtinctFire, fangxiuning, Ferran Pallarès Roca, Filipe Brandenburger, | |
6458 | Filippo Falezza, Finn, Florian Klink, Florian Mayer, Franck Bui, | |
6459 | Frantisek Sumsal, gaurav, Georg Müller, Gergely Polonkai, Giedrius | |
6460 | Statkevičius, Gigadoc2, gogogogi, Gaurav Singh, gzjsgdsb, Hans de | |
6461 | Goede, Haochen Tong, ianhi, ignapk, Jakov Smolic, James T. Lee, Jan | |
6462 | Janssen, Jan Klötzke, Jan Palus, Jay Burger, Jeremy Cline, Jérémy | |
6463 | Rosen, Jian-Hong Pan, Jiri Slaby, Joel Shapiro, Joerg Behrmann, Jörg | |
6464 | Thalheim, Jouke Witteveen, Kai-Heng Feng, Kenny Levinsen, Kevin | |
6465 | Kuehler, Kumar Kartikeya Dwivedi, layderv, laydervus, Lénaïc Huard, | |
6466 | Lennart Poettering, Lidong Zhong, Luca Boccassi, Luca BRUNO, Lucas | |
6467 | Werkmeister, Lukas Klingsbo, Lukáš Nykrýn, Łukasz Stelmach, Maciej | |
6468 | S. Szmigiero, MadMcCrow, Marc-André Lureau, Marcel Holtmann, Marc | |
6469 | Kleine-Budde, Martin Hundebøll, Matthew Leeds, Matt Ranostay, Maxim | |
6470 | Fomin, MaxVerevkin, Michael Biebl, Michael Chapman, Michael Gubbels, | |
6471 | Michael Marley, Michał Bartoszkiewicz, Michal Koutný, Michal Sekletár, | |
6472 | Mike Gilbert, Mike Kazantsev, Mikhail Novosyolov, ml, Motiejus Jakštys, | |
6473 | nabijaczleweli, nerdopolis, Niccolò Maggioni, Niklas Hambüchen, Norbert | |
6474 | Lange, Paul Cercueil, pelzvieh, Peter Hutterer, Piero La Terza, Pieter | |
6475 | Lexis, Piotr Drąg, Rafael Fontenelle, Richard Petri, Ronan Pigott, Ross | |
6476 | Lagerwall, Rubens Figueiredo, satmandu, Sean-StarLabs, Sebastian | |
6477 | Jennen, sterlinghughes, Surhud More, Susant Sahani, szb512, Thomas | |
6478 | Haller, Tobias Hunger, Tom, Tomáš Pospíšek, Tomer Shechner, Tom Hughes, | |
6479 | Topi Miettinen, Tudor Roman, Uwe Kleine-König, Valery0xff, Vito Caputo, | |
6480 | Vladimir Panteleev, Vladyslav Tronko, Wen Yang, Yegor Vialov, Yigal | |
6481 | Korman, Yi Gao, YmrDtnJu, Yuri Chornoivan, Yu Watanabe, Zbigniew | |
6482 | Jędrzejewski-Szmek, Zhu Li, Дамјан Георгиевски, наб | |
6483 | ||
6484 | – Warsaw, 2020-07-30 | |
976efe1b | 6485 | |
901d1ce8 | 6486 | CHANGES WITH 245: |
723822f0 | 6487 | |
68410195 | 6488 | * A new tool "systemd-repart" has been added, that operates as an |
2ad98889 ZJS |
6489 | idempotent declarative repartitioner for GPT partition tables. |
6490 | Specifically, a set of partitions that must or may exist can be | |
6491 | configured via drop-in files, and during every boot the partition | |
6492 | table on disk is compared with these files, creating missing | |
6493 | partitions or growing existing ones based on configurable relative | |
6494 | and absolute size constraints. The tool is strictly incremental, | |
6495 | i.e. does not delete, shrink or move partitions, but only adds and | |
6496 | grows them. The primary use-case is OS images that ship in minimized | |
6497 | form, that on first boot are grown to the size of the underlying | |
6498 | block device or augmented with additional partitions. For example, | |
6499 | the root partition could be extended to cover the whole disk, or a | |
6500 | swap or /home partitions could be added on first boot. It can also be | |
6501 | used for systems that use an A/B update scheme but ship images with | |
6502 | just the A partition, with B added on first boot. The tool is | |
6503 | primarily intended to be run in the initrd, shortly before | |
6504 | transitioning into the host OS, but can also be run after the | |
6505 | transition took place. It automatically discovers the disk backing | |
6506 | the root file system, and should hence not require any additional | |
6507 | configuration besides the partition definition drop-ins. If no | |
6508 | configuration drop-ins are present, no action is taken. | |
68410195 LP |
6509 | |
6510 | * A new component "userdb" has been added, along with a small daemon | |
a0223c30 | 6511 | "systemd-userdbd.service" and a client tool "userdbctl". The framework |
68410195 LP |
6512 | allows defining rich user and group records in a JSON format, |
6513 | extending on the classic "struct passwd" and "struct group" | |
6514 | structures. Various components in systemd have been updated to | |
6515 | process records in this format, including systemd-logind and | |
6516 | pam-systemd. The user records are intended to be extensible, and | |
6517 | allow setting various resource management, security and runtime | |
6518 | parameters that shall be applied to processes and sessions of the | |
6519 | user as they log in. This facility is intended to allow associating | |
6520 | such metadata directly with user/group records so that they can be | |
6521 | produced, extended and consumed in unified form. We hope that | |
6522 | eventually frameworks such as sssd will generate records this way, so | |
6523 | that for the first time resource management and various other | |
6524 | per-user settings can be configured in LDAP directories and then | |
6525 | provided to systemd (specifically to systemd-logind and pam-system) | |
2ad98889 | 6526 | to apply on login. For further details see: |
68410195 LP |
6527 | |
6528 | https://systemd.io/USER_RECORD | |
6529 | https://systemd.io/GROUP_RECORD | |
6530 | https://systemd.io/USER_GROUP_API | |
6531 | ||
9a4940bf | 6532 | * A small new service systemd-homed.service has been added, that may be |
2ad98889 ZJS |
6533 | used to securely manage home directories with built-in encryption. |
6534 | The complete user record data is unified with the home directory, | |
6535 | thus making home directories naturally migratable. Its primary | |
6536 | back-end is based on LUKS volumes, but fscrypt, plain directories, | |
6537 | and other storage schemes are also supported. This solves a couple of | |
6538 | problems we saw with traditional ways to manage home directories, in | |
6539 | particular when it comes to encryption. For further discussion of | |
6540 | this, see the video of Lennart's talk at AllSystemsGo! 2019: | |
9a4940bf LP |
6541 | |
6542 | https://media.ccc.de/v/ASG2019-164-reinventing-home-directories | |
6543 | ||
6544 | For further details about the format and expectations on home | |
6545 | directories this new daemon makes, see: | |
6546 | ||
6547 | https://systemd.io/HOME_DIRECTORY | |
6548 | ||
6549 | * systemd-journald is now multi-instantiable. In addition to the main | |
6550 | instance systemd-journald.service there's now a template unit | |
2ad98889 ZJS |
6551 | systemd-journald@.service, with each instance defining a new named |
6552 | log 'namespace' (whose name is specified via the instance part of the | |
6553 | unit name). A new unit file setting LogNamespace= has been added, | |
6554 | taking such a namespace name, that assigns services to the specified | |
6555 | log namespaces. As each log namespace is serviced by its own | |
6556 | independent journal daemon, this functionality may be used to improve | |
6557 | performance and increase isolation of applications, at the price of | |
6558 | losing global message ordering. Each instance of journald has a | |
6559 | separate set of configuration files, with possibly different disk | |
6560 | usage limitations and other settings. | |
6561 | ||
6562 | journalctl now takes a new option --namespace= to show logs from a | |
6563 | specific log namespace. The sd-journal.h API gained | |
6564 | sd_journal_open_namespace() for opening the log stream of a specific | |
6565 | log namespace. systemd-journald also gained the ability to exit on | |
6566 | idle, which is useful in the context of log namespaces, as this means | |
6567 | log daemons for log namespaces can be activated automatically on | |
6568 | demand and will stop automatically when no longer used, minimizing | |
6569 | resource usage. | |
9a4940bf | 6570 | |
723822f0 | 6571 | * When systemd-tmpfiles copies a file tree using the 'C' line type it |
2ad98889 | 6572 | will now label every copied file according to the SELinux database. |
723822f0 LP |
6573 | |
6574 | * When systemd/PID 1 detects it is used in the initrd it will now boot | |
6575 | into initrd.target rather than default.target by default. This should | |
6576 | make it simpler to build initrds with systemd as for many cases the | |
6577 | only difference between a host OS image and an initrd image now is | |
2ad98889 | 6578 | the presence of the /etc/initrd-release file. |
723822f0 LP |
6579 | |
6580 | * A new kernel command line option systemd.cpu_affinity= is now | |
6581 | understood. It's equivalent to the CPUAffinity= option in | |
6582 | /etc/systemd/system.conf and allows setting the CPU mask for PID 1 | |
2ad98889 | 6583 | itself and the default for all other processes. |
723822f0 | 6584 | |
2ad98889 ZJS |
6585 | * When systemd/PID 1 is reloaded (with systemctl daemon-reload or |
6586 | equivalent), the SELinux database is now reloaded, ensuring that | |
723822f0 LP |
6587 | sockets and other file system objects are generated taking the new |
6588 | database into account. | |
6589 | ||
d54bb638 ZJS |
6590 | * systemd/PID 1 accepts a new "systemd.show-status=error" setting, and |
6591 | "quiet" has been changed to imply that instead of | |
6592 | "systemd.show-status=auto". In this mode, only messages about errors | |
6593 | and significant delays in boot are shown on the console. | |
6594 | ||
2ad98889 | 6595 | * The sd-event.h API gained native support for the new Linux "pidfd" |
723822f0 LP |
6596 | concept. This permits watching processes using file descriptors |
6597 | instead of PID numbers, which fixes a number of races and makes | |
2ad98889 | 6598 | process supervision more robust and efficient. All of systemd's |
723822f0 LP |
6599 | components will now use pidfds if the kernel supports it for process |
6600 | watching, with the exception of PID 1 itself, unfortunately. We hope | |
6601 | to move PID 1 to exclusively using pidfds too eventually, but this | |
6602 | requires some more kernel work first. (Background: PID 1 watches | |
6603 | processes using waitid() with the P_ALL flag, and that does not play | |
6604 | together nicely with pidfds yet.) | |
6605 | ||
6606 | * Closely related to this, the sd-event.h API gained two new calls | |
6607 | sd_event_source_send_child_signal() (for sending a signal to a | |
6608 | watched process) and sd_event_source_get_child_process_own() (for | |
2ad98889 ZJS |
6609 | marking a process so that it is killed automatically whenever the |
6610 | event source watching it is freed). | |
723822f0 | 6611 | |
60ed2dcf | 6612 | * systemd-networkd gained support for configuring Token Bucket Filter |
2ad98889 ZJS |
6613 | (TBF) parameters in its qdisc configuration support. Similarly, |
6614 | support for Stochastic Fairness Queuing (SFQ), Controlled-Delay | |
69f17347 | 6615 | Active Queue Management (CoDel), and Fair Queue (FQ) has been added. |
723822f0 LP |
6616 | |
6617 | * systemd-networkd gained support for Intermediate Functional Block | |
6618 | (IFB) network devices. | |
6619 | ||
6620 | * systemd-networkd gained support for configuring multi-path IP routes, | |
6621 | using the new MultiPathRoute= setting in the [Route] section. | |
6622 | ||
2ad98889 ZJS |
6623 | * systemd-networkd's DHCPv4 client has been updated to support a new |
6624 | SendDecline= option. If enabled, duplicate address detection is done | |
6625 | after a DHCP offer is received from the server. If a conflict is | |
6626 | detected, the address is declined. The DHCPv4 client also gained | |
723822f0 LP |
6627 | support for a new RouteMTUBytes= setting that allows to configure the |
6628 | MTU size to be used for routes generated from DHCPv4 leases. | |
6629 | ||
6630 | * The PrefixRoute= setting in systemd-networkd's [Address] section of | |
6631 | .network files has been deprecated, and replaced by AddPrefixRoute=, | |
2ad98889 | 6632 | with its sense inverted. |
723822f0 LP |
6633 | |
6634 | * The Gateway= setting of [Route] sections of .network files gained | |
427928ca ZJS |
6635 | support for a special new value "_dhcp". If set, the configured |
6636 | static route uses the gateway host configured via DHCP. | |
723822f0 | 6637 | |
573e58f6 | 6638 | * New User= and SuppressPrefixLength= settings have been implemented |
f05c0615 YW |
6639 | for the [RoutingPolicyRule] section of .network files to configure |
6640 | source routing based on UID ranges and prefix length, respectively. | |
9a4940bf | 6641 | |
88b2a950 ZJS |
6642 | * The Type= match property of .link files has been generalized to |
6643 | always match the device type shown by 'networkctl status', even for | |
6644 | devices where udev does not set DEVTYPE=. This allows e.g. Type=ether | |
6645 | to be used. | |
6646 | ||
2ad98889 ZJS |
6647 | * sd-bus gained a new API call sd_bus_message_sensitive() that marks a |
6648 | D-Bus message object as "sensitive". Those objects are erased from | |
6649 | memory when they are freed. This concept is intended to be used for | |
6650 | messages that contain security sensitive data. A new flag | |
6651 | SD_BUS_VTABLE_SENSITIVE has been introduced as well to mark methods | |
6652 | in sd-bus vtables, causing any incoming and outgoing messages of | |
6653 | those methods to be implicitly marked as "sensitive". | |
723822f0 | 6654 | |
9a4940bf | 6655 | * sd-bus gained a new API call sd_bus_message_dump() for dumping the |
2ad98889 | 6656 | contents of a message (or parts thereof) to standard output for |
9a4940bf LP |
6657 | debugging purposes. |
6658 | ||
2ad98889 ZJS |
6659 | * systemd-sysusers gained support for creating users with the primary |
6660 | group named differently than the user. | |
9a4940bf | 6661 | |
723822f0 LP |
6662 | * systemd-growfs (i.e. the x-systemd.growfs mount option in /etc/fstab) |
6663 | gained support for growing XFS partitions. Previously it supported | |
6664 | only ext4 and btrfs partitions. | |
6665 | ||
6666 | * The support for /etc/crypttab gained a new x-initrd.attach option. If | |
2ad98889 ZJS |
6667 | set, the specified encrypted volume is unlocked already in the |
6668 | initrd. This concept corresponds to the x-initrd.mount option in | |
723822f0 LP |
6669 | /etc/fstab. |
6670 | ||
6671 | * systemd-cryptsetup gained native support for unlocking encrypted | |
6672 | volumes utilizing PKCS#11 smartcards, i.e. for example to bind | |
2ad98889 | 6673 | encryption of volumes to YubiKeys. This is exposed in the new |
723822f0 LP |
6674 | pkcs11-uri= option in /etc/crypttab. |
6675 | ||
68410195 LP |
6676 | * The /etc/fstab support in systemd now supports two new mount options |
6677 | x-systemd.{required,wanted}-by=, for explicitly configuring the units | |
6678 | that the specified mount shall be pulled in by, in place of | |
6679 | the usual local-fs.target/remote-fs.target. | |
6680 | ||
723822f0 LP |
6681 | * The https://systemd.io/ web site has been relaunched, directly |
6682 | populated with most of the documentation included in the systemd | |
2ad98889 ZJS |
6683 | repository. systemd also acquired a new logo, thanks to Tobias |
6684 | Bernard. | |
723822f0 LP |
6685 | |
6686 | * systemd-udevd gained support for managing "alternative" network | |
6687 | interface names, as supported by new Linux kernels. For the first | |
6688 | time this permits assigning multiple (and longer!) names to a network | |
6689 | interface. systemd-udevd will now by default assign the names | |
2ad98889 ZJS |
6690 | generated via all supported naming schemes to each interface. This |
6691 | may be further tweaked with .link files and the AlternativeName= and | |
6692 | AlternativeNamesPolicy= settings. Other components of systemd have | |
6693 | been updated to support the new alternative names wherever | |
6694 | appropriate. For example, systemd-nspawn will now generate | |
6695 | alternative interface names for the host-facing side of container | |
6696 | veth links based on the full container name without truncation. | |
723822f0 LP |
6697 | |
6698 | * systemd-nspawn interface naming logic has been updated in another way | |
6699 | too: if the main interface name (i.e. as opposed to new-style | |
2ad98889 ZJS |
6700 | "alternative" names) based on the container name is truncated, a |
6701 | simple hashing scheme is used to give different interface names to | |
6702 | multiple containers whose names all begin with the same prefix. Since | |
6703 | this changes the primary interface names pointing to containers if | |
6704 | truncation happens, the old scheme may still be requested by | |
6705 | selecting an older naming scheme, via the net.naming-scheme= kernel | |
6706 | command line option. | |
723822f0 | 6707 | |
68410195 LP |
6708 | * PrivateUsers= in service files now works in services run by the |
6709 | systemd --user per-user instance of the service manager. | |
6710 | ||
6711 | * A new per-service sandboxing option ProtectClock= has been added that | |
6712 | locks down write access to the system clock. It takes away device | |
2ad98889 ZJS |
6713 | node access to /dev/rtc as well as the system calls that set the |
6714 | system clock and the CAP_SYS_TIME and CAP_WAKE_ALARM capabilities. | |
6715 | Note that this option does not affect access to auxiliary services | |
6716 | that allow changing the clock, for example access to | |
6717 | systemd-timedated. | |
68410195 LP |
6718 | |
6719 | * The systemd-id128 tool gained a new "show" verb for listing or | |
da890466 | 6720 | resolving a number of well-known UUIDs/128-bit IDs, currently mostly |
68410195 | 6721 | GPT partition table types. |
723822f0 LP |
6722 | |
6723 | * The Discoverable Partitions Specification has been updated to support | |
6724 | /var and /var/tmp partition discovery. Support for this has been | |
6725 | added to systemd-gpt-auto-generator. For details see: | |
6726 | ||
6727 | https://systemd.io/DISCOVERABLE_PARTITIONS | |
6728 | ||
6729 | * "systemctl list-unit-files" has been updated to show a new column | |
6730 | with the suggested enablement state based on the vendor preset files | |
6731 | for the respective units. | |
6732 | ||
68410195 LP |
6733 | * "systemctl" gained a new option "--with-dependencies". If specified |
6734 | commands such as "systemctl status" or "systemctl cat" will now show | |
6735 | all specified units along with all units they depend on. | |
6736 | ||
723822f0 LP |
6737 | * networkctl gained support for showing per-interface logs in its |
6738 | "status" output. | |
6739 | ||
a100fe3c | 6740 | * systemd-networkd-wait-online gained support for specifying the maximum |
573e58f6 ZJS |
6741 | operational state to wait for, and to wait for interfaces to |
6742 | disappear. | |
6743 | ||
723822f0 LP |
6744 | * The [Match] section of .link and .network files now supports a new |
6745 | option PermanentMACAddress= which may be used to check against the | |
6746 | permanent MAC address of a network device even if a randomized MAC | |
6747 | address is used. | |
6748 | ||
69f17347 ZJS |
6749 | * The [TrafficControlQueueingDiscipline] section in .network files has |
6750 | been renamed to [NetworkEmulator] with the "NetworkEmulator" prefix | |
6751 | dropped from the individual setting names. | |
68410195 | 6752 | |
823b0352 ZJS |
6753 | * Any .link and .network files that have an empty [Match] section (this |
6754 | also includes empty and commented-out files) will now be | |
6755 | rejected. systemd-udev and systemd-networkd started warning about | |
6756 | such files in version 243. | |
6757 | ||
2ad98889 | 6758 | * systemd-logind will now validate access to the operation of changing |
98ab0dae | 6759 | the virtual terminal via a polkit action. By default, only users |
2ad98889 | 6760 | with at least one session on a local VT are granted permission. |
68410195 | 6761 | |
2ad98889 ZJS |
6762 | * When systemd sets up PAM sessions that invoked service processes |
6763 | shall run in, the pam_setcred() API is now invoked, thus permitting | |
6764 | PAM modules to set additional credentials for the processes. | |
68410195 | 6765 | |
d54bb638 ZJS |
6766 | * portablectl attach/detach verbs now accept --now and --enable options |
6767 | to combine attachment with enablement and invocation, or detachment | |
6768 | with stopping and disablement. | |
6769 | ||
2ca17c78 ZJS |
6770 | * UPGRADE ISSUE: a bug where some jobs were trimmed as redundant was |
6771 | fixed, which in turn exposed bugs in unit configuration of services | |
6772 | which have Type=oneshot and should only run once, but do not have | |
6773 | RemainAfterExit=yes set. Without RemainAfterExit=yes, a one-shot | |
6774 | service may be started again after exiting successfully, for example | |
6775 | as a dependency in another transaction. Affected services included | |
6776 | some internal systemd services (most notably | |
6777 | systemd-vconsole-setup.service, which was updated to have | |
6778 | RemainAfterExit=yes), and plymouth-start.service. Please ensure that | |
6779 | plymouth has been suitably updated or patched before upgrading to | |
6780 | this systemd release. See | |
6781 | https://bugzilla.redhat.com/show_bug.cgi?id=1807771 for some | |
6782 | additional discussion. | |
6783 | ||
9c4d3d79 ZJS |
6784 | Contributions from: AJ Bagwell, Alin Popa, Andreas Rammhold, Anita |
6785 | Zhang, Ansgar Burchardt, Antonio Russo, Arian van Putten, Ashley Davis, | |
6786 | Balint Reczey, Bart Willems, Bastien Nocera, Benjamin Dahlhoff, Charles | |
6787 | (Chas) Williams, cheese1, Chris Down, Chris Murphy, Christian Ehrhardt, | |
6788 | Christian Göttsche, cvoinf, Daan De Meyer, Daniele Medri, Daniel Rusek, | |
6789 | Daniel Shahaf, Dann Frazier, Dan Streetman, Dariusz Gadomski, David | |
6790 | Michael, Dimitri John Ledkov, Emmanuel Bourg, Evgeny Vereshchagin, | |
6791 | ezst036, Felipe Sateler, Filipe Brandenburger, Florian Klink, Franck | |
6792 | Bui, Fran Dieguez, Frantisek Sumsal, Greg "GothAck" Miell, Guilhem | |
6793 | Lettron, Guillaume Douézan-Grard, Hans de Goede, HATAYAMA Daisuke, Iain | |
6794 | Lane, James Buren, Jan Alexander Steffens (heftig), Jérémy Rosen, Jin | |
6795 | Park, Jun'ichi Nomura, Kai Krakow, Kevin Kuehler, Kevin P. Fleming, | |
6796 | Lennart Poettering, Leonid Bloch, Leonid Evdokimov, lothrond, Luca | |
6797 | Boccassi, Lukas K, Lynn Kirby, Mario Limonciello, Mark Deneen, Matthew | |
6798 | Leeds, Michael Biebl, Michal Koutný, Michal Sekletár, Mike Auty, Mike | |
6799 | Gilbert, mtron, nabijaczleweli, Naïm Favier, Nate Jones, Norbert Lange, | |
6800 | Oliver Giles, Paul Davey, Paul Menzel, Peter Hutterer, Piotr Drąg, Rafa | |
6801 | Couto, Raphael, rhn, Robert Scheck, Rocka, Romain Naour, Ryan Attard, | |
6802 | Sascha Dewald, Shengjing Zhu, Slava Kardakov, Spencer Michaels, Sylvain | |
6803 | Plantefeve, Stanislav Angelovič, Susant Sahani, Thomas Haller, Thomas | |
6804 | Schmitt, Timo Schlüßler, Timo Wilken, Tobias Bernard, Tobias Klauser, | |
6805 | Tobias Stoeckmann, Topi Miettinen, tsia, WataruMatsuoka, Wieland | |
6806 | Hoffmann, Wilhelm Schuster, Will Fleming, xduugu, Yong Cong Sin, Yuri | |
6807 | Chornoivan, Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek, Zeyu | |
6808 | DONG | |
901d1ce8 | 6809 | |
03985d06 | 6810 | – Warsaw, 2020-03-06 |
723822f0 | 6811 | |
bdf2357c | 6812 | CHANGES WITH 244: |
ee50dada ZJS |
6813 | |
6814 | * Support for the cpuset cgroups v2 controller has been added. | |
6815 | Processes may be restricted to specific CPUs using the new | |
6816 | AllowedCPUs= setting, and to specific memory NUMA nodes using the new | |
6817 | AllowedMemoryNodes= setting. | |
6818 | ||
6819 | * The signal used in restart jobs (as opposed to e.g. stop jobs) may | |
1e904320 | 6820 | now be configured using a new RestartKillSignal= setting. This |
ee50dada ZJS |
6821 | allows units which signals to request termination to implement |
6822 | different behaviour when stopping in preparation for a restart. | |
6823 | ||
6824 | * "systemctl clean" may now be used also for socket, mount, and swap | |
6825 | units. | |
6826 | ||
6827 | * systemd will also read configuration options from the EFI variable | |
6828 | SystemdOptions. This may be used to configure systemd behaviour when | |
6829 | modifying the kernel command line is inconvenient, but configuration | |
6830 | on disk is read too late, for example for the options related to | |
2536752d | 6831 | cgroup hierarchy setup. 'bootctl systemd-efi-options' may be used to |
ee50dada ZJS |
6832 | set the EFI variable. |
6833 | ||
6834 | * systemd will now disable printk ratelimits in early boot. This should | |
6835 | allow us to capture more logs from the early boot phase where normal | |
6836 | storage is not available and the kernel ring buffer is used for | |
6837 | logging. Configuration on the kernel command line has higher priority | |
6838 | and overrides the systemd setting. | |
6839 | ||
6840 | systemd programs which log to /dev/kmsg directly use internal | |
6841 | ratelimits to prevent runaway logging. (Normally this is only used | |
6842 | during early boot, so in practice this change has very little | |
6843 | effect.) | |
6844 | ||
3e1db806 AZ |
6845 | * Unit files now support top level dropin directories of the form |
6846 | <unit_type>.d/ (e.g. service.d/) that may be used to add configuration | |
6847 | that affects all corresponding unit files. | |
ee50dada | 6848 | |
852b7272 ZJS |
6849 | * systemctl gained support for 'stop --job-mode=triggering' which will |
6850 | stop the specified unit and any units which could trigger it. | |
6851 | ||
6852 | * Unit status display now includes units triggering and triggered by | |
6853 | the unit being shown. | |
6854 | ||
ee50dada ZJS |
6855 | * The RuntimeMaxSec= setting is now supported by scopes, not just |
6856 | .service units. This is particularly useful for PAM sessions which | |
6857 | create a scope unit for the user login. systemd.runtime_max_sec= | |
6858 | setting may used with the pam_systemd module to limit the duration | |
6859 | of the PAM session, for example for time-limited logins. | |
6860 | ||
852b7272 | 6861 | * A new @pkey system call group is now defined to make it easier to |
6b000af4 | 6862 | allow-list memory protection syscalls for containers and services |
852b7272 ZJS |
6863 | which need to use them. |
6864 | ||
6865 | * systemd-udevd: removed the 30s timeout for killing stale workers on | |
6866 | exit. systemd-udevd now waits for workers to finish. The hard-coded | |
6867 | exit timeout of 30s was too short for some large installations, where | |
6868 | driver initialization could be prematurely interrupted during initrd | |
6869 | processing if the root file system had been mounted and init was | |
6870 | preparing to switch root. If udevd is run without systemd and workers | |
6871 | are hanging while udevd receives an exit signal, udevd will now exit | |
6872 | when udev.event_timeout is reached for the last hanging worker. With | |
6873 | systemd, the exit timeout can additionally be configured using | |
6874 | TimeoutStopSec= in systemd-udevd.service. | |
6875 | ||
ee50dada ZJS |
6876 | * udev now provides a program (fido_id) that identifies FIDO CTAP1 |
6877 | ("U2F")/CTAP2 security tokens based on the usage declared in their | |
6878 | report and descriptor and outputs suitable environment variables. | |
6b000af4 | 6879 | This replaces the externally maintained allow lists of all known |
ee50dada ZJS |
6880 | security tokens that were used previously. |
6881 | ||
6b000af4 | 6882 | * Automatically generated autosuspend udev rules for allow-listed |
ee50dada | 6883 | devices have been imported from the Chromium OS project. This should |
b7db8b7b | 6884 | improve power saving with many more devices. |
ee50dada ZJS |
6885 | |
6886 | * udev gained a new "CONST{key}=value" setting that allows matching | |
6887 | against system-wide constants without forking a helper binary. | |
6888 | Currently "arch" and "virt" keys are supported. | |
6889 | ||
6890 | * udev now opens CDROMs in non-exclusive mode when querying their | |
6891 | capabilities. This should fix issues where other programs trying to | |
6892 | use the CDROM cannot gain access to it, but carries a risk of | |
6893 | interfering with programs writing to the disk, if they did not open | |
6894 | the device in exclusive mode as they should. | |
6895 | ||
6896 | * systemd-networkd does not create a default route for IPv4 link local | |
6897 | addressing anymore. The creation of the route was unexpected and was | |
6898 | breaking routing in various cases, but people who rely on it being | |
6899 | created implicitly will need to adjust. Such a route may be requested | |
6900 | with DefaultRouteOnDevice=yes. | |
6901 | ||
6902 | Similarly, systemd-networkd will not assign a link-local IPv6 address | |
6903 | when IPv6 link-local routing is not enabled. | |
6904 | ||
6905 | * Receive and transmit buffers may now be configured on links with | |
6906 | the new RxBufferSize= and TxBufferSize= settings. | |
6907 | ||
6908 | * systemd-networkd may now advertise additional IPv6 routes. A new | |
6909 | [IPv6RoutePrefix] section with Route= and LifetimeSec= options is | |
6910 | now supported. | |
6911 | ||
6912 | * systemd-networkd may now configure "next hop" routes using the | |
6913 | [NextHop] section and Gateway= and Id= settings. | |
6914 | ||
6915 | * systemd-networkd will now retain DHCP config on restarts by default | |
6916 | (but this may be overridden using the KeepConfiguration= setting). | |
6917 | The default for SendRelease= has been changed to true. | |
6918 | ||
6919 | * The DHCPv4 client now uses the OPTION_INFORMATION_REFRESH_TIME option | |
6920 | received from the server. | |
6921 | ||
6922 | The client will use the received SIP server list if UseSIP=yes is | |
6923 | set. | |
6924 | ||
6925 | The client may be configured to request specific options from the | |
6926 | server using a new RequestOptions= setting. | |
6927 | ||
852b7272 ZJS |
6928 | The client may be configured to send arbitrary options to the server |
6929 | using a new SendOption= setting. | |
6930 | ||
ee50dada ZJS |
6931 | A new IPServiceType= setting has been added to configure the "IP |
6932 | service type" value used by the client. | |
6933 | ||
6934 | * The DHCPv6 client learnt a new PrefixDelegationHint= option to | |
6935 | request prefix hints in the DHCPv6 solicitation. | |
6936 | ||
852b7272 | 6937 | * The DHCPv4 server may be configured to send arbitrary options using |
88b86003 | 6938 | a new SendOption= setting. |
852b7272 ZJS |
6939 | |
6940 | * The DHCPv4 server may now be configured to emit SIP server list using | |
6941 | the new EmitSIP= and SIP= settings. | |
6942 | ||
ee50dada ZJS |
6943 | * systemd-networkd and networkctl may now renew DHCP leases on demand. |
6944 | networkctl has a new 'networkctl renew' verb. | |
6945 | ||
6946 | * systemd-networkd may now reconfigure links on demand. networkctl | |
6947 | gained two new verbs: "reload" will reload the configuration, and | |
6948 | "reconfigure DEVICE…" will reconfigure one or more devices. | |
6949 | ||
6950 | * .network files may now match on SSID and BSSID of a wireless network, | |
6951 | i.e. the access point name and hardware address using the new SSID= | |
6952 | and BSSID= options. networkctl will display the current SSID and | |
6953 | BSSID for wireless links. | |
6954 | ||
6955 | .network files may also match on the wireless network type using the | |
f36e6a4a | 6956 | new WLANInterfaceType= option. |
ee50dada ZJS |
6957 | |
6958 | * systemd-networkd now includes default configuration that enables | |
6959 | link-local addressing when connected to an ad-hoc wireless network. | |
6960 | ||
ee50dada ZJS |
6961 | * systemd-networkd may configure the Traffic Control queueing |
6962 | disciplines in the kernel using the new | |
6963 | [TrafficControlQueueingDiscipline] section and Parent=, | |
6964 | NetworkEmulatorDelaySec=, NetworkEmulatorDelayJitterSec=, | |
6878c022 YW |
6965 | NetworkEmulatorPacketLimit=, NetworkEmulatorLossRate=, |
6966 | NetworkEmulatorDuplicateRate= settings. | |
ee50dada ZJS |
6967 | |
6968 | * systemd-tmpfiles gained a new w+ setting to append to files. | |
6969 | ||
6970 | * systemd-analyze dump will now report when the memory configuration in | |
6971 | the kernel does not match what systemd has configured (usually, | |
6972 | because some external program has modified the kernel configuration | |
6973 | on its own). | |
6974 | ||
852b7272 ZJS |
6975 | * systemd-analyze gained a new --base-time= switch instructs the |
6976 | 'calendar' verb to resolve times relative to that timestamp instead | |
6977 | of the present time. | |
6978 | ||
ee50dada ZJS |
6979 | * journalctl --update-catalog now produces deterministic output (making |
6980 | reproducible image builds easier). | |
6981 | ||
6982 | * A new devicetree-overlay setting is now documented in the Boot Loader | |
6983 | Specification. | |
6984 | ||
6985 | * The default value of the WatchdogSec= setting used in systemd | |
6986 | services (the ones bundled with the project itself) may be set at | |
6987 | configuration time using the -Dservice-watchdog= setting. If set to | |
6988 | empty, the watchdogs will be disabled. | |
6989 | ||
0e4daba1 MK |
6990 | * systemd-resolved validates IP addresses in certificates now when GnuTLS |
6991 | is being used. | |
597f905c | 6992 | |
852b7272 ZJS |
6993 | * libcryptsetup >= 2.0.1 is now required. |
6994 | ||
6995 | * A configuration option -Duser-path= may be used to override the $PATH | |
6996 | used by the user service manager. The default is again to use the same | |
6997 | path as the system manager. | |
6998 | ||
168e131b | 6999 | * The systemd-id128 tool gained a new switch "-u" (or "--uuid") for |
da890466 | 7000 | outputting the 128-bit IDs in UUID format (i.e. in the "canonical |
168e131b LP |
7001 | representation"). |
7002 | ||
7003 | * Service units gained a new sandboxing option ProtectKernelLogs= which | |
7004 | makes sure the program cannot get direct access to the kernel log | |
7005 | buffer anymore, i.e. the syslog() system call (not to be confused | |
7006 | with the API of the same name in libc, which is not affected), the | |
7007 | /proc/kmsg and /dev/kmsg nodes and the CAP_SYSLOG capability are made | |
7008 | inaccessible to the service. It's recommended to enable this setting | |
7009 | for all services that should not be able to read from or write to the | |
7010 | kernel log buffer, which are probably almost all. | |
7011 | ||
bdf2357c | 7012 | Contributions from: Aaron Plattner, Alcaro, Anita Zhang, Balint Reczey, |
8bf2a311 LP |
7013 | Bastien Nocera, Baybal Ni, Benjamin Bouvier, Benjamin Gilbert, Carlo |
7014 | Teubner, cbzxt, Chen Qi, Chris Down, Christian Rebischke, Claudio | |
7015 | Zumbo, ClydeByrdIII, crashfistfight, Cyprien Laplace, Daniel Edgecumbe, | |
7016 | Daniel Gorbea, Daniel Rusek, Daniel Stuart, Dan Streetman, David | |
7017 | Pedersen, David Tardon, Dimitri John Ledkov, Dominique Martinet, Donald | |
7018 | A. Cupp Jr, Evgeny Vereshchagin, Fabian Henneke, Filipe Brandenburger, | |
7019 | Franck Bui, Frantisek Sumsal, Georg Müller, Hans de Goede, Haochen | |
7020 | Tong, HATAYAMA Daisuke, Iwan Timmer, Jan Janssen, Jan Kundrát, Jan | |
7021 | Synacek, Jan Tojnar, Jay Strict, Jérémy Rosen, Jóhann B. Guðmundsson, | |
7022 | Jonas Jelten, Jonas Thelemann, Justin Trudell, J. Xing, Kai-Heng Feng, | |
7023 | Kenneth D'souza, Kevin Becker, Kevin Kuehler, Lennart Poettering, | |
7024 | Léonard Gérard, Lorenz Bauer, Luca Boccassi, Maciej Stanczew, Mario | |
7025 | Limonciello, Marko Myllynen, Mark Stosberg, Martin Wilck, matthiasroos, | |
7026 | Michael Biebl, Michael Olbrich, Michael Tretter, Michal Sekletar, | |
7027 | Michal Sekletár, Michal Suchanek, Mike Gilbert, Mike Kazantsev, Nicolas | |
7028 | Douma, nikolas, Norbert Lange, pan93412, Pascal de Bruijn, Paul Menzel, | |
7029 | Pavel Hrdina, Peter Wu, Philip Withnall, Piotr Drąg, Rafael Fontenelle, | |
7030 | Renaud Métrich, Riccardo Schirone, RoadrunnerWMC, Ronan Pigott, Ryan | |
7031 | Attard, Sebastian Wick, Serge, Siddharth Chandrasekara, Steve Ramage, | |
7032 | Steve Traylen, Susant Sahani, Thibault Nélis, Tim Teichmann, Tom | |
7033 | Fitzhenry, Tommy J, Torsten Hilbrich, Vito Caputo, ypf791, Yu Watanabe, | |
7034 | Zach Smith, Zbigniew Jędrzejewski-Szmek | |
7035 | ||
7036 | – Warsaw, 2019-11-29 | |
bdf2357c | 7037 | |
efb536d0 | 7038 | CHANGES WITH 243: |
6af90583 | 7039 | |
0338934f LP |
7040 | * This release enables unprivileged programs (i.e. requiring neither |
7041 | setuid nor file capabilities) to send ICMP Echo (i.e. ping) requests | |
08b59539 | 7042 | by turning on the "net.ipv4.ping_group_range" sysctl of the Linux |
0338934f LP |
7043 | kernel for the whole UNIX group range, i.e. all processes. This |
7044 | change should be reasonably safe, as the kernel support for it was | |
7045 | specifically implemented to allow safe access to ICMP Echo for | |
7046 | processes lacking any privileges. If this is not desirable, it can be | |
7047 | disabled again by setting the parameter to "1 0". | |
7048 | ||
4cd82631 | 7049 | * Previously, filters defined with SystemCallFilter= would have the |
08b59539 CR |
7050 | effect that any calling of an offending system call would terminate |
7051 | the calling thread. This behaviour never made much sense, since | |
7052 | killing individual threads of unsuspecting processes is likely to | |
7053 | create more problems than it solves. With this release the default | |
7054 | action changed from killing the thread to killing the whole | |
7055 | process. For this to work correctly both a kernel version (>= 4.14) | |
7056 | and a libseccomp version (>= 2.4.0) supporting this new seccomp | |
7057 | action is required. If an older kernel or libseccomp is used the old | |
7058 | behaviour continues to be used. This change does not affect any | |
7059 | services that have no system call filters defined, or that use | |
7060 | SystemCallErrorNumber= (and thus see EPERM or another error instead | |
7061 | of being killed when calling an offending system call). Note that | |
7062 | systemd documentation always claimed that the whole process is | |
7063 | killed. With this change behaviour is thus adjusted to match the | |
7064 | documentation. | |
4cd82631 | 7065 | |
4860f5c2 ZJS |
7066 | * On 64 bit systems, the "kernel.pid_max" sysctl is now bumped to |
7067 | 4194304 by default, i.e. the full 22bit range the kernel allows, up | |
da890466 | 7068 | from the old 16-bit range. This should improve security and |
4860f5c2 ZJS |
7069 | robustness, as PID collisions are made less likely (though certainly |
7070 | still possible). There are rumours this might create compatibility | |
6af90583 LP |
7071 | problems, though at this moment no practical ones are known to |
7072 | us. Downstream distributions are hence advised to undo this change in | |
7073 | their builds if they are concerned about maximum compatibility, but | |
7074 | for everybody else we recommend leaving the value bumped. Besides | |
7075 | improving security and robustness this should also simplify things as | |
7076 | the maximum number of allowed concurrent tasks was previously bounded | |
4860f5c2 ZJS |
7077 | by both "kernel.pid_max" and "kernel.threads-max" and now effectively |
7078 | only a single knob is left ("kernel.threads-max"). There have been | |
7079 | concerns that usability is affected by this change because larger PID | |
7080 | numbers are harder to type, but we believe the change from 5 digits | |
7081 | to 7 digits doesn't hamper usability. | |
6af90583 | 7082 | |
2875a36b LP |
7083 | * MemoryLow= and MemoryMin= gained hierarchy-aware counterparts, |
7084 | DefaultMemoryLow= and DefaultMemoryMin=, which can be used to | |
acdb4b52 CD |
7085 | hierarchically set default memory protection values for a particular |
7086 | subtree of the unit hierarchy. | |
7087 | ||
22bf131b CD |
7088 | * Memory protection directives can now take a value of zero, allowing |
7089 | explicit opting out of a default value propagated by an ancestor. | |
7090 | ||
87cf1f8f LP |
7091 | * systemd now defaults to the "unified" cgroup hierarchy setup during |
7092 | build-time, i.e. -Ddefault-hierarchy=unified is now the build-time | |
7093 | default. Previously, -Ddefault-hierarchy=hybrid was the default. This | |
7094 | change reflects the fact that cgroupsv2 support has matured | |
7095 | substantially in both systemd and in the kernel, and is clearly the | |
7096 | way forward. Downstream production distributions might want to | |
7097 | continue to use -Ddefault-hierarchy=hybrid (or even =legacy) for | |
7098 | their builds as unfortunately the popular container managers have not | |
7099 | caught up with the kernel API changes. | |
7100 | ||
51b568f7 ZJS |
7101 | * Man pages are not built by default anymore (html pages were already |
7102 | disabled by default), to make development builds quicker. When | |
7103 | building systemd for a full installation with documentation, meson | |
4860f5c2 ZJS |
7104 | should be called with -Dman=true and/or -Dhtml=true as appropriate. |
7105 | The default was changed based on the assumption that quick one-off or | |
7106 | repeated development builds are much more common than full optimized | |
7107 | builds for installation, and people need to pass various other | |
7108 | options to when doing "proper" builds anyway, so the gain from making | |
7109 | development builds quicker is bigger than the one time disruption for | |
7110 | packagers. | |
51b568f7 ZJS |
7111 | |
7112 | Two scripts are created in the *build* directory to generate and | |
7113 | preview man and html pages on demand, e.g.: | |
7114 | ||
7115 | build/man/man systemctl | |
7116 | build/man/html systemd.index | |
7117 | ||
e110599b | 7118 | * libidn2 is used by default if both libidn2 and libidn are installed. |
4860f5c2 | 7119 | Please use -Dlibidn=true if libidn is preferred. |
e110599b | 7120 | |
2875a36b | 7121 | * The D-Bus "wire format" of the CPUAffinity= attribute is changed on |
9120e2bf ZJS |
7122 | big-endian machines. Before, bytes were written and read in native |
7123 | machine order as exposed by the native libc __cpu_mask interface. | |
7124 | Now, little-endian order is always used (CPUs 0–7 are described by | |
7125 | bits 0–7 in byte 0, CPUs 8–15 are described by byte 1, and so on). | |
7126 | This change fixes D-Bus calls that cross endianness boundary. | |
7127 | ||
2875a36b LP |
7128 | The presentation format used for CPUAffinity= by "systemctl show" and |
7129 | "systemd-analyze dump" is changed to present CPU indices instead of | |
7130 | the raw __cpu_mask bitmask. For example, CPUAffinity=0-1 would be | |
7131 | shown as CPUAffinity=03000000000000000000000000000… (on | |
7132 | little-endian) or CPUAffinity=00000000000000300000000000000… (on | |
7133 | 64-bit big-endian), and is now shown as CPUAffinity=0-1, matching the | |
7134 | input format. The maximum integer that will be printed in the new | |
7135 | format is 8191 (four digits), while the old format always used a very | |
7136 | long number (with the length varying by architecture), so they can be | |
7137 | unambiguously distinguished. | |
9120e2bf | 7138 | |
51b568f7 ZJS |
7139 | * /usr/sbin/halt.local is no longer supported. Implementation in |
7140 | distributions was inconsistent and it seems this functionality was | |
7141 | very rarely used. | |
7142 | ||
7143 | To replace this functionality, users should: | |
7144 | - either define a new unit and make it a dependency of final.target | |
7145 | (systemctl add-wants final.target my-halt-local.service) | |
7146 | - or move the shutdown script to /usr/lib/systemd/system-shutdown/ | |
7147 | and ensure that it accepts "halt", "poweroff", "reboot", and | |
7148 | "kexec" as an argument, see the description in systemd-shutdown(8). | |
7149 | ||
7150 | * When a [Match] section in .link or .network file is empty (contains | |
7151 | no match patterns), a warning will be emitted. Please add any "match | |
9120e2bf | 7152 | all" pattern instead, e.g. OriginalName=* or Name=* in case all |
51b568f7 ZJS |
7153 | interfaces should really be matched. |
7154 | ||
b070c7c0 | 7155 | * A new setting NUMAPolicy= may be used to set process memory |
08b59539 CR |
7156 | allocation policy. This setting can be specified in |
7157 | /etc/systemd/system.conf and hence will set the default policy for | |
7158 | PID1. The default policy can be overridden on a per-service | |
7159 | basis. The related setting NUMAMask= is used to specify NUMA node | |
7160 | mask that should be associated with the selected policy. | |
2875a36b LP |
7161 | |
7162 | * PID 1 will now listen to Out-Of-Memory (OOM) events the kernel | |
eebaa724 | 7163 | generates when processes it manages are reaching their memory limits, |
2875a36b LP |
7164 | and will place their units in a special state, and optionally kill or |
7165 | stop the whole unit. | |
7166 | ||
7167 | * The service manager will now expose bus properties for the IO | |
7168 | resources used by units. This information is also shown in "systemctl | |
7169 | status" now (for services that have IOAccounting=yes set). Moreover, | |
7170 | the IO accounting data is included in the resource log message | |
7171 | generated whenever a unit stops. | |
7172 | ||
201632e3 | 7173 | * Units may now configure an explicit timeout to wait for when killed |
2875a36b | 7174 | with SIGABRT, for example when a service watchdog is hit. Previously, |
201632e3 ZJS |
7175 | the regular TimeoutStopSec= timeout was applied in this case too — |
7176 | now a separate timeout may be set using TimeoutAbortSec=. | |
2875a36b LP |
7177 | |
7178 | * Services may now send a special WATCHDOG=trigger message with | |
7179 | sd_notify() to trigger an immediate "watchdog missed" event, and thus | |
4860f5c2 | 7180 | trigger service termination. This is useful both for testing watchdog |
2875a36b LP |
7181 | handling, but also for defining error paths in services, that shall |
7182 | be handled the same way as watchdog events. | |
7183 | ||
7184 | * There are two new per-unit settings IPIngressFilterPath= and | |
7185 | IPEgressFilterPath= which allow configuration of a BPF program | |
7186 | (usually by specifying a path to a program uploaded to /sys/fs/bpf/) | |
7187 | to apply to the IP packet ingress/egress path of all processes of a | |
7188 | unit. This is useful to allow running systemd services with BPF | |
7189 | programs set up externally. | |
7190 | ||
7191 | * systemctl gained a new "clean" verb for removing the state, cache, | |
7192 | runtime or logs directories of a service while it is terminated. The | |
7193 | new verb may also be used to remove the state maintained on disk for | |
7194 | timer units that have Persistent= configured. | |
7195 | ||
7196 | * During the last phase of shutdown systemd will now automatically | |
7197 | increase the log level configured in the "kernel.printk" sysctl so | |
7198 | that any relevant loggable events happening during late shutdown are | |
7199 | made visible. Previously, loggable events happening so late during | |
7200 | shutdown were generally lost if the "kernel.printk" sysctl was set to | |
7201 | high thresholds, as regular logging daemons are terminated at that | |
7202 | time and thus nothing is written to disk. | |
7203 | ||
7204 | * If processes terminated during the last phase of shutdown do not exit | |
7205 | quickly systemd will now show their names after a short time, to make | |
201632e3 | 7206 | debugging easier. After a longer timeout they are forcibly killed, |
2875a36b LP |
7207 | as before. |
7208 | ||
7209 | * journalctl (and the other tools that display logs) will now highlight | |
7210 | warnings in yellow (previously, both LOG_NOTICE and LOG_WARNING where | |
7211 | shown in bright bold, now only LOG_NOTICE is). Moreover, audit logs | |
7212 | are now shown in blue color, to separate them visually from regular | |
7213 | logs. References to configuration files are now turned into clickable | |
7214 | links on terminals that support that. | |
7215 | ||
7216 | * systemd-journald will now stop logging to /var/log/journal during | |
7217 | shutdown when /var/ is on a separate mount, so that it can be | |
7218 | unmounted safely during shutdown. | |
7219 | ||
7220 | * systemd-resolved gained support for a new 'strict' DNS-over-TLS mode. | |
7221 | ||
29db4c3a LP |
7222 | * systemd-resolved "Cache=" configuration option in resolved.conf has |
7223 | been extended to also accept the 'no-negative' value. Previously, | |
7224 | only a boolean option was allowed (yes/no), having yes as the | |
4860f5c2 ZJS |
7225 | default. If this option is set to 'no-negative', negative answers are |
7226 | not cached while the old cache heuristics are used positive answers. | |
7227 | The default remains unchanged. | |
37d7a7d9 | 7228 | |
2875a36b LP |
7229 | * The predictable naming scheme for network devices now supports |
7230 | generating predictable names for "netdevsim" devices. | |
7231 | ||
4860f5c2 ZJS |
7232 | Moreover, the "en" prefix was dropped from the ID_NET_NAME_ONBOARD |
7233 | udev property. | |
7234 | ||
7235 | Those two changes form a new net.naming-policy-scheme= entry. | |
7236 | Distributions which want to preserve naming stability may want to set | |
7237 | the -Ddefault-net-naming-scheme= configuration option. | |
7238 | ||
2875a36b LP |
7239 | * systemd-networkd now supports MACsec, nlmon, IPVTAP and Xfrm |
7240 | interfaces natively. | |
7241 | ||
7242 | * systemd-networkd's bridge FDB support now allows configuration of a | |
7243 | destination address for each entry (Destination=), as well as the | |
7244 | VXLAN VNI (VNI=), as well as an option to declare what an entry is | |
7245 | associated with (AssociatedWith=). | |
7246 | ||
7247 | * systemd-networkd's DHCPv4 support now understands a new MaxAttempts= | |
33db1b90 | 7248 | option for configuring the maximum number of DHCP lease requests. It |
6b000af4 | 7249 | also learnt a new BlackList= option for deny-listing DHCP servers (a |
08b59539 CR |
7250 | similar setting has also been added to the IPv6 RA client), as well |
7251 | as a SendRelease= option for configuring whether to send a DHCP | |
7252 | RELEASE message when terminating. | |
2875a36b LP |
7253 | |
7254 | * systemd-networkd's DHCPv4 and DHCPv6 stacks can now be configured | |
29db4c3a LP |
7255 | separately in the [DHCPv4] and [DHCPv6] sections. |
7256 | ||
7257 | * systemd-networkd's DHCP support will now optionally create an | |
7258 | implicit host route to the DNS server specified in the DHCP lease, in | |
7259 | addition to the routes listed explicitly in the lease. This should | |
7260 | ensure that in multi-homed systems DNS traffic leaves the systems on | |
7261 | the interface that acquired the DNS server information even if other | |
7262 | routes such as default routes exist. This behaviour may be turned on | |
7263 | with the new RoutesToDNS= option. | |
2875a36b LP |
7264 | |
7265 | * systemd-networkd's VXLAN support gained a new option | |
eebaa724 | 7266 | GenericProtocolExtension= for enabling VXLAN Generic Protocol |
2875a36b LP |
7267 | Extension support, as well as IPDoNotFragment= for setting the IP |
7268 | "Don't fragment" bit on outgoing packets. A similar option has been | |
7269 | added to the GENEVE support. | |
7270 | ||
7271 | * In systemd-networkd's [Route] section you may now configure | |
7272 | FastOpenNoCookie= for configuring per-route TCP fast-open support, as | |
7273 | well as TTLPropagate= for configuring Label Switched Path (LSP) TTL | |
7274 | propagation. The Type= setting now supports local, broadcast, | |
7275 | anycast, multicast, any, xresolve routes, too. | |
7276 | ||
7277 | * systemd-networkd's [Network] section learnt a new option | |
7278 | DefaultRouteOnDevice= for automatically configuring a default route | |
7279 | onto the network device. | |
7280 | ||
7281 | * systemd-networkd's bridging support gained two new options ProxyARP= | |
7282 | and ProxyARPWifi= for configuring proxy ARP behaviour as well as | |
9f0d45e4 YW |
7283 | MulticastRouter= for configuring multicast routing behaviour. A new |
7284 | option MulticastIGMPVersion= may be used to change bridge's multicast | |
7285 | Internet Group Management Protocol (IGMP) version. | |
2875a36b LP |
7286 | |
7287 | * systemd-networkd's FooOverUDP support gained the ability to configure | |
7288 | local and peer IP addresses via Local= and Peer=. A new option | |
7289 | PeerPort= may be used to configure the peer's IP port. | |
7290 | ||
7291 | * systemd-networkd's TUN support gained a new setting VnetHeader= for | |
7292 | tweaking Generic Segment Offload support. | |
7293 | ||
70fcda85 ZJS |
7294 | * The address family for policy rules may be specified using the new |
7295 | Family= option in the [RoutingPolicyRule] section. | |
7296 | ||
2875a36b LP |
7297 | * networkctl gained a new "delete" command for removing virtual network |
7298 | devices, as well as a new "--stats" switch for showing device | |
7299 | statistics. | |
7300 | ||
13cb62af YW |
7301 | * networkd.conf gained a new setting SpeedMeter= and |
7302 | SpeedMeterIntervalSec=, to measure bitrate of network interfaces. The | |
7303 | measured speed may be shown by 'networkctl status'. | |
7304 | ||
4860f5c2 ZJS |
7305 | * "networkctl status" now displays MTU and queue lengths, and more |
7306 | detailed information about VXLAN and bridge devices. | |
7307 | ||
2875a36b LP |
7308 | * systemd-networkd's .network and .link files gained a new Property= |
7309 | setting in the [Match] section, to match against devices with | |
7310 | specific udev properties. | |
7311 | ||
7312 | * systemd-networkd's tunnel support gained a new option | |
7313 | AssignToLoopback= for selecting whether to use the loopback device | |
7314 | "lo" as underlying device. | |
7315 | ||
70183735 | 7316 | * systemd-networkd's MACAddress= setting in the [Neighbor] section has |
2875a36b LP |
7317 | been renamed to LinkLayerAddress=, and it now allows configuration of |
7318 | IP addresses, too. | |
7319 | ||
efb536d0 ZJS |
7320 | * systemd-networkd's handling of the kernel's disable_ipv6 sysctl is |
7321 | simplified: systemd-networkd will disable the sysctl (enable IPv6) if | |
7322 | IPv6 configuration (static or DHCPv6) was found for a given | |
7323 | interface. It will not touch the sysctl otherwise. | |
7324 | ||
7325 | * The order of entries is $PATH used by the user manager instance was | |
7326 | changed to put bin/ entries before the corresponding sbin/ entries. | |
7327 | It is recommended to not rely on this order, and only ever have one | |
7328 | binary with a given name in the system paths under /usr. | |
7329 | ||
29db4c3a LP |
7330 | * A new tool systemd-network-generator has been added that may generate |
7331 | .network, .netdev and .link files from IP configuration specified on | |
4860f5c2 | 7332 | the kernel command line in the format used by Dracut. |
29db4c3a | 7333 | |
2875a36b LP |
7334 | * The CriticalConnection= setting in .network files is now deprecated, |
7335 | and replaced by a new KeepConfiguration= setting which allows more | |
7336 | detailed configuration of the IP configuration to keep in place. | |
7337 | ||
4860f5c2 ZJS |
7338 | * systemd-analyze gained a few new verbs: |
7339 | ||
7340 | - "systemd-analyze timestamp" parses and converts timestamps. This is | |
7341 | similar to the existing "systemd-analyze calendar" command which | |
7342 | does the same for recurring calendar events. | |
7343 | ||
7344 | - "systemd-analyze timespan" parses and converts timespans (i.e. | |
7345 | durations as opposed to points in time). | |
7346 | ||
7347 | - "systemd-analyze condition" will parse and test ConditionXYZ= | |
7348 | expressions. | |
7349 | ||
7350 | - "systemd-analyze exit-status" will parse and convert exit status | |
7351 | codes to their names and back. | |
7352 | ||
7353 | - "systemd-analyze unit-files" will print a list of all unit | |
7354 | file paths and unit aliases. | |
7355 | ||
7356 | * SuccessExitStatus=, RestartPreventExitStatus=, and | |
7357 | RestartForceExitStatus= now accept exit status names (e.g. "DATAERR" | |
7358 | is equivalent to "65"). Those exit status name mappings may be | |
cc560ac0 | 7359 | displayed with the systemd-analyze exit-status verb describe above. |
2875a36b LP |
7360 | |
7361 | * systemd-logind now exposes a per-session SetBrightness() bus call, | |
7362 | which may be used to securely change the brightness of a kernel | |
7363 | brightness device, if it belongs to the session's seat. By using this | |
7364 | call unprivileged clients can make changes to "backlight" and "leds" | |
4860f5c2 ZJS |
7365 | devices securely with strict requirements on session membership. |
7366 | Desktop environments may use this to generically make brightness | |
7367 | changes to such devices without shipping private SUID binaries or | |
7368 | udev rules for that purpose. | |
2875a36b LP |
7369 | |
7370 | * "udevadm info" gained a --wait-for-initialization switch to wait for | |
7371 | a device to be initialized. | |
7372 | ||
7373 | * systemd-hibernate-resume-generator will now look for resumeflags= on | |
7374 | the kernel command line, which is similar to rootflags= and may be | |
4860f5c2 | 7375 | used to configure device timeout for the hibernation device. |
2875a36b LP |
7376 | |
7377 | * sd-event learnt a new API call sd_event_source_disable_unref() for | |
7378 | disabling and unref'ing an event source in a single function. A | |
7379 | related call sd_event_source_disable_unrefp() has been added for use | |
4860f5c2 | 7380 | with gcc's cleanup extension. |
2875a36b LP |
7381 | |
7382 | * The sd-id128.h public API gained a new definition | |
da890466 | 7383 | SD_ID128_UUID_FORMAT_STR for formatting a 128-bit ID in UUID format |
2875a36b LP |
7384 | with printf(). |
7385 | ||
7386 | * "busctl introspect" gained a new switch --xml-interface for dumping | |
7387 | XML introspection data unmodified. | |
7388 | ||
7389 | * PID 1 may now show the unit name instead of the unit description | |
7390 | string in its status output during boot. This may be configured in | |
7391 | the StatusUnitFormat= setting in /etc/systemd/system.conf or the | |
7392 | kernel command line option systemd.status_unit_format=. | |
7393 | ||
907ddcd3 | 7394 | * PID 1 now understands a new option KExecWatchdogSec= in |
4860f5c2 ZJS |
7395 | /etc/systemd/system.conf to set a watchdog timeout for kexec reboots. |
7396 | Previously watchdog functionality was only available for regular | |
7397 | reboots. The new setting defaults to off, because we don't know in | |
7398 | the general case if the watchdog will be reset after kexec (some | |
7399 | drivers do reset it, but not all), and the new userspace might not be | |
7400 | configured to handle the watchdog. | |
7401 | ||
7402 | Moreover, the old ShutdownWatchdogSec= setting has been renamed to | |
7403 | RebootWatchdogSec= to more clearly communicate what it is about. The | |
7404 | old name is still accepted for compatibility. | |
907ddcd3 | 7405 | |
2875a36b | 7406 | * The systemd.debug_shell kernel command line option now optionally |
4860f5c2 ZJS |
7407 | takes a tty name to spawn the debug shell on, which allows a |
7408 | different tty to be selected than the built-in default. | |
2875a36b | 7409 | |
a4d5848a AZ |
7410 | * Service units gained a new ExecCondition= setting which will run |
7411 | before ExecStartPre= and either continue execution of the unit (for | |
7412 | clean exit codes), stop execution without marking the unit failed | |
7413 | (for exit codes 1 through 254), or stop execution and fail the unit | |
4860f5c2 | 7414 | (for exit code 255 or abnormal termination). |
a4d5848a | 7415 | |
29db4c3a | 7416 | * A new service systemd-pstore.service has been added that pulls data |
08b59539 | 7417 | from /sys/fs/pstore/ and saves it to /var/lib/pstore for later |
29db4c3a LP |
7418 | review. |
7419 | ||
7420 | * timedatectl gained new verbs for configuring per-interface NTP | |
7421 | service configuration for systemd-timesyncd. | |
7422 | ||
7423 | * "localectl list-locales" won't list non-UTF-8 locales anymore. It's | |
a18a3aac | 7424 | 2019. (You can set non-UTF-8 locales though, if you know their name.) |
29db4c3a | 7425 | |
b64c47c0 LP |
7426 | * If variable assignments in sysctl.d/ files are prefixed with "-" any |
7427 | failures to apply them are now ignored. | |
7428 | ||
a18a3aac LP |
7429 | * systemd-random-seed.service now optionally credits entropy when |
7430 | applying the seed to the system. Set $SYSTEMD_RANDOM_SEED_CREDIT to | |
7431 | true for the service to enable this behaviour, but please consult the | |
7432 | documentation first, since this comes with a couple of caveats. | |
7433 | ||
4860f5c2 ZJS |
7434 | * systemd-random-seed.service is now a synchronization point for full |
7435 | initialization of the kernel's entropy pool. Services that require | |
7436 | /dev/urandom to be correctly initialized should be ordered after this | |
7437 | service. | |
a18a3aac LP |
7438 | |
7439 | * The systemd-boot boot loader has been updated to optionally maintain | |
7440 | a random seed file in the EFI System Partition (ESP). During the boot | |
7441 | phase, this random seed is read and updated with a new seed | |
4860f5c2 | 7442 | cryptographically derived from it. Another derived seed is passed to |
a18a3aac LP |
7443 | the OS. The latter seed is then credited to the kernel's entropy pool |
7444 | very early during userspace initialization (from PID 1). This allows | |
7445 | systems to boot up with a fully initialized kernel entropy pool from | |
7446 | earliest boot on, and thus entirely removes all entropy pool | |
7447 | initialization delays from systems using systemd-boot. Special care | |
7448 | is taken to ensure different seeds are derived on system images | |
4860f5c2 ZJS |
7449 | replicated to multiple systems. "bootctl status" will show whether |
7450 | a seed was received from the boot loader. | |
7451 | ||
7452 | * bootctl gained two new verbs: | |
7453 | ||
7454 | - "bootctl random-seed" will generate the file in ESP and an EFI | |
7455 | variable to allow a random seed to be passed to the OS as described | |
7456 | above. | |
7457 | ||
7458 | - "bootctl is-installed" checks whether systemd-boot is currently | |
7459 | installed. | |
a18a3aac | 7460 | |
4860f5c2 ZJS |
7461 | * bootctl will warn if it detects that boot entries are misconfigured |
7462 | (for example if the kernel image was removed without purging the | |
7463 | bootloader entry). | |
a18a3aac LP |
7464 | |
7465 | * A new document has been added describing systemd's use and support | |
7466 | for the kernel's entropy pool subsystem: | |
7467 | ||
7468 | https://systemd.io/RANDOM_SEEDS | |
7469 | ||
48fd50f7 LP |
7470 | * When the system is hibernated the swap device to write the |
7471 | hibernation image to is now automatically picked from all available | |
7472 | swap devices, preferring the swap device with the highest configured | |
7473 | priority over all others, and picking the device with the most free | |
7474 | space if there are multiple devices with the highest priority. | |
7475 | ||
7476 | * /etc/crypttab support has learnt a new keyfile-timeout= per-device | |
86b52a39 | 7477 | option that permits selecting the timeout how long to wait for a |
48fd50f7 LP |
7478 | device with an encryption key before asking for the password. |
7479 | ||
2dbc45ae KK |
7480 | * IOWeight= has learnt to properly set the IO weight when using the |
7481 | BFQ scheduler officially found in kernels 5.0+. | |
7482 | ||
6d8cf864 ZJS |
7483 | * A new mailing list has been created for reporting of security issues: |
7484 | systemd-security@redhat.com. For mode details, see | |
7485 | https://systemd.io/CONTRIBUTING#security-vulnerability-reports. | |
7486 | ||
f21e2ecb ZJS |
7487 | Contributions from: Aaron Barany, Adrian Bunk, Alan Jenkins, Albrecht |
7488 | Lohofener, Andrej Valek, Anita Zhang, Arian van Putten, Balint Reczey, | |
7489 | Bastien Nocera, Ben Boeckel, Benjamin Robin, camoz, Chen Qi, Chris | |
03985d06 ZJS |
7490 | Chiu, Chris Down, Christian Göttsche, Christian Kellner, Clinton Roy, |
7491 | Connor Reeder, Daniel Black, Daniel Lublin, Daniele Medri, Dan | |
7492 | Streetman, Dave Reisner, Dave Ross, David Art, David Tardon, Debarshi | |
7493 | Ray, Dimitri John Ledkov, Dominick Grift, Donald Buczek, Douglas | |
7494 | Christman, Eric DeVolder, EtherGraf, Evgeny Vereshchagin, Feldwor, | |
7495 | Felix Riemann, Florian Dollinger, Francesco Pennica, Franck Bui, | |
7496 | Frantisek Sumsal, Franz Pletz, frederik, Hans de Goede, Iago López | |
7497 | Galeiras, Insun Pyo, Ivan Shapovalov, Iwan Timmer, Jack, Jakob | |
7498 | Unterwurzacher, Jan Chren, Jan Klötzke, Jan Losinski, Jan Pokorný, Jan | |
7499 | Synacek, Jan-Michael Brummer, Jeka Pats, Jeremy Soller, Jérémy Rosen, | |
7500 | Jiri Pirko, Joe Lin, Joerg Behrmann, Joe Richey, Jóhann B. Guðmundsson, | |
7501 | Johannes Christ, Johannes Schmitz, Jonathan Rouleau, Jorge Niedbalski, | |
7502 | Jörg Thalheim, Kai Krakow, Kai Lüke, Karel Zak, Kashyap Chamarthy, | |
f21e2ecb ZJS |
7503 | Krayushkin Konstantin, Lennart Poettering, Lubomir Rintel, Luca |
7504 | Boccassi, Luís Ferreira, Marc-André Lureau, Markus Felten, Martin Pitt, | |
7505 | Matthew Leeds, Mattias Jernberg, Michael Biebl, Michael Olbrich, | |
7506 | Michael Prokop, Michael Stapelberg, Michael Zhivich, Michal Koutný, | |
7507 | Michal Sekletar, Mike Gilbert, Milan Broz, Miroslav Lichvar, mpe85, | |
7508 | Mr-Foo, Network Silence, Oliver Harley, pan93412, Paul Menzel, pEJipE, | |
03985d06 ZJS |
7509 | Peter A. Bigot, Philip Withnall, Piotr Drąg, Rafael Fontenelle, Robert |
7510 | Scheck, Roberto Santalla, Ronan Pigott, root, RussianNeuroMancer, | |
7511 | Sebastian Jennen, shinygold, Shreyas Behera, Simon Schricker, Susant | |
7512 | Sahani, Thadeu Lima de Souza Cascardo, Theo Ouzhinski, Thiebaud | |
7513 | Weksteen, Thomas Haller, Thomas Weißschuh, Tomas Mraz, Tommi Rantala, | |
7514 | Topi Miettinen, VD-Lycos, ven, Vladimir Yerilov, Wieland Hoffmann, | |
7515 | William A. Kennington III, William Wold, Xi Ruoyao, Yuri Chornoivan, | |
7516 | Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek, Zhang Xianwei | |
29db4c3a | 7517 | |
efb536d0 | 7518 | – Camerino, 2019-09-03 |
6af90583 | 7519 | |
d822bd4e | 7520 | CHANGES WITH 242: |
9b89e602 ZJS |
7521 | |
7522 | * In .link files, MACAddressPolicy=persistent (the default) is changed | |
7523 | to cover more devices. For devices like bridges, tun, tap, bond, and | |
7524 | similar interfaces that do not have other identifying information, | |
7525 | the interface name is used as the basis for persistent seed for MAC | |
7526 | and IPv4LL addresses. The way that devices that were handled | |
fd9baae8 ZJS |
7527 | previously is not changed, and this change is about covering more |
7528 | devices then previously by the "persistent" policy. | |
9b89e602 ZJS |
7529 | |
7530 | MACAddressPolicy=random may be used to force randomized MACs and | |
7531 | IPv4LL addresses for a device if desired. | |
7532 | ||
7533 | Hint: the log output from udev (at debug level) was enhanced to | |
7534 | clarify what policy is followed and which attributes are used. | |
7535 | `SYSTEMD_LOG_LEVEL=debug udevadm test-builtin net_setup_link /sys/class/net/<name>` | |
7536 | may be used to view this. | |
7537 | ||
c111cd98 YW |
7538 | Hint: if a bridge interface is created without any slaves, and gains |
7539 | a slave later, then now the bridge does not inherit slave's MAC. | |
7540 | To inherit slave's MAC, for example, create the following file: | |
7541 | ``` | |
7542 | # /etc/systemd/network/98-bridge-inherit-mac.link | |
7543 | [Match] | |
7544 | Type=bridge | |
7545 | ||
7546 | [Link] | |
7547 | MACAddressPolicy=none | |
7548 | ``` | |
7549 | ||
9b89e602 ZJS |
7550 | * The .device units generated by systemd-fstab-generator and other |
7551 | generators do not automatically pull in the corresponding .mount unit | |
7552 | as a Wants= dependency. This means that simply plugging in the device | |
7553 | will not cause the mount unit to be started automatically. But please | |
5787c509 LP |
7554 | note that the mount unit may be started for other reasons, in |
7555 | particular if it is part of local-fs.target, and any unit which | |
7556 | (transitively) depends on local-fs.target is started. | |
9b89e602 | 7557 | |
5787c509 LP |
7558 | * networkctl list/status/lldp now accept globbing wildcards for network |
7559 | interface names to match against all existing interfaces. | |
7560 | ||
7561 | * The $PIDFILE environment variable is set to point the absolute path | |
7562 | configured with PIDFile= for processes of that service. | |
9b89e602 ZJS |
7563 | |
7564 | * The fallback DNS server list was augmented with Cloudflare public DNS | |
7565 | servers. Use `-Ddns-servers=` to set a different fallback. | |
7566 | ||
7567 | * A new special target usb-gadget.target will be started automatically | |
7568 | when a USB Device Controller is detected (which means that the system | |
7569 | is a USB peripheral). | |
7570 | ||
5787c509 LP |
7571 | * A new unit setting CPUQuotaPeriodSec= assigns the time period |
7572 | relatively to which the CPU time quota specified by CPUQuota= is | |
7573 | measured. | |
9b89e602 | 7574 | |
5787c509 | 7575 | * A new unit setting ProtectHostname= may be used to prevent services |
9b89e602 ZJS |
7576 | from modifying hostname information (even if they otherwise would |
7577 | have privileges to do so). | |
7578 | ||
5787c509 | 7579 | * A new unit setting NetworkNamespacePath= may be used to specify a |
9b89e602 ZJS |
7580 | namespace for service or socket units through a path referring to a |
7581 | Linux network namespace pseudo-file. | |
7582 | ||
5787c509 LP |
7583 | * The PrivateNetwork= setting and JoinsNamespaceOf= dependencies now |
7584 | have an effect on .socket units: when used the listening socket is | |
7585 | created within the configured network namespace instead of the host | |
7586 | namespace. | |
7587 | ||
7588 | * ExecStart= command lines in unit files may now be prefixed with ':' | |
7589 | in which case environment variable substitution is | |
7590 | disabled. (Supported for the other ExecXYZ= settings, too.) | |
7591 | ||
2eb466fc LP |
7592 | * .timer units gained two new boolean settings OnClockChange= and |
7593 | OnTimezoneChange= which may be used to also trigger a unit when the | |
7594 | system clock is changed or the local timezone is | |
7595 | modified. systemd-run has been updated to make these options easily | |
7596 | accessible from the command line for transient timers. | |
7597 | ||
7598 | * Two new conditions for units have been added: ConditionMemory= may be | |
7599 | used to conditionalize a unit based on installed system | |
7600 | RAM. ConditionCPUs= may be used to conditionalize a unit based on | |
39e445c9 | 7601 | installed CPU cores. |
2eb466fc LP |
7602 | |
7603 | * The @default system call filter group understood by SystemCallFilter= | |
7604 | has been updated to include the new rseq() system call introduced in | |
7605 | kernel 4.15. | |
7606 | ||
ab80eca1 ZJS |
7607 | * A new time-set.target has been added that indicates that the system |
7608 | time has been set from a local source (possibly imprecise). The | |
7609 | existing time-sync.target is stronger and indicates that the time has | |
7610 | been synchronized with a precise external source. Services where | |
7611 | approximate time is sufficient should use the new target. | |
7612 | ||
2eb466fc LP |
7613 | * "systemctl start" (and related commands) learnt a new |
7614 | --show-transaction option. If specified brief information about all | |
7615 | jobs queued because of the requested operation is shown. | |
7616 | ||
5787c509 LP |
7617 | * systemd-networkd recognizes a new operation state 'enslaved', used |
7618 | (instead of 'degraded' or 'carrier') for interfaces which form a | |
7619 | bridge, bond, or similar, and an new 'degraded-carrier' operational | |
7620 | state used for the bond or bridge master interface when one of the | |
7621 | enslaved devices is not operational. | |
9b89e602 | 7622 | |
5787c509 LP |
7623 | * .network files learnt the new IgnoreCarrierLoss= option for leaving |
7624 | networks configured even if the carrier is lost. | |
7625 | ||
7626 | * The RequiredForOnline= setting in .network files may now specify a | |
9b89e602 | 7627 | minimum operational state required for the interface to be considered |
5787c509 LP |
7628 | "online" by systemd-networkd-wait-online. Related to this |
7629 | systemd-networkd-wait-online gained a new option --operational-state= | |
7630 | to configure the same, and its --interface= option was updated to | |
7631 | optionally also take an operational state specific for an interface. | |
9b89e602 | 7632 | |
2eb466fc LP |
7633 | * systemd-networkd-wait-online gained a new setting --any for waiting |
7634 | for only one of the requested interfaces instead of all of them. | |
7635 | ||
9b89e602 ZJS |
7636 | * systemd-networkd now implements L2TP tunnels. |
7637 | ||
5787c509 LP |
7638 | * Two new .network settings UseAutonomousPrefix= and UseOnLinkPrefix= |
7639 | may be used to cause autonomous and onlink prefixes received in IPv6 | |
9b89e602 ZJS |
7640 | Router Advertisements to be ignored. |
7641 | ||
5787c509 LP |
7642 | * New MulticastFlood=, NeighborSuppression=, and Learning= .network |
7643 | file settings may be used to tweak bridge behaviour. | |
7644 | ||
7645 | * The new TripleSampling= option in .network files may be used to | |
7646 | configure CAN triple sampling. | |
9b89e602 | 7647 | |
78bb2866 YW |
7648 | * A new .netdev settings PrivateKeyFile= and PresharedKeyFile= may be |
7649 | used to point to private or preshared key for a WireGuard interface. | |
9b89e602 | 7650 | |
5787c509 LP |
7651 | * /etc/crypttab now supports the same-cpu-crypt and |
7652 | submit-from-crypt-cpus options to tweak encryption work scheduling | |
7653 | details. | |
9b89e602 ZJS |
7654 | |
7655 | * systemd-tmpfiles will now take a BSD file lock before operating on a | |
7656 | contents of directory. This may be used to temporarily exclude | |
7657 | directories from aging by taking the same lock (useful for example | |
7658 | when extracting a tarball into /tmp or /var/tmp as a privileged user, | |
7659 | which might create files with really old timestamps, which | |
5787c509 LP |
7660 | nevertheless should not be deleted). For further details, see: |
7661 | ||
7662 | https://systemd.io/TEMPORARY_DIRECTORIES | |
9b89e602 | 7663 | |
70d8401d LP |
7664 | * systemd-tmpfiles' h line type gained support for the |
7665 | FS_PROJINHERIT_FL ('P') file attribute (introduced in kernel 4.5), | |
7666 | controlling project quota inheritance. | |
7667 | ||
9b89e602 ZJS |
7668 | * sd-boot and bootctl now implement support for an Extended Boot Loader |
7669 | (XBOOTLDR) partition, that is intended to be mounted to /boot, in | |
7670 | addition to the ESP partition mounted to /efi or /boot/efi. | |
7671 | Configuration file fragments, kernels, initrds and other EFI images | |
7672 | to boot will be loaded from both the ESP and XBOOTLDR partitions. | |
7673 | The XBOOTLDR partition was previously described by the Boot Loader | |
5787c509 LP |
7674 | Specification, but implementation was missing in sd-boot. Support for |
7675 | this concept allows using the sd-boot boot loader in more | |
7676 | conservative scenarios where the boot loader itself is placed in the | |
7677 | ESP but the kernels to boot (and their metadata) in a separate | |
7678 | partition. | |
9b89e602 | 7679 | |
5787c509 LP |
7680 | * A system may now be booted with systemd.volatile=overlay on the |
7681 | kernel command line, which causes the root file system to be set up | |
7682 | an overlayfs mount combining the root-only root directory with a | |
7683 | writable tmpfs. In this setup, the underlying root device is not | |
7684 | modified, and any changes are lost at reboot. | |
9b89e602 | 7685 | |
5787c509 LP |
7686 | * Similar, systemd-nspawn can now boot containers with a volatile |
7687 | overlayfs root with the new --volatile=overlay switch. | |
9b89e602 ZJS |
7688 | |
7689 | * systemd-nspawn can now consume OCI runtime bundles using a new | |
7690 | --oci-bundle= option. This implementation is fully usable, with most | |
7691 | features in the specification implemented, but since this a lot of | |
7692 | new code and functionality, this feature should most likely not | |
7693 | be used in production yet. | |
7694 | ||
5787c509 LP |
7695 | * systemd-nspawn now supports various options described by the OCI |
7696 | runtime specification on the command-line and in .nspawn files: | |
9b89e602 | 7697 | --inaccessible=/Inaccessible= may be used to mask parts of the file |
5787c509 | 7698 | system tree, --console=/--pipe may be used to configure how standard |
9b89e602 ZJS |
7699 | input, output, and error are set up. |
7700 | ||
7701 | * busctl learned the `emit` verb to generate D-Bus signals. | |
7702 | ||
7703 | * systemd-analyze cat-config may be used to gather and display | |
7704 | configuration spread over multiple files, for example system and user | |
7705 | presets, tmpfiles.d, sysusers.d, udev rules, etc. | |
7706 | ||
5787c509 LP |
7707 | * systemd-analyze calendar now takes an optional new parameter |
7708 | --iterations= which may be used to show a maximum number of iterations | |
7709 | the specified expression will elapse next. | |
7710 | ||
7711 | * The sd-bus C API gained support for naming method parameters in the | |
7712 | introspection data. | |
7713 | ||
7714 | * systemd-logind gained D-Bus APIs to specify the "reboot parameter" | |
7715 | the reboot() system call expects. | |
7716 | ||
7717 | * journalctl learnt a new --cursor-file= option that points to a file | |
9b89e602 ZJS |
7718 | from which a cursor should be loaded in the beginning and to which |
7719 | the updated cursor should be stored at the end. | |
7720 | ||
7721 | * ACRN hypervisor and Windows Subsystem for Linux (WSL) are now | |
7722 | detected by systemd-detect-virt (and may also be used in | |
7723 | ConditionVirtualization=). | |
7724 | ||
7725 | * The behaviour of systemd-logind may now be modified with environment | |
7726 | variables $SYSTEMD_REBOOT_TO_FIRMWARE_SETUP, | |
7727 | $SYSTEMD_REBOOT_TO_BOOT_LOADER_MENU, and | |
7728 | $SYSTEMD_REBOOT_TO_BOOT_LOADER_ENTRY. They cause logind to either | |
7729 | skip the relevant operation completely (when set to false), or to | |
7730 | create a flag file in /run/systemd (when set to true), instead of | |
7731 | actually commencing the real operation when requested. The presence | |
7732 | of /run/systemd/reboot-to-firmware-setup, | |
7733 | /run/systemd/reboot-to-boot-loader-menu, and | |
7734 | /run/systemd/reboot-to-boot-loader-entry, may be used by alternative | |
7735 | boot loader implementations to replace some steps logind performs | |
7736 | during reboot with their own operations. | |
7737 | ||
7738 | * systemctl can be used to request a reboot into the boot loader menu | |
5787c509 LP |
7739 | or a specific boot loader entry with the new --boot-load-menu= and |
7740 | --boot-loader-entry= options to a reboot command. (This requires a | |
7741 | boot loader that supports this, for example sd-boot.) | |
9b89e602 ZJS |
7742 | |
7743 | * kernel-install will no longer unconditionally create the output | |
7744 | directory (e.g. /efi/<machine-id>/<kernel-version>) for boot loader | |
7745 | snippets, but will do only if the machine-specific parent directory | |
7746 | (i.e. /efi/<machine-id>/) already exists. bootctl has been modified | |
7747 | to create this parent directory during sd-boot installation. | |
7748 | ||
7749 | This makes it easier to use kernel-install with plugins which support | |
7750 | a different layout of the bootloader partitions (for example grub2). | |
7751 | ||
a3134241 | 7752 | * During package installation (with `ninja install`), we would create |
1fa3ba90 PM |
7753 | symlinks for getty@tty1.service, systemd-networkd.service, |
7754 | systemd-networkd.socket, systemd-resolved.service, | |
7755 | remote-cryptsetup.target, remote-fs.target, | |
a3134241 ZJS |
7756 | systemd-networkd-wait-online.service, and systemd-timesyncd.service |
7757 | in /etc, as if `systemctl enable` was called for those units, to make | |
7758 | the system usable immediately after installation. Now this is not | |
7759 | done anymore, and instead calling `systemctl preset-all` is | |
7760 | recommended after the first installation of systemd. | |
7761 | ||
bf65b7e0 LP |
7762 | * A new boolean sandboxing option RestrictSUIDSGID= has been added that |
7763 | is built on seccomp. When turned on creation of SUID/SGID files is | |
7764 | prohibited. | |
7765 | ||
7766 | * The NoNewPrivileges= and the new RestrictSUIDSGID= options are now | |
7767 | implied if DynamicUser= is turned on for a service. This hardens | |
7768 | these services, so that they neither can benefit from nor create | |
7769 | SUID/SGID executables. This is a minor compatibility breakage, given | |
7770 | that when DynamicUser= was first introduced SUID/SGID behaviour was | |
7771 | unaffected. However, the security benefit of these two options is | |
7772 | substantial, and the setting is still relatively new, hence we opted | |
7773 | to make it mandatory for services with dynamic users. | |
7774 | ||
5b2fc74f LP |
7775 | Contributions from: Adam Jackson, Alexander Tsoy, Andrey Yashkin, |
7776 | Andrzej Pietrasiewicz, Anita Zhang, Balint Reczey, Beniamino Galvani, | |
7777 | Ben Iofel, Benjamin Berg, Benjamin Dahlhoff, Chris, Chris Morin, | |
7778 | Christopher Wong, Claudius Ellsel, Clemens Gruber, dana, Daniel Black, | |
7779 | Davide Cavalca, David Michael, David Rheinsberg, emersion, Evgeny | |
7780 | Vereshchagin, Filipe Brandenburger, Franck Bui, Frantisek Sumsal, | |
7781 | Giacinto Cifelli, Hans de Goede, Hugo Kindel, Ignat Korchagin, Insun | |
c3287a42 LP |
7782 | Pyo, Jan Engelhardt, Jonas Dorel, Jonathan Lebon, Jonathon Kowalski, |
7783 | Jörg Sommer, Jörg Thalheim, Jussi Pakkanen, Kai-Heng Feng, Lennart | |
7784 | Poettering, Lubomir Rintel, Luís Ferreira, Martin Pitt, Matthias | |
7785 | Klumpp, Michael Biebl, Michael Niewöhner, Michael Olbrich, Michal | |
ab80eca1 ZJS |
7786 | Sekletar, Mike Lothian, Paul Menzel, Piotr Drąg, Riccardo Schirone, |
7787 | Robin Elvedi, Roman Kulikov, Ronald Tschalär, Ross Burton, Ryan | |
1e5d2d65 ZJS |
7788 | Gonzalez, Sebastian Krzyszkowiak, Stephane Chazelas, StKob, Susant |
7789 | Sahani, Sylvain Plantefève, Szabolcs Fruhwald, Taro Yamada, Theo | |
7790 | Ouzhinski, Thomas Haller, Tobias Jungel, Tom Yan, Tony Asleson, Topi | |
7791 | Miettinen, unixsysadmin, Van Laser, Vesa Jääskeläinen, Yu, Li-Yu, | |
7792 | Yu Watanabe, Zbigniew Jędrzejewski-Szmek | |
c3287a42 | 7793 | |
1e5d2d65 | 7794 | — Warsaw, 2019-04-11 |
bf65b7e0 | 7795 | |
d0f71749 | 7796 | CHANGES WITH 241: |
b4ff3dbb ZJS |
7797 | |
7798 | * The default locale can now be configured at compile time. Otherwise, | |
7799 | a suitable default will be selected automatically (one of C.UTF-8, | |
7800 | en_US.UTF-8, and C). | |
7801 | ||
7802 | * The version string shown by systemd and other tools now includes the | |
7803 | git commit hash when built from git. An override may be specified | |
7804 | during compilation, which is intended to be used by distributions to | |
7805 | include the package release information. | |
7806 | ||
7807 | * systemd-cat can now filter standard input and standard error streams | |
7808 | for different syslog priorities using the new --stderr-priority= | |
7809 | option. | |
7810 | ||
7811 | * systemd-journald and systemd-journal-remote reject entries which | |
7812 | contain too many fields (CVE-2018-16865) and set limits on the | |
7813 | process' command line length (CVE-2018-16864). | |
7814 | ||
7815 | * $DBUS_SESSION_BUS_ADDRESS environment variable is set by pam_systemd | |
7816 | again. | |
7817 | ||
08e1fe42 ZJS |
7818 | * A new network device NamePolicy "keep" is implemented for link files, |
7819 | and used by default in 99-default.link (the fallback configuration | |
7820 | provided by systemd). With this policy, if the network device name | |
7821 | was already set by userspace, the device will not be renamed again. | |
7822 | This matches the naming scheme that was implemented before | |
7823 | systemd-240. If naming-scheme < 240 is specified, the "keep" policy | |
7824 | is also enabled by default, even if not specified. Effectively, this | |
7825 | means that if naming-scheme >= 240 is specified, network devices will | |
7826 | be renamed according to the configuration, even if they have been | |
7827 | renamed already, if "keep" is not specified as the naming policy in | |
7828 | the .link file. The 99-default.link file provided by systemd includes | |
7829 | "keep" for backwards compatibility, but it is recommended for user | |
7830 | installed .link files to *not* include it. | |
7831 | ||
7832 | The "kernel" policy, which keeps kernel names declared to be | |
7833 | "persistent", now works again as documented. | |
7834 | ||
ba7a6b8c LP |
7835 | * kernel-install script now optionally takes the paths to one or more |
7836 | initrd files, and passes them to all plugins. | |
bd36ef0a | 7837 | |
57c03b1e LP |
7838 | * The mincore() system call has been dropped from the @system-service |
7839 | system call filter group, as it is pretty exotic and may potentially | |
7840 | used for side-channel attacks. | |
7841 | ||
774d6375 ZJS |
7842 | * -fPIE is dropped from compiler and linker options. Please specify |
7843 | -Db_pie=true option to meson to build position-independent | |
bd36ef0a YW |
7844 | executables. Note that the meson option is supported since meson-0.49. |
7845 | ||
27325875 LW |
7846 | * The fs.protected_regular and fs.protected_fifos sysctls, which were |
7847 | added in Linux 4.19 to make some data spoofing attacks harder, are | |
7848 | now enabled by default. While this will hopefully improve the | |
7849 | security of most installations, it is technically a backwards | |
7850 | incompatible change; to disable these sysctls again, place the | |
7851 | following lines in /etc/sysctl.d/60-protected.conf or a similar file: | |
7852 | ||
7853 | fs.protected_regular = 0 | |
7854 | fs.protected_fifos = 0 | |
7855 | ||
7856 | Note that the similar hardlink and symlink protection has been | |
7857 | enabled since v199, and may be disabled likewise. | |
7858 | ||
a77f438b LT |
7859 | * The files read from the EnvironmentFile= setting in unit files now |
7860 | parse backslashes inside quotes literally, matching the behaviour of | |
7861 | POSIX shells. | |
7862 | ||
ba7a6b8c LP |
7863 | * udevadm trigger, udevadm control, udevadm settle and udevadm monitor |
7864 | now automatically become NOPs when run in a chroot() environment. | |
7865 | ||
7866 | * The tmpfiles.d/ "C" line type will now copy directory trees not only | |
7867 | when the destination is so far missing, but also if it already exists | |
7868 | as a directory and is empty. This is useful to cater for systems | |
7869 | where directory trees are put together from multiple separate mount | |
7870 | points but otherwise empty. | |
7871 | ||
7872 | * A new function sd_bus_close_unref() (and the associated | |
7873 | sd_bus_close_unrefp()) has been added to libsystemd, that combines | |
7874 | sd_bus_close() and sd_bus_unref() in one. | |
7875 | ||
7876 | * udevadm control learnt a new option for --ping for testing whether a | |
7877 | systemd-udevd instance is running and reacting. | |
7878 | ||
ecebd1ec YW |
7879 | * udevadm trigger learnt a new option for --wait-daemon for waiting |
7880 | systemd-udevd daemon to be initialized. | |
7881 | ||
d0f71749 LP |
7882 | Contributions from: Aaron Plattner, Alberts Muktupāvels, Alex Mayer, |
7883 | Ayman Bagabas, Beniamino Galvani, Burt P, Chris Down, Chris Lamb, Chris | |
7884 | Morin, Christian Hesse, Claudius Ellsel, dana, Daniel Axtens, Daniele | |
7885 | Medri, Dave Reisner, David Santamaría Rogado, Diego Canuhe, Dimitri | |
7886 | John Ledkov, Evgeny Vereshchagin, Fabrice Fontaine, Filipe | |
7887 | Brandenburger, Franck Bui, Frantisek Sumsal, govwin, Hans de Goede, | |
7888 | James Hilliard, Jan Engelhardt, Jani Uusitalo, Jan Janssen, Jan | |
7889 | Synacek, Jonathan McDowell, Jonathan Roemer, Jonathon Kowalski, Joost | |
7890 | Heitbrink, Jörg Thalheim, Lance, Lennart Poettering, Louis Taylor, | |
7891 | Lucas Werkmeister, Mantas Mikulėnas, Marc-Antoine Perennou, | |
7892 | marvelousblack, Michael Biebl, Michael Sloan, Michal Sekletar, Mike | |
7893 | Auty, Mike Gilbert, Mikhail Kasimov, Neil Brown, Niklas Hambüchen, | |
7894 | Patrick Williams, Paul Seyfert, Peter Hutterer, Philip Withnall, Roger | |
7895 | James, Ronnie P. Thomas, Ryan Gonzalez, Sam Morris, Stephan Edel, | |
7896 | Stephan Gerhold, Susant Sahani, Taro Yamada, Thomas Haller, Topi | |
7897 | Miettinen, YiFei Zhu, YmrDtnJu, YunQiang Su, Yu Watanabe, Zbigniew | |
7898 | Jędrzejewski-Szmek, zsergeant77, Дамјан Георгиевски | |
7899 | ||
36d28ebc | 7900 | — Berlin, 2019-02-14 |
ba7a6b8c | 7901 | |
32673162 | 7902 | CHANGES WITH 240: |
fcb97512 | 7903 | |
e68a35a7 ZJS |
7904 | * NoNewPrivileges=yes has been set for all long-running services |
7905 | implemented by systemd. Previously, this was problematic due to | |
7906 | SELinux (as this would also prohibit the transition from PID1's label | |
7907 | to the service's label). This restriction has since been lifted, but | |
7908 | an SELinux policy update is required. | |
7909 | (See e.g. https://github.com/fedora-selinux/selinux-policy/pull/234.) | |
7910 | ||
aa2437e2 YW |
7911 | * DynamicUser=yes is dropped from systemd-networkd.service, |
7912 | systemd-resolved.service and systemd-timesyncd.service, which was | |
7913 | enabled in v239 for systemd-networkd.service and systemd-resolved.service, | |
7914 | and since v236 for systemd-timesyncd.service. The users and groups | |
7915 | systemd-network, systemd-resolve and systemd-timesync are created | |
7916 | by systemd-sysusers again. Distributors or system administrators | |
7917 | may need to create these users and groups if they not exist (or need | |
7918 | to re-enable DynamicUser= for those units) while upgrading systemd. | |
787a133f YW |
7919 | Also, the clock file for systemd-timesyncd may need to move from |
7920 | /var/lib/private/systemd/timesync/clock to /var/lib/systemd/timesync/clock. | |
aa2437e2 | 7921 | |
b1a082cd ZJS |
7922 | * When unit files are loaded from disk, previously systemd would |
7923 | sometimes (depending on the unit loading order) load units from the | |
7924 | target path of symlinks in .wants/ or .requires/ directories of other | |
7925 | units. This meant that unit could be loaded from different paths | |
7926 | depending on whether the unit was requested explicitly or as a | |
7927 | dependency of another unit, not honouring the priority of directories | |
7928 | in search path. It also meant that it was possible to successfully | |
7929 | load and start units which are not found in the unit search path, as | |
7930 | long as they were requested as a dependency and linked to from | |
7931 | .wants/ or .requires/. The target paths of those symlinks are not | |
7932 | used for loading units anymore and the unit file must be found in | |
7933 | the search path. | |
7934 | ||
fcb97512 | 7935 | * A new service type has been added: Type=exec. It's very similar to |
421e3b45 | 7936 | Type=simple but ensures the service manager will wait for both fork() |
fcb97512 LP |
7937 | and execve() of the main service binary to complete before proceeding |
7938 | with follow-up units. This is primarily useful so that the manager | |
7939 | propagates any errors in the preparation phase of service execution | |
7940 | back to the job that requested the unit to be started. For example, | |
7941 | consider a service that has ExecStart= set to a file system binary | |
421e3b45 ZJS |
7942 | that doesn't exist. With Type=simple starting the unit would be |
7943 | considered instantly successful, as only fork() has to complete | |
7944 | successfully and the manager does not wait for execve(), and hence | |
7945 | its failure is seen "too late". With the new Type=exec service type | |
7946 | starting the unit will fail, as the manager will wait for the | |
7947 | execve() and notice its failure, which is then propagated back to the | |
7948 | start job. | |
fcb97512 LP |
7949 | |
7950 | NOTE: with the next release 241 of systemd we intend to change the | |
7951 | systemd-run tool to default to Type=exec for transient services | |
7952 | started by it. This should be mostly safe, but in specific corner | |
7953 | cases might result in problems, as the systemd-run tool will then | |
6b1ab752 | 7954 | block on NSS calls (such as user name look-ups due to User=) done |
fcb97512 LP |
7955 | between the fork() and execve(), which under specific circumstances |
7956 | might cause problems. It is recommended to specify "-p Type=simple" | |
7957 | explicitly in the few cases where this applies. For regular, | |
7958 | non-transient services (i.e. those defined with unit files on disk) | |
7959 | we will continue to default to Type=simple. | |
7960 | ||
0972c1ae LP |
7961 | * The Linux kernel's current default RLIMIT_NOFILE resource limit for |
7962 | userspace processes is set to 1024 (soft) and 4096 | |
7963 | (hard). Previously, systemd passed this on unmodified to all | |
7964 | processes it forked off. With this systemd release the hard limit | |
0abf9492 | 7965 | systemd passes on is increased to 512K, overriding the kernel's |
0972c1ae LP |
7966 | defaults and substantially increasing the number of simultaneous file |
7967 | descriptors unprivileged userspace processes can allocate. Note that | |
7968 | the soft limit remains at 1024 for compatibility reasons: the | |
7969 | traditional UNIX select() call cannot deal with file descriptors >= | |
7970 | 1024 and increasing the soft limit globally might thus result in | |
7971 | programs unexpectedly allocating a high file descriptor and thus | |
7972 | failing abnormally when attempting to use it with select() (of | |
7973 | course, programs shouldn't use select() anymore, and prefer | |
7974 | poll()/epoll, but the call unfortunately remains undeservedly popular | |
7975 | at this time). This change reflects the fact that file descriptor | |
7976 | handling in the Linux kernel has been optimized in more recent | |
7977 | kernels and allocating large numbers of them should be much cheaper | |
7978 | both in memory and in performance than it used to be. Programs that | |
7979 | want to take benefit of the increased limit have to "opt-in" into | |
421e3b45 ZJS |
7980 | high file descriptors explicitly by raising their soft limit. Of |
7981 | course, when they do that they must acknowledge that they cannot use | |
7982 | select() anymore (and neither can any shared library they use — or | |
7983 | any shared library used by any shared library they use and so on). | |
7984 | Which default hard limit is most appropriate is of course hard to | |
7985 | decide. However, given reports that ~300K file descriptors are used | |
7986 | in real-life applications we believe 512K is sufficiently high as new | |
7987 | default for now. Note that there are also reports that using very | |
7988 | high hard limits (e.g. 1G) is problematic: some software allocates | |
7989 | large arrays with one element for each potential file descriptor | |
7990 | (Java, …) — a high hard limit thus triggers excessively large memory | |
7991 | allocations in these applications. Hopefully, the new default of 512K | |
7992 | is a good middle ground: higher than what real-life applications | |
7993 | currently need, and low enough for avoid triggering excessively large | |
7994 | allocations in problematic software. (And yes, somebody should fix | |
7995 | Java.) | |
0972c1ae | 7996 | |
a8b627aa LP |
7997 | * The fs.nr_open and fs.file-max sysctls are now automatically bumped |
7998 | to the highest possible values, as separate accounting of file | |
7999 | descriptors is no longer necessary, as memcg tracks them correctly as | |
8000 | part of the memory accounting anyway. Thus, from the four limits on | |
8001 | file descriptors currently enforced (fs.file-max, fs.nr_open, | |
8002 | RLIMIT_NOFILE hard, RLIMIT_NOFILE soft) we turn off the first two, | |
8003 | and keep only the latter two. A set of build-time options | |
a579d42a | 8004 | (-Dbump-proc-sys-fs-file-max=false and -Dbump-proc-sys-fs-nr-open=false) |
a8b627aa LP |
8005 | has been added to revert this change in behaviour, which might be |
8006 | an option for systems that turn off memcg in the kernel. | |
8007 | ||
4f7dc24f LP |
8008 | * When no /etc/locale.conf file exists (and hence no locale settings |
8009 | are in place), systemd will now use the "C.UTF-8" locale by default, | |
8010 | and set LANG= to it. This locale is supported by various | |
8011 | distributions including Fedora, with clear indications that upstream | |
8012 | glibc is going to make it available too. This locale enables UTF-8 | |
8013 | mode by default, which appears appropriate for 2018. | |
8014 | ||
230450d4 LR |
8015 | * The "net.ipv4.conf.all.rp_filter" sysctl will now be set to 2 by |
8016 | default. This effectively switches the RFC3704 Reverse Path filtering | |
8017 | from Strict mode to Loose mode. This is more appropriate for hosts | |
8018 | that have multiple links with routes to the same networks (e.g. | |
8019 | a client with a Wi-Fi and Ethernet both connected to the internet). | |
8020 | ||
6b1ab752 | 8021 | Consult the kernel documentation for details on this sysctl: |
0e685823 | 8022 | https://docs.kernel.org/networking/ip-sysctl.html |
195d181c | 8023 | |
e447ffe4 SI |
8024 | * The v239 change to turn on "net.ipv4.tcp_ecn" by default has been |
8025 | reverted. | |
230450d4 | 8026 | |
23305a29 CD |
8027 | * CPUAccounting=yes no longer enables the CPU controller when using |
8028 | kernel 4.15+ and the unified cgroup hierarchy, as required accounting | |
8029 | statistics are now provided independently from the CPU controller. | |
8030 | ||
6b1ab752 | 8031 | * Support for disabling a particular cgroup controller within a sub-tree |
a8467688 CD |
8032 | has been added through the DisableControllers= directive. |
8033 | ||
8f044cf9 CD |
8034 | * cgroup_no_v1=all on the kernel command line now also implies |
8035 | using the unified cgroup hierarchy, unless one explicitly passes | |
8036 | systemd.unified_cgroup_hierarchy=0 on the kernel command line. | |
8037 | ||
6b1ab752 LP |
8038 | * The new "MemoryMin=" unit file property may now be used to set the |
8039 | memory usage protection limit of processes invoked by the unit. This | |
4e1dfa45 | 8040 | controls the cgroup v2 memory.min attribute. Similarly, the new |
6b1ab752 | 8041 | "IODeviceLatencyTargetSec=" property has been added, wrapping the new |
4e1dfa45 | 8042 | cgroup v2 io.latency cgroup property for configuring per-service I/O |
6b1ab752 LP |
8043 | latency. |
8044 | ||
4e1dfa45 CD |
8045 | * systemd now supports the cgroup v2 devices BPF logic, as counterpart |
8046 | to the cgroup v1 "devices" cgroup controller. | |
6b1ab752 LP |
8047 | |
8048 | * systemd-escape now is able to combine --unescape with --template. It | |
8049 | also learnt a new option --instance for extracting and unescaping the | |
8050 | instance part of a unit name. | |
8051 | ||
8052 | * sd-bus now provides the sd_bus_message_readv() which is similar to | |
8053 | sd_bus_message_read() but takes a va_list object. The pair | |
8054 | sd_bus_set_method_call_timeout() and sd_bus_get_method_call_timeout() | |
421e3b45 | 8055 | has been added for configuring the default method call timeout to |
6b1ab752 LP |
8056 | use. sd_bus_error_move() may be used to efficiently move the contents |
8057 | from one sd_bus_error structure to another, invalidating the | |
8058 | source. sd_bus_set_close_on_exit() and sd_bus_get_close_on_exit() may | |
8059 | be used to control whether a bus connection object is automatically | |
8060 | flushed when an sd-event loop is exited. | |
8061 | ||
8062 | * When processing classic BSD syslog log messages, journald will now | |
8063 | save the original time-stamp string supplied in the new | |
8064 | SYSLOG_TIMESTAMP= journal field. This permits consumers to | |
8065 | reconstruct the original BSD syslog message more correctly. | |
8066 | ||
8067 | * StandardOutput=/StandardError= in service files gained support for | |
8068 | new "append:…" parameters, for connecting STDOUT/STDERR of a service | |
8069 | to a file, and appending to it. | |
8070 | ||
8071 | * The signal to use as last step of killing of unit processes is now | |
8072 | configurable. Previously it was hard-coded to SIGKILL, which may now | |
8073 | be overridden with the new KillSignal= setting. Note that this is the | |
46b028f2 | 8074 | signal used when regular termination (i.e. SIGTERM) does not suffice. |
421e3b45 ZJS |
8075 | Similarly, the signal used when aborting a program in case of a |
8076 | watchdog timeout may now be configured too (WatchdogSignal=). | |
6b1ab752 LP |
8077 | |
8078 | * The XDG_SESSION_DESKTOP environment variable may now be configured in | |
8079 | the pam_systemd argument line, using the new desktop= switch. This is | |
8080 | useful to initialize it properly from a display manager without | |
8081 | having to touch C code. | |
8082 | ||
421e3b45 ZJS |
8083 | * Most configuration options that previously accepted percentage values |
8084 | now also accept permille values with the '‰' suffix (instead of '%'). | |
6b1ab752 | 8085 | |
6b1ab752 LP |
8086 | * systemd-resolved may now optionally use OpenSSL instead of GnuTLS for |
8087 | DNS-over-TLS. | |
8088 | ||
8089 | * systemd-resolved's configuration file resolved.conf gained a new | |
8090 | option ReadEtcHosts= which may be used to turn off processing and | |
8091 | honoring /etc/hosts entries. | |
8092 | ||
8093 | * The "--wait" switch may now be passed to "systemctl | |
8094 | is-system-running", in which case the tool will synchronously wait | |
8095 | until the system finished start-up. | |
8096 | ||
8097 | * hostnamed gained a new bus call to determine the DMI product UUID. | |
8098 | ||
8099 | * On x86-64 systemd will now prefer using the RDRAND processor | |
8100 | instruction over /dev/urandom whenever it requires randomness that | |
8101 | neither has to be crypto-grade nor should be reproducible. This | |
8102 | should substantially reduce the amount of entropy systemd requests | |
8103 | from the kernel during initialization on such systems, though not | |
8104 | reduce it to zero. (Why not zero? systemd still needs to allocate | |
8105 | UUIDs and such uniquely, which require high-quality randomness.) | |
8106 | ||
8107 | * networkd gained support for Foo-Over-UDP, ERSPAN and ISATAP | |
8108 | tunnels. It also gained a new option ForceDHCPv6PDOtherInformation= | |
8109 | for forcing the "Other Information" bit in IPv6 RA messages. The | |
d6131be9 | 8110 | bonding logic gained four new options AdActorSystemPriority=, |
6b1ab752 | 8111 | AdUserPortKey=, AdActorSystem= for configuring various 802.3ad |
d6131be9 YW |
8112 | aspects, and DynamicTransmitLoadBalancing= for enabling dynamic |
8113 | shuffling of flows. The tunnel logic gained a new | |
8114 | IPv6RapidDeploymentPrefix= option for configuring IPv6 Rapid | |
8115 | Deployment. The policy rule logic gained four new options IPProtocol=, | |
8116 | SourcePort= and DestinationPort=, InvertRule=. The bridge logic gained | |
8117 | support for the MulticastToUnicast= option. networkd also gained | |
8118 | support for configuring static IPv4 ARP or IPv6 neighbor entries. | |
6b1ab752 LP |
8119 | |
8120 | * .preset files (as read by 'systemctl preset') may now be used to | |
8121 | instantiate services. | |
8122 | ||
8123 | * /etc/crypttab now understands the sector-size= option to configure | |
8124 | the sector size for an encrypted partition. | |
8125 | ||
8126 | * Key material for encrypted disks may now be placed on a formatted | |
421e3b45 ZJS |
8127 | medium, and referenced from /etc/crypttab by the UUID of the file |
8128 | system, followed by "=" suffixed by the path to the key file. | |
6b1ab752 LP |
8129 | |
8130 | * The "collect" udev component has been removed without replacement, as | |
421e3b45 | 8131 | it is neither used nor maintained. |
6b1ab752 LP |
8132 | |
8133 | * When the RuntimeDirectory=, StateDirectory=, CacheDirectory=, | |
8134 | LogsDirectory=, ConfigurationDirectory= settings are used in a | |
8135 | service the executed processes will now receive a set of environment | |
421e3b45 ZJS |
8136 | variables containing the full paths of these directories. |
8137 | Specifically, RUNTIME_DIRECTORY=, STATE_DIRECTORY, CACHE_DIRECTORY, | |
8138 | LOGS_DIRECTORY, CONFIGURATION_DIRECTORY are now set if these options | |
8139 | are used. Note that these options may be used multiple times per | |
8140 | service in which case the resulting paths will be concatenated and | |
8141 | separated by colons. | |
6b1ab752 LP |
8142 | |
8143 | * Predictable interface naming has been extended to cover InfiniBand | |
8144 | NICs. They will be exposed with an "ib" prefix. | |
8145 | ||
8146 | * tmpfiles.d/ line types may now be suffixed with a '-' character, in | |
8147 | which case the respective line failing is ignored. | |
8148 | ||
8149 | * .link files may now be used to configure the equivalent to the | |
8150 | "ethtool advertise" commands. | |
8151 | ||
8152 | * The sd-device.h and sd-hwdb.h APIs are now exported, as an | |
8153 | alternative to libudev.h. Previously, the latter was just an internal | |
8154 | wrapper around the former, but now these two APIs are exposed | |
8155 | directly. | |
8156 | ||
8157 | * sd-id128.h gained a new function sd_id128_get_boot_app_specific() | |
8158 | which calculates an app-specific boot ID similar to how | |
8159 | sd_id128_get_machine_app_specific() generates an app-specific machine | |
8160 | ID. | |
8161 | ||
8162 | * A new tool systemd-id128 has been added that can be used to determine | |
da890466 | 8163 | and generate various 128-bit IDs. |
6b1ab752 LP |
8164 | |
8165 | * /etc/os-release gained two new standardized fields DOCUMENTATION_URL= | |
8166 | and LOGO=. | |
8167 | ||
8168 | * systemd-hibernate-resume-generator will now honor the "noresume" | |
8169 | kernel command line option, in which case it will bypass resuming | |
8170 | from any hibernated image. | |
8171 | ||
8172 | * The systemd-sleep.conf configuration file gained new options | |
8173 | AllowSuspend=, AllowHibernation=, AllowSuspendThenHibernate=, | |
8174 | AllowHybridSleep= for prohibiting specific sleep modes even if the | |
421e3b45 | 8175 | kernel exports them. |
6b1ab752 LP |
8176 | |
8177 | * portablectl is now officially supported and has thus moved to | |
8178 | /usr/bin/. | |
8179 | ||
8180 | * bootctl learnt the two new commands "set-default" and "set-oneshot" | |
8181 | for setting the default boot loader item to boot to (either | |
8182 | persistently or only for the next boot). This is currently only | |
8183 | compatible with sd-boot, but may be implemented on other boot loaders | |
8184 | too, that follow the boot loader interface. The updated interface is | |
8185 | now documented here: | |
8186 | ||
8187 | https://systemd.io/BOOT_LOADER_INTERFACE | |
8188 | ||
8189 | * A new kernel command line option systemd.early_core_pattern= is now | |
8190 | understood which may be used to influence the core_pattern PID 1 | |
8191 | installs during early boot. | |
8192 | ||
8193 | * busctl learnt two new options -j and --json= for outputting method | |
8194 | call replies, properties and monitoring output in JSON. | |
8195 | ||
8196 | * journalctl's JSON output now supports simple ANSI coloring as well as | |
8197 | a new "json-seq" mode for generating RFC7464 output. | |
8198 | ||
8199 | * Unit files now support the %g/%G specifiers that resolve to the UNIX | |
8200 | group/GID of the service manager runs as, similar to the existing | |
8201 | %u/%U specifiers that resolve to the UNIX user/UID. | |
8202 | ||
8203 | * systemd-logind learnt a new global configuration option | |
8204 | UserStopDelaySec= that may be set in logind.conf. It specifies how | |
8205 | long the systemd --user instance shall remain started after a user | |
8206 | logs out. This is useful to speed up repetitive re-connections of the | |
8207 | same user, as it means the user's service manager doesn't have to be | |
8208 | stopped/restarted on each iteration, but can be reused between | |
8209 | subsequent options. This setting defaults to 10s. systemd-logind also | |
8210 | exports two new properties on its Manager D-Bus objects indicating | |
421e3b45 ZJS |
8211 | whether the system's lid is currently closed, and whether the system |
8212 | is on AC power. | |
6b1ab752 LP |
8213 | |
8214 | * systemd gained support for a generic boot counting logic, which | |
8215 | generically permits automatic reverting to older boot loader entries | |
8216 | if newer updated ones don't work. The boot loader side is implemented | |
8217 | in sd-boot, but is kept open for other boot loaders too. For details | |
8218 | see: | |
8219 | ||
8220 | https://systemd.io/AUTOMATIC_BOOT_ASSESSMENT | |
8221 | ||
8222 | * The SuccessAction=/FailureAction= unit file settings now learnt two | |
8223 | new parameters: "exit" and "exit-force", which result in immediate | |
8224 | exiting of the service manager, and are only useful in systemd --user | |
8225 | and container environments. | |
8226 | ||
8227 | * Unit files gained support for a pair of options | |
8228 | FailureActionExitStatus=/SuccessActionExitStatus= for configuring the | |
8229 | exit status to use as service manager exit status when | |
8230 | SuccessAction=/FailureAction= is set to exit or exit-force. | |
8231 | ||
8232 | * A pair of LogRateLimitIntervalSec=/LogRateLimitBurst= per-service | |
8233 | options may now be used to configure the log rate limiting applied by | |
8234 | journald per-service. | |
8235 | ||
8236 | * systemd-analyze gained a new verb "timespan" for parsing and | |
8237 | normalizing time span values (i.e. strings like "5min 7s 8us"). | |
8238 | ||
8239 | * systemd-analyze also gained a new verb "security" for analyzing the | |
8240 | security and sand-boxing settings of services in order to determine an | |
8241 | "exposure level" for them, indicating whether a service would benefit | |
8242 | from more sand-boxing options turned on for them. | |
8243 | ||
8244 | * "systemd-analyze syscall-filter" will now also show system calls | |
8245 | supported by the local kernel but not included in any of the defined | |
8246 | groups. | |
8247 | ||
8248 | * .nspawn files now understand the Ephemeral= setting, matching the | |
8249 | --ephemeral command line switch. | |
8250 | ||
8251 | * sd-event gained the new APIs sd_event_source_get_floating() and | |
8252 | sd_event_source_set_floating() for controlling whether a specific | |
8253 | event source is "floating", i.e. destroyed along with the even loop | |
8254 | object itself. | |
8255 | ||
8256 | * Unit objects on D-Bus gained a new "Refs" property that lists all | |
421e3b45 ZJS |
8257 | clients that currently have a reference on the unit (to ensure it is |
8258 | not unloaded). | |
6b1ab752 LP |
8259 | |
8260 | * The JoinControllers= option in system.conf is no longer supported, as | |
8261 | it didn't work correctly, is hard to support properly, is legacy (as | |
4e1dfa45 | 8262 | the concept only exists on cgroup v1) and apparently wasn't used. |
6b1ab752 LP |
8263 | |
8264 | * Journal messages that are generated whenever a unit enters the failed | |
421e3b45 ZJS |
8265 | state are now tagged with a unique MESSAGE_ID. Similarly, messages |
8266 | generated whenever a service process exits are now made recognizable, | |
5238e957 | 8267 | too. A tagged message is also emitted whenever a unit enters the |
421e3b45 | 8268 | "dead" state on success. |
6b1ab752 LP |
8269 | |
8270 | * systemd-run gained a new switch --working-directory= for configuring | |
8271 | the working directory of the service to start. A shortcut -d is | |
8272 | equivalent, setting the working directory of the service to the | |
8273 | current working directory of the invoking program. The new --shell | |
8274 | (or just -S) option has been added for invoking the $SHELL of the | |
8275 | caller as a service, and implies --pty --same-dir --wait --collect | |
421e3b45 | 8276 | --service-type=exec. Or in other words, "systemd-run -S" is now the |
6b1ab752 LP |
8277 | quickest way to quickly get an interactive in a fully clean and |
8278 | well-defined system service context. | |
8279 | ||
8280 | * machinectl gained a new verb "import-fs" for importing an OS tree | |
8281 | from a directory. Moreover, when a directory or tarball is imported | |
8282 | and single top-level directory found with the OS itself below the OS | |
8283 | tree is automatically mangled and moved one level up. | |
8284 | ||
421e3b45 ZJS |
8285 | * systemd-importd will no longer set up an implicit btrfs loop-back |
8286 | file system on /var/lib/machines. If one is already set up, it will | |
8287 | continue to be used. | |
6b1ab752 LP |
8288 | |
8289 | * A new generator "systemd-run-generator" has been added. It will | |
8290 | synthesize a unit from one or more program command lines included in | |
8291 | the kernel command line. This is very useful in container managers | |
8292 | for example: | |
8293 | ||
8294 | # systemd-nspawn -i someimage.raw -b systemd.run='"some command line"' | |
8295 | ||
8296 | This will run "systemd-nspawn" on an image, invoke the specified | |
421e3b45 ZJS |
8297 | command line and immediately shut down the container again, returning |
8298 | the command line's exit code. | |
6b1ab752 | 8299 | |
421e3b45 | 8300 | * The block device locking logic is now documented: |
6b1ab752 LP |
8301 | |
8302 | https://systemd.io/BLOCK_DEVICE_LOCKING | |
8303 | ||
8304 | * loginctl and machinectl now optionally output the various tables in | |
8305 | JSON using the --output= switch. It is our intention to add similar | |
8306 | support to systemctl and all other commands. | |
8307 | ||
8308 | * udevadm's query and trigger verb now optionally take a .device unit | |
8309 | name as argument. | |
8310 | ||
8311 | * systemd-udevd's network naming logic now understands a new | |
421e3b45 | 8312 | net.naming-scheme= kernel command line switch, which may be used to |
6b1ab752 LP |
8313 | pick a specific version of the naming scheme. This helps stabilizing |
8314 | interface names even as systemd/udev are updated and the naming logic | |
8315 | is improved. | |
8316 | ||
67081438 | 8317 | * sd-id128.h learnt two new auxiliary helpers: sd_id128_is_allf() and |
da890466 | 8318 | SD_ID128_ALLF to test if a 128-bit ID is set to all 0xFF bytes, and to |
67081438 LP |
8319 | initialize one to all 0xFF. |
8320 | ||
144d7f1d LP |
8321 | * After loading the SELinux policy systemd will now recursively relabel |
8322 | all files and directories listed in | |
8323 | /run/systemd/relabel-extra.d/*.relabel (which should be simple | |
8324 | newline separated lists of paths) in addition to the ones it already | |
8325 | implicitly relabels in /run, /dev and /sys. After the relabelling is | |
8326 | completed the *.relabel files (and /run/systemd/relabel-extra.d/) are | |
8327 | removed. This is useful to permit initrds (i.e. code running before | |
8328 | the SELinux policy is in effect) to generate files in the host | |
8329 | filesystem safely and ensure that the correct label is applied during | |
8330 | the transition to the host OS. | |
8331 | ||
98a7b55a LP |
8332 | * KERNEL API BREAKAGE: Linux kernel 4.18 changed behaviour regarding |
8333 | mknod() handling in user namespaces. Previously mknod() would always | |
8334 | fail with EPERM in user namespaces. Since 4.18 mknod() will succeed | |
8335 | but device nodes generated that way cannot be opened, and attempts to | |
8336 | open them result in EPERM. This breaks the "graceful fallback" logic | |
8337 | in systemd's PrivateDevices= sand-boxing option. This option is | |
8338 | implemented defensively, so that when systemd detects it runs in a | |
8339 | restricted environment (such as a user namespace, or an environment | |
8340 | where mknod() is blocked through seccomp or absence of CAP_SYS_MKNOD) | |
8341 | where device nodes cannot be created the effect of PrivateDevices= is | |
8342 | bypassed (following the logic that 2nd-level sand-boxing is not | |
8343 | essential if the system systemd runs in is itself already sand-boxed | |
8344 | as a whole). This logic breaks with 4.18 in container managers where | |
8345 | user namespacing is used: suddenly PrivateDevices= succeeds setting | |
8346 | up a private /dev/ file system containing devices nodes — but when | |
8347 | these are opened they don't work. | |
8348 | ||
d238709c | 8349 | At this point it is recommended that container managers utilizing |
98a7b55a LP |
8350 | user namespaces that intend to run systemd in the payload explicitly |
8351 | block mknod() with seccomp or similar, so that the graceful fallback | |
8352 | logic works again. | |
8353 | ||
8354 | We are very sorry for the breakage and the requirement to change | |
8355 | container configurations for newer kernels. It's purely caused by an | |
8356 | incompatible kernel change. The relevant kernel developers have been | |
8357 | notified about this userspace breakage quickly, but they chose to | |
8358 | ignore it. | |
8359 | ||
455027c9 ZJS |
8360 | * PermissionsStartOnly= setting is deprecated (but is still supported |
8361 | for backwards compatibility). The same functionality is provided by | |
8362 | the more flexible "+", "!", and "!!" prefixes to ExecStart= and other | |
8363 | commands. | |
8364 | ||
b4ff3dbb ZJS |
8365 | * $DBUS_SESSION_BUS_ADDRESS environment variable is not set by |
8366 | pam_systemd anymore. | |
8367 | ||
08e1fe42 ZJS |
8368 | * The naming scheme for network devices was changed to always rename |
8369 | devices, even if they were already renamed by userspace. The "kernel" | |
8370 | policy was changed to only apply as a fallback, if no other naming | |
8371 | policy took effect. | |
8372 | ||
bd36ef0a YW |
8373 | * The requirements to build systemd is bumped to meson-0.46 and |
8374 | python-3.5. | |
8375 | ||
6b1ab752 LP |
8376 | Contributions from: afg, Alan Jenkins, Aleksei Timofeyev, Alexander |
8377 | Filippov, Alexander Kurtz, Alexey Bogdanenko, Andreas Henriksson, | |
8378 | Andrew Jorgensen, Anita Zhang, apnix-uk, Arkan49, Arseny Maslennikov, | |
8379 | asavah, Asbjørn Apeland, aszlig, Bastien Nocera, Ben Boeckel, Benedikt | |
c37e2358 LP |
8380 | Morbach, Benjamin Berg, Bruce Zhang, Carlo Caione, Cedric Viou, Chen |
8381 | Qi, Chris Chiu, Chris Down, Chris Morin, Christian Rebischke, Claudius | |
8382 | Ellsel, Colin Guthrie, dana, Daniel, Daniele Medri, Daniel Kahn | |
8383 | Gillmor, Daniel Rusek, Daniel van Vugt, Dariusz Gadomski, Dave Reisner, | |
8384 | David Anderson, Davide Cavalca, David Leeds, David Malcolm, David | |
1742aae2 ZJS |
8385 | Strauss, David Tardon, Dimitri John Ledkov, Dmitry Torokhov, dj-kaktus, |
8386 | Dongsu Park, Elias Probst, Emil Soleyman, Erik Kooistra, Ervin Peters, | |
8387 | Evgeni Golov, Evgeny Vereshchagin, Fabrice Fontaine, Faheel Ahmad, | |
8388 | Faizal Luthfi, Felix Yan, Filipe Brandenburger, Franck Bui, Frank | |
8389 | Schaefer, Frantisek Sumsal, Gautier Husson, Gianluca Boiano, Giuseppe | |
8390 | Scrivano, glitsj16, Hans de Goede, Harald Hoyer, Harry Mallon, Harshit | |
8391 | Jain, Helmut Grohne, Henry Tung, Hui Yiqun, imayoda, Insun Pyo, Iwan | |
8392 | Timmer, Jan Janssen, Jan Pokorný, Jan Synacek, Jason A. Donenfeld, | |
8393 | javitoom, Jérémy Nouhaud, Jeremy Su, Jiuyang Liu, João Paulo Rechi | |
8394 | Vita, Joe Hershberger, Joe Rayhawk, Joerg Behrmann, Joerg Steffens, | |
8395 | Jonas Dorel, Jon Ringle, Josh Soref, Julian Andres Klode, Jun Bo Bi, | |
8396 | Jürg Billeter, Keith Busch, Khem Raj, Kirill Marinushkin, Larry | |
8397 | Bernstone, Lennart Poettering, Lion Yang, Li Song, Lorenz | |
8398 | Hübschle-Schneider, Lubomir Rintel, Lucas Werkmeister, Ludwin Janvier, | |
8399 | Lukáš Nykrýn, Luke Shumaker, mal, Marc-Antoine Perennou, Marcin | |
8400 | Skarbek, Marco Trevisan (Treviño), Marian Cepok, Mario Hros, Marko | |
8401 | Myllynen, Markus Grimm, Martin Pitt, Martin Sobotka, Martin Wilck, | |
8402 | Mathieu Trudel-Lapierre, Matthew Leeds, Michael Biebl, Michael Olbrich, | |
8403 | Michael 'pbone' Pobega, Michael Scherer, Michal Koutný, Michal | |
8404 | Sekletar, Michal Soltys, Mike Gilbert, Mike Palmer, Muhammet Kara, Neal | |
8405 | Gompa, Neil Brown, Network Silence, Niklas Tibbling, Nikolas Nyby, | |
8406 | Nogisaka Sadata, Oliver Smith, Patrik Flykt, Pavel Hrdina, Paweł | |
8407 | Szewczyk, Peter Hutterer, Piotr Drąg, Ray Strode, Reinhold Mueller, | |
8408 | Renaud Métrich, Roman Gushchin, Ronny Chevalier, Rubén Suárez Alvarez, | |
8409 | Ruixin Bao, RussianNeuroMancer, Ryutaroh Matsumoto, Saleem Rashid, Sam | |
8410 | Morris, Samuel Morris, Sandy Carter, scootergrisen, Sébastien Bacher, | |
8411 | Sergey Ptashnick, Shawn Landden, Shengyao Xue, Shih-Yuan Lee | |
8412 | (FourDollars), Silvio Knizek, Sjoerd Simons, Stasiek Michalski, Stephen | |
8413 | Gallagher, Steven Allen, Steve Ramage, Susant Sahani, Sven Joachim, | |
8414 | Sylvain Plantefève, Tanu Kaskinen, Tejun Heo, Thiago Macieira, Thomas | |
8415 | Blume, Thomas Haller, Thomas H. P. Andersen, Tim Ruffing, TJ, Tobias | |
8416 | Jungel, Todd Walton, Tommi Rantala, Tomsod M, Tony Novak, Tore | |
8417 | Anderson, Trevonn, Victor Laskurain, Victor Tapia, Violet Halo, Vojtech | |
8418 | Trefny, welaq, William A. Kennington III, William Douglas, Wyatt Ward, | |
8419 | Xiang Fan, Xi Ruoyao, Xuanwo, Yann E. Morin, YmrDtnJu, Yu Watanabe, | |
8420 | Zbigniew Jędrzejewski-Szmek, Zhang Xianwei, Zsolt Dollenstein | |
8421 | ||
8422 | — Warsaw, 2018-12-21 | |
6b1ab752 | 8423 | |
e8498f82 | 8424 | CHANGES WITH 239: |
019cb3ab SH |
8425 | |
8426 | * NETWORK INTERFACE DEVICE NAMING CHANGES: systemd-udevd's "net_id" | |
d69f5282 ZJS |
8427 | builtin will name network interfaces differently than in previous |
8428 | versions for virtual network interfaces created with SR-IOV and NPAR | |
8429 | and for devices where the PCI network controller device does not have | |
8430 | a slot number associated. | |
8431 | ||
8432 | SR-IOV virtual devices are now named based on the name of the parent | |
8433 | interface, with a suffix of "v<N>", where <N> is the virtual device | |
8434 | number. Previously those virtual devices were named as if completely | |
8435 | independent. | |
8436 | ||
8437 | The ninth and later NPAR virtual devices will be named following the | |
8438 | scheme used for the first eight NPAR partitions. Previously those | |
8439 | devices were not renamed and the kernel default (eth<n>) was used. | |
8440 | ||
8441 | "net_id" will also generate names for PCI devices where the PCI | |
8442 | network controller device does not have an associated slot number | |
8443 | itself, but one of its parents does. Previously those devices were | |
8444 | not renamed and the kernel default (eth<n>) was used. | |
019cb3ab | 8445 | |
6e2d744b YW |
8446 | * AF_INET and AF_INET6 are dropped from RestrictAddressFamilies= in |
8447 | systemd-logind.service. Since v235, IPAddressDeny=any has been set to | |
fe903cf4 LP |
8448 | the unit. So, it is expected that the default behavior of |
8449 | systemd-logind is not changed. However, if distribution packagers or | |
8450 | administrators disabled or modified IPAddressDeny= setting by a | |
8451 | drop-in config file, then it may be necessary to update the file to | |
8452 | re-enable AF_INET and AF_INET6 to support network user name services, | |
8453 | e.g. NIS. | |
8454 | ||
8455 | * When the RestrictNamespaces= unit property is specified multiple | |
8456 | times, then the specified types are merged now. Previously, only the | |
8457 | last assignment was used. So, if distribution packagers or | |
8458 | administrators modified the setting by a drop-in config file, then it | |
8459 | may be necessary to update the file. | |
e0eee477 | 8460 | |
1fc83d09 LP |
8461 | * When OnFailure= is used in combination with Restart= on a service |
8462 | unit, then the specified units will no longer be triggered on | |
8463 | failures that result in restarting. Previously, the specified units | |
8464 | would be activated each time the unit failed, even when the unit was | |
8465 | going to be restarted automatically. This behaviour contradicted the | |
8466 | documentation. With this release the code is adjusted to match the | |
8467 | documentation. | |
8468 | ||
41a4c3ec LP |
8469 | * systemd-tmpfiles will now print a notice whenever it encounters |
8470 | tmpfiles.d/ lines referencing the /var/run/ directory. It will | |
8471 | recommend reworking them to use the /run/ directory instead (for | |
5cadf58e ZJS |
8472 | which /var/run/ is simply a symlinked compatibility alias). This way |
8473 | systemd-tmpfiles can properly detect line conflicts and merge lines | |
8474 | referencing the same file by two paths, without having to access | |
8475 | them. | |
41a4c3ec | 8476 | |
ce55bd5e ZJS |
8477 | * systemctl disable/unmask/preset/preset-all cannot be used with |
8478 | --runtime. Previously this was allowed, but resulted in unintuitive | |
fe903cf4 LP |
8479 | behaviour that wasn't useful. systemctl disable/unmask will now undo |
8480 | both runtime and persistent enablement/masking, i.e. it will remove | |
8481 | any relevant symlinks both in /run and /etc. | |
ce55bd5e | 8482 | |
e01d9e21 | 8483 | * Note that all long-running system services shipped with systemd will |
6b000af4 | 8484 | now default to a system call allow list (rather than a deny list, as |
e01d9e21 LP |
8485 | before). In particular, systemd-udevd will now enforce one too. For |
8486 | most cases this should be safe, however downstream distributions | |
8487 | which disabled sandboxing of systemd-udevd (specifically the | |
8488 | MountFlags= setting), might want to disable this security feature | |
6b000af4 | 8489 | too, as the default allow-listing will prohibit all mount, swap, |
e01d9e21 LP |
8490 | reboot and clock changing operations from udev rules. |
8491 | ||
5cadf58e ZJS |
8492 | * sd-boot acquired new loader configuration settings to optionally turn |
8493 | off Windows and MacOS boot partition discovery as well as | |
8494 | reboot-into-firmware menu items. It is also able to pick a better | |
8495 | screen resolution for HiDPI systems, and now provides loader | |
41a4c3ec LP |
8496 | configuration settings to change the resolution explicitly. |
8497 | ||
c9299be2 IT |
8498 | * systemd-resolved now supports DNS-over-TLS. It's still |
8499 | turned off by default, use DNSOverTLS=opportunistic to turn it on in | |
c086ce8c LP |
8500 | resolved.conf. We intend to make this the default as soon as couple |
8501 | of additional techniques for optimizing the initial latency caused by | |
8502 | establishing a TLS/TCP connection are implemented. | |
8503 | ||
73c718a9 YW |
8504 | * systemd-resolved.service and systemd-networkd.service now set |
8505 | DynamicUser=yes. The users systemd-resolve and systemd-network are | |
abc291aa LP |
8506 | not created by systemd-sysusers anymore. |
8507 | ||
8508 | NOTE: This has a chance of breaking nss-ldap and similar NSS modules | |
5238e957 | 8509 | that embed a network facing module into any process using getpwuid() |
abc291aa LP |
8510 | or related call: the dynamic allocation of the user ID for |
8511 | systemd-resolved.service means the service manager has to check NSS | |
8512 | if the user name is already taken when forking off the service. Since | |
8513 | the user in the common case won't be defined in /etc/passwd the | |
8514 | lookup is likely to trigger nss-ldap which in turn might use NSS to | |
8515 | ask systemd-resolved for hostname lookups. This will hence result in | |
8516 | a deadlock: a user name lookup in order to start | |
38b38500 | 8517 | systemd-resolved.service will result in a hostname lookup for which |
abc291aa LP |
8518 | systemd-resolved.service needs to be started already. There are |
8519 | multiple ways to work around this problem: pre-allocate the | |
8520 | "systemd-resolve" user on such systems, so that nss-ldap won't be | |
8521 | triggered; or use a different NSS package that doesn't do networking | |
8522 | in-process but provides a local asynchronous name cache; or configure | |
8523 | the NSS package to avoid lookups for UIDs in the range `pkg-config | |
8524 | systemd --variable=dynamicuidmin` … `pkg-config systemd | |
8525 | --variable=dynamicuidmax`, so that it does not consider itself | |
8526 | authoritative for the same UID range systemd allocates dynamic users | |
8527 | from. | |
73c718a9 | 8528 | |
41a4c3ec LP |
8529 | * The systemd-resolve tool has been renamed to resolvectl (it also |
8530 | remains available under the old name, for compatibility), and its | |
8531 | interface is now verb-based, similar in style to the other <xyz>ctl | |
5cadf58e ZJS |
8532 | tools, such as systemctl or loginctl. |
8533 | ||
75da262a LP |
8534 | * The resolvectl/systemd-resolve tool also provides 'resolvconf' |
8535 | compatibility. It may be symlinked under the 'resolvconf' name, in | |
5cadf58e ZJS |
8536 | which case it will take arguments and input compatible with the |
8537 | Debian and FreeBSD resolvconf tool. | |
41a4c3ec LP |
8538 | |
8539 | * Support for suspend-then-hibernate has been added, i.e. a sleep mode | |
3f9a0a52 | 8540 | where the system initially suspends, and after a timeout resumes and |
41a4c3ec LP |
8541 | hibernates again. |
8542 | ||
8543 | * networkd's ClientIdentifier= now accepts a new option "duid-only". If | |
dc3f9b5e YW |
8544 | set the client will only send a DUID as client identifier. (EDIT: the |
8545 | option was broken, and was dropped in v255.) | |
41a4c3ec LP |
8546 | |
8547 | * The nss-systemd glibc NSS module will now enumerate dynamic users and | |
8548 | groups in effect. Previously, it could resolve UIDs/GIDs to user | |
8549 | names/groups and vice versa, but did not support enumeration. | |
8550 | ||
8551 | * journald's Compress= configuration setting now optionally accepts a | |
8552 | byte threshold value. All journal objects larger than this threshold | |
8553 | will be compressed, smaller ones will not. Previously this threshold | |
8554 | was not configurable and set to 512. | |
8555 | ||
5cadf58e ZJS |
8556 | * A new system.conf setting NoNewPrivileges= is now available which may |
8557 | be used to turn off acquisition of new privileges system-wide | |
8558 | (i.e. set Linux' PR_SET_NO_NEW_PRIVS for PID 1 itself, and thus also | |
8559 | for all its children). Note that turning this option on means setuid | |
8560 | binaries and file system capabilities lose their special powers. | |
8561 | While turning on this option is a big step towards a more secure | |
8562 | system, doing so is likely to break numerous pre-existing UNIX tools, | |
8563 | in particular su and sudo. | |
41a4c3ec LP |
8564 | |
8565 | * A new service systemd-time-sync-wait.service has been added. If | |
8566 | enabled it will delay the time-sync.target unit at boot until time | |
bc99dac5 | 8567 | synchronization has been received from the network. This |
41a4c3ec LP |
8568 | functionality is useful on systems lacking a local RTC or where it is |
8569 | acceptable that the boot process shall be delayed by external network | |
8570 | services. | |
8571 | ||
8572 | * When hibernating, systemd will now inform the kernel of the image | |
8573 | write offset, on kernels new enough to support this. This means swap | |
8574 | files should work for hibernation now. | |
8575 | ||
5cadf58e ZJS |
8576 | * When loading unit files, systemd will now look for drop-in unit files |
8577 | extensions in additional places. Previously, for a unit file name | |
41a4c3ec LP |
8578 | "foo-bar-baz.service" it would look for dropin files in |
8579 | "foo-bar-baz.service.d/*.conf". Now, it will also look in | |
8580 | "foo-bar-.service.d/*.conf" and "foo-.service.d/", i.e. at the | |
8581 | service name truncated after all inner dashes. This scheme allows | |
8582 | writing drop-ins easily that apply to a whole set of unit files at | |
8583 | once. It's particularly useful for mount and slice units (as their | |
5cadf58e ZJS |
8584 | naming is prefix based), but is also useful for service and other |
8585 | units, for packages that install multiple unit files at once, | |
41a4c3ec | 8586 | following a strict naming regime of beginning the unit file name with |
5cadf58e ZJS |
8587 | the package's name. Two new specifiers are now supported in unit |
8588 | files to match this: %j and %J are replaced by the part of the unit | |
8589 | name following the last dash. | |
8590 | ||
8591 | * Unit files and other configuration files that support specifier | |
88099359 | 8592 | expansion now understand another three new specifiers: %T and %V will |
5cadf58e | 8593 | resolve to /tmp and /var/tmp respectively, or whatever temporary |
88099359 ZJS |
8594 | directory has been set for the calling user. %E will expand to either |
8595 | /etc (for system units) or $XDG_CONFIG_HOME (for user units). | |
41a4c3ec LP |
8596 | |
8597 | * The ExecStart= lines of unit files are no longer required to | |
8598 | reference absolute paths. If non-absolute paths are specified the | |
8599 | specified binary name is searched within the service manager's | |
5cadf58e ZJS |
8600 | built-in $PATH, which may be queried with 'systemd-path |
8601 | search-binaries-default'. It's generally recommended to continue to | |
8602 | use absolute paths for all binaries specified in unit files. | |
41a4c3ec | 8603 | |
c7f93e28 ZJS |
8604 | * Units gained a new load state "bad-setting", which is used when a |
8605 | unit file was loaded, but contained fatal errors which prevent it | |
ba1dc1a1 LP |
8606 | from being started (for example, a service unit has been defined |
8607 | lacking both ExecStart= and ExecStop= lines). | |
c7f93e28 | 8608 | |
41a4c3ec LP |
8609 | * coredumpctl's "gdb" verb has been renamed to "debug", in order to |
8610 | support alternative debuggers, for example lldb. The old name | |
8611 | continues to be available however, for compatibility reasons. Use the | |
5cadf58e ZJS |
8612 | new --debugger= switch or the $SYSTEMD_DEBUGGER environment variable |
8613 | to pick an alternative debugger instead of the default gdb. | |
41a4c3ec LP |
8614 | |
8615 | * systemctl and the other tools will now output escape sequences that | |
8616 | generate proper clickable hyperlinks in various terminal emulators | |
8617 | where useful (for example, in the "systemctl status" output you can | |
8618 | now click on the unit file name to quickly open it in the | |
8619 | editor/viewer of your choice). Note that not all terminal emulators | |
8620 | support this functionality yet, but many do. Unfortunately, the | |
8621 | "less" pager doesn't support this yet, hence this functionality is | |
8622 | currently automatically turned off when a pager is started (which | |
8623 | happens quite often due to auto-paging). We hope to remove this | |
5cadf58e ZJS |
8624 | limitation as soon as "less" learns these escape sequences. This new |
8625 | behaviour may also be turned off explicitly with the $SYSTEMD_URLIFY | |
8626 | environment variable. For details on these escape sequences see: | |
41a4c3ec LP |
8627 | https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda |
8628 | ||
8629 | * networkd's .network files now support a new IPv6MTUBytes= option for | |
8630 | setting the MTU used by IPv6 explicitly as well as a new MTUBytes= | |
8631 | option in the [Route] section to configure the MTU to use for | |
8632 | specific routes. It also gained support for configuration of the DHCP | |
8633 | "UserClass" option through the new UserClass= setting. It gained | |
8634 | three new options in the new [CAN] section for configuring CAN | |
8635 | networks. The MULTICAST and ALLMULTI interface flags may now be | |
8636 | controlled explicitly with the new Multicast= and AllMulticast= | |
8637 | settings. | |
8638 | ||
8639 | * networkd will now automatically make use of the kernel's route | |
8640 | expiration feature, if it is available. | |
8641 | ||
5cadf58e ZJS |
8642 | * udevd's .link files now support setting the number of receive and |
8643 | transmit channels, using the RxChannels=, TxChannels=, | |
8644 | OtherChannels=, CombinedChannels= settings. | |
8645 | ||
8646 | * Support for UDPSegmentationOffload= has been removed, given its | |
8647 | limited support in hardware, and waning software support. | |
41a4c3ec LP |
8648 | |
8649 | * networkd's .netdev files now support creating "netdevsim" interfaces. | |
8650 | ||
8651 | * PID 1 learnt a new bus call GetUnitByControlGroup() which may be used | |
8652 | to query the unit belonging to a specific kernel control group. | |
8653 | ||
5cadf58e | 8654 | * systemd-analyze gained a new verb "cat-config", which may be used to |
41a4c3ec LP |
8655 | dump the contents of any configuration file, with all its matching |
8656 | drop-in files added in, and honouring the usual search and masking | |
8657 | logic applied to systemd configuration files. For example use | |
8658 | "systemd-analyze cat-config systemd/system.conf" to get the complete | |
8659 | system configuration file of systemd how it would be loaded by PID 1 | |
5cadf58e ZJS |
8660 | itself. Similar to this, various tools such as systemd-tmpfiles or |
8661 | systemd-sysusers, gained a new option "--cat-config", which does the | |
41a4c3ec LP |
8662 | corresponding operation for their own configuration settings. For |
8663 | example, "systemd-tmpfiles --cat-config" will now output the full | |
8664 | list of tmpfiles.d/ lines in place. | |
8665 | ||
704ae536 YW |
8666 | * timedatectl gained three new verbs: "show" shows bus properties of |
8667 | systemd-timedated, "timesync-status" shows the current NTP | |
8668 | synchronization state of systemd-timesyncd, and "show-timesync" | |
8669 | shows bus properties of systemd-timesyncd. | |
41a4c3ec LP |
8670 | |
8671 | * systemd-timesyncd gained a bus interface on which it exposes details | |
8672 | about its state. | |
8673 | ||
73c718a9 YW |
8674 | * A new environment variable $SYSTEMD_TIMEDATED_NTP_SERVICES is now |
8675 | understood by systemd-timedated. It takes a colon-separated list of | |
8676 | unit names of NTP client services. The list is used by | |
8677 | "timedatectl set-ntp". | |
8678 | ||
41a4c3ec LP |
8679 | * systemd-nspawn gained a new --rlimit= switch for setting initial |
8680 | resource limits for the container payload. There's a new switch | |
5cadf58e | 8681 | --hostname= to explicitly override the container's hostname. A new |
41a4c3ec LP |
8682 | --no-new-privileges= switch may be used to control the |
8683 | PR_SET_NO_NEW_PRIVS flag for the container payload. A new | |
8684 | --oom-score-adjust= switch controls the OOM scoring adjustment value | |
8685 | for the payload. The new --cpu-affinity= switch controls the CPU | |
8686 | affinity of the container payload. The new --resolv-conf= switch | |
8687 | allows more detailed control of /etc/resolv.conf handling of the | |
5cadf58e | 8688 | container. Similarly, the new --timezone= switch allows more detailed |
41a4c3ec LP |
8689 | control of /etc/localtime handling of the container. |
8690 | ||
5cadf58e | 8691 | * systemd-detect-virt gained a new --list switch, which will print a |
41a4c3ec LP |
8692 | list of all currently known VM and container environments. |
8693 | ||
5cadf58e | 8694 | * Support for "Portable Services" has been added, see |
41a4c3ec | 8695 | doc/PORTABLE_SERVICES.md for details. Currently, the support is still |
5cadf58e ZJS |
8696 | experimental, but this is expected to change soon. Reflecting this |
8697 | experimental state, the "portablectl" binary is not installed into | |
41a4c3ec LP |
8698 | /usr/bin yet. The binary has to be called with the full path |
8699 | /usr/lib/systemd/portablectl instead. | |
8700 | ||
8701 | * journalctl's and systemctl's -o switch now knows a new log output | |
8702 | mode "with-unit". The output it generates is very similar to the | |
8703 | regular "short" mode, but displays the unit name instead of the | |
8704 | syslog tag for each log line. Also, the date is shown with timezone | |
8705 | information. This mode is probably more useful than the classic | |
8706 | "short" output mode for most purposes, except where pixel-perfect | |
8707 | compatibility with classic /var/log/messages formatting is required. | |
8708 | ||
8709 | * A new --dump-bus-properties switch has been added to the systemd | |
8710 | binary, which may be used to dump all supported D-Bus properties. | |
c7f93e28 ZJS |
8711 | (Options which are still supported, but are deprecated, are *not* |
8712 | shown.) | |
41a4c3ec | 8713 | |
41a4c3ec LP |
8714 | * sd-bus gained a set of new calls: |
8715 | sd_bus_slot_set_floating()/sd_bus_slot_get_floating() may be used to | |
8716 | enable/disable the "floating" state of a bus slot object, | |
8717 | i.e. whether the slot object pins the bus it is allocated for into | |
8718 | memory or if the bus slot object gets disconnected when the bus goes | |
8719 | away. sd_bus_open_with_description(), | |
8720 | sd_bus_open_user_with_description(), | |
8721 | sd_bus_open_system_with_description() may be used to allocate bus | |
8722 | objects and set their description string already during allocation. | |
8723 | ||
8724 | * sd-event gained support for watching inotify events from the event | |
8725 | loop, in an efficient way, sharing inotify handles between multiple | |
8726 | users. For this a new function sd_event_add_inotify() has been added. | |
8727 | ||
8728 | * sd-event and sd-bus gained support for calling special user-supplied | |
8729 | destructor functions for userdata pointers associated with | |
c7f93e28 ZJS |
8730 | sd_event_source, sd_bus_slot, and sd_bus_track objects. For this new |
8731 | functions sd_bus_slot_set_destroy_callback, | |
8732 | sd_bus_slot_get_destroy_callback, sd_bus_track_set_destroy_callback, | |
8733 | sd_bus_track_get_destroy_callback, | |
8734 | sd_event_source_set_destroy_callback, | |
8735 | sd_event_source_get_destroy_callback have been added. | |
41a4c3ec LP |
8736 | |
8737 | * The "net.ipv4.tcp_ecn" sysctl will now be turned on by default. | |
8738 | ||
8739 | * PID 1 will now automatically reschedule .timer units whenever the | |
5cadf58e | 8740 | local timezone changes. (They previously got rescheduled |
41a4c3ec LP |
8741 | automatically when the system clock changed.) |
8742 | ||
8743 | * New documentation has been added to document cgroups delegation, | |
8744 | portable services and the various code quality tools we have set up: | |
8745 | ||
a8a27374 SK |
8746 | https://github.com/systemd/systemd/blob/master/docs/CGROUP_DELEGATION.md |
8747 | https://github.com/systemd/systemd/blob/master/docs/PORTABLE_SERVICES.md | |
8748 | https://github.com/systemd/systemd/blob/master/docs/CODE_QUALITY.md | |
41a4c3ec | 8749 | |
d6906108 LP |
8750 | * The Boot Loader Specification has been added to the source tree. |
8751 | ||
a8a27374 | 8752 | https://github.com/systemd/systemd/blob/master/docs/BOOT_LOADER_SPECIFICATION.md |
d6906108 LP |
8753 | |
8754 | While moving it into our source tree we have updated it and further | |
8755 | changes are now accepted through the usual github PR workflow. | |
8756 | ||
41a4c3ec LP |
8757 | * pam_systemd will now look for PAM userdata fields systemd.memory_max, |
8758 | systemd.tasks_max, systemd.cpu_weight, systemd.io_weight set by | |
8759 | earlier PAM modules. The data in these fields is used to initialize | |
8760 | the session scope's resource properties. Thus external PAM modules | |
8761 | may now configure per-session limits, for example sourced from | |
8762 | external user databases. | |
8763 | ||
8764 | * socket units with Accept=yes will now maintain a "refused" counter in | |
8765 | addition to the existing "accepted" counter, counting connections | |
8766 | refused due to the enforced limits. | |
8767 | ||
8768 | * The "systemd-path search-binaries-default" command may now be use to | |
8769 | query the default, built-in $PATH PID 1 will pass to the services it | |
8770 | manages. | |
8771 | ||
c49a7cbd LP |
8772 | * A new unit file setting PrivateMounts= has been added. It's a boolean |
8773 | option. If enabled the unit's processes are invoked in their own file | |
8774 | system namespace. Note that this behaviour is also implied if any | |
8775 | other file system namespacing options (such as PrivateTmp=, | |
8776 | PrivateDevices=, ProtectSystem=, …) are used. This option is hence | |
8777 | primarily useful for services that do not use any of the other file | |
8778 | system namespacing options. One such service is systemd-udevd.service | |
5238e957 | 8779 | where this is now used by default. |
c49a7cbd | 8780 | |
57ab451e ZJS |
8781 | * ConditionSecurity= gained a new value "uefi-secureboot" that is true |
8782 | when the system is booted in UEFI "secure mode". | |
8783 | ||
c7668c1c LP |
8784 | * A new unit "system-update-pre.target" is added, which defines an |
8785 | optional synchronization point for offline system updates, as | |
8786 | implemented by the pre-existing "system-update.target" unit. It | |
8787 | allows ordering services before the service that executes the actual | |
8788 | update process in a generic way. | |
8789 | ||
f26ad321 ZJS |
8790 | * Systemd now emits warnings whenever .include syntax is used. |
8791 | ||
41a4c3ec | 8792 | Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale, |
ec53d48c | 8793 | Alexander Kurtz, Alex Gartrell, Anssi Hannula, Arnaud Rebillout, Brian |
bb6f071f LP |
8794 | J. Murrell, Bruno Vernay, Chris Lamb, Chris Lesiak, Christian Brauner, |
8795 | Christian Hesse, Christian Rebischke, Colin Guthrie, Daniel Dao, Daniel | |
8796 | Lin, Danylo Korostil, Davide Cavalca, David Tardon, Dimitri John | |
8797 | Ledkov, Dmitriy Geels, Douglas Christman, Elia Geretto, emelenas, Emil | |
8798 | Velikov, Evgeny Vereshchagin, Felipe Sateler, Feng Sun, Filipe | |
8799 | Brandenburger, Franck Bui, futpib, Giuseppe Scrivano, Guillem Jover, | |
8800 | guixxx, Hannes Reinecke, Hans de Goede, Harald Hoyer, Henrique Dante de | |
8801 | Almeida, Hiram van Paassen, Ian Miell, Igor Gnatenko, Ivan Shapovalov, | |
8802 | Iwan Timmer, James Cowgill, Jan Janssen, Jan Synacek, Jared Kazimir, | |
8803 | Jérémy Rosen, João Paulo Rechi Vita, Joost Heitbrink, Jui-Chi Ricky | |
8804 | Liang, Jürg Billeter, Kai-Heng Feng, Karol Augustin, Kay Sievers, | |
8805 | Krzysztof Nowicki, Lauri Tirkkonen, Lennart Poettering, Leonard König, | |
8806 | Long Li, Luca Boccassi, Lucas Werkmeister, Marcel Hoppe, Marc | |
8807 | Kleine-Budde, Mario Limonciello, Martin Jansa, Martin Wilck, Mathieu | |
8808 | Malaterre, Matteo F. Vescovi, Matthew McGinn, Matthias-Christian Ott, | |
8809 | Michael Biebl, Michael Olbrich, Michael Prokop, Michal Koutný, Michal | |
8810 | Sekletar, Mike Gilbert, Mikhail Kasimov, Milan Broz, Milan Pässler, | |
8811 | Mladen Pejaković, Muhammet Kara, Nicolas Boichat, Omer Katz, Paride | |
8812 | Legovini, Paul Menzel, Paul Milliken, Pavel Hrdina, Peter A. Bigot, | |
8813 | Peter D'Hoye, Peter Hutterer, Peter Jones, Philip Sequeira, Philip | |
61d0025d | 8814 | Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de Araujo, |
bb6f071f LP |
8815 | Ronny Chevalier, Rosen Penev, Rubén Suárez Alvarez, Ryan Gonzalez, |
8816 | Salvo Tomaselli, Sebastian Reichel, Sergey Ptashnick, Sergio Lindo | |
8817 | Mansilla, Stefan Schweter, Stephen Hemminger, Stuart Hayes, Susant | |
8818 | Sahani, Sylvain Plantefève, Thomas H. P. Andersen, Tobias Jungel, | |
8819 | Tomasz Torcz, Vito Caputo, Will Dietz, Will Thompson, Wim van Mourik, | |
8820 | Yu Watanabe, Zbigniew Jędrzejewski-Szmek | |
41a4c3ec | 8821 | |
e8498f82 | 8822 | — Berlin, 2018-06-22 |
41a4c3ec | 8823 | |
c657bff1 | 8824 | CHANGES WITH 238: |
e0c46a73 LP |
8825 | |
8826 | * The MemoryAccounting= unit property now defaults to on. After | |
8827 | discussions with the upstream control group maintainers we learnt | |
8828 | that the negative impact of cgroup memory accounting on current | |
8829 | kernels is finally relatively minimal, so that it should be safe to | |
444d5863 ZJS |
8830 | enable this by default without affecting system performance. Besides |
8831 | memory accounting only task accounting is turned on by default, all | |
8832 | other forms of resource accounting (CPU, IO, IP) remain off for now, | |
8833 | because it's not clear yet that their impact is small enough to move | |
8834 | from opt-in to opt-out. We recommend downstreams to leave memory | |
07a35e84 | 8835 | accounting on by default if kernel 4.14 or higher is primarily |
444d5863 ZJS |
8836 | used. On very resource constrained systems or when support for old |
8837 | kernels is a necessity, -Dmemory-accounting-default=false can be used | |
8838 | to revert this change. | |
e0c46a73 | 8839 | |
313c32c3 ZJS |
8840 | * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update, |
8841 | %udev_rules_update) and the journal catalog (%journal_catalog_update) | |
8842 | from the upgrade scriptlets of individual packages now do nothing. | |
8843 | Transfiletriggers have been added which will perform those updates | |
8844 | once at the end of the transaction. | |
8845 | ||
8846 | Similar transfiletriggers have been added to execute any sysctl.d | |
8847 | and binfmt.d rules. Thus, it should be unnecessary to provide any | |
8848 | scriptlets to execute this configuration from package installation | |
8849 | scripts. | |
8850 | ||
8851 | * systemd-sysusers gained a mode where the configuration to execute is | |
8852 | specified on the command line, but this configuration is not executed | |
8853 | directly, but instead it is merged with the configuration on disk, | |
8854 | and the result is executed. This is useful for package installation | |
8855 | scripts which want to create the user before installing any files on | |
8856 | disk (in case some of those files are owned by that user), while | |
8857 | still allowing local admin overrides. | |
8858 | ||
07a35e84 | 8859 | This functionality is exposed to rpm scriptlets through a new |
313c32c3 ZJS |
8860 | %sysusers_create_package macro. Old %sysusers_create and |
8861 | %sysusers_create_inline macros are deprecated. | |
8862 | ||
8863 | A transfiletrigger for sysusers.d configuration is now installed, | |
07a35e84 | 8864 | which means that it should be unnecessary to call systemd-sysusers from |
313c32c3 ZJS |
8865 | package installation scripts, unless the package installs any files |
8866 | owned by those newly-created users, in which case | |
8867 | %sysusers_create_package should be used. | |
8868 | ||
8869 | * Analogous change has been done for systemd-tmpfiles: it gained a mode | |
8870 | where the command-line configuration is merged with the configuration | |
8871 | on disk. This is exposed as the new %tmpfiles_create_package macro, | |
8872 | and %tmpfiles_create is deprecated. A transfiletrigger is installed | |
8873 | for tmpfiles.d, hence it should be unnecessary to call systemd-tmpfiles | |
8874 | from package installation scripts. | |
8875 | ||
8876 | * sysusers.d configuration for a user may now also specify the group | |
8877 | number, in addition to the user number ("u username 123:456"), or | |
8878 | without the user number ("u username -:456"). | |
8879 | ||
8880 | * Configution items for systemd-sysusers can now be specified as | |
8881 | positional arguments when the new --inline switch is used. | |
8882 | ||
8883 | * The login shell of users created through sysusers.d may now be | |
8884 | specified (previously, it was always /bin/sh for root and | |
8885 | /sbin/nologin for other users). | |
8886 | ||
8887 | * systemd-analyze gained a new --global switch to look at global user | |
8888 | configuration. It also gained a unit-paths verb to list the unit load | |
8889 | paths that are compiled into systemd (which can be used with | |
8890 | --systemd, --user, or --global). | |
8891 | ||
8892 | * udevadm trigger gained a new --settle/-w option to wait for any | |
8893 | triggered events to finish (but just those, and not any other events | |
8894 | which are triggered meanwhile). | |
8895 | ||
8896 | * The action that systemd-logind takes when the lid is closed and the | |
8897 | machine is connected to external power can now be configured using | |
8898 | HandleLidSwitchExternalPower= in logind.conf. Previously, this action | |
8899 | was determined by HandleLidSwitch=, and, for backwards compatibility, | |
8900 | is still is, if HandleLidSwitchExternalPower= is not explicitly set. | |
8901 | ||
8902 | * journalctl will periodically call sd_journal_process() to make it | |
8903 | resilient against inotify queue overruns when journal files are | |
8904 | rotated very quickly. | |
8905 | ||
8906 | * Two new functions in libsystemd — sd_bus_get_n_queued_read and | |
8907 | sd_bus_get_n_queued_write — may be used to check the number of | |
8908 | pending bus messages. | |
8909 | ||
8910 | * systemd gained a new | |
8911 | org.freedesktop.systemd1.Manager.AttachProcessesToUnit dbus call | |
8912 | which can be used to migrate foreign processes to scope and service | |
8913 | units. The primary user for this new API is systemd itself: the | |
8914 | systemd --user instance uses this call of the systemd --system | |
8915 | instance to migrate processes if it itself gets the request to | |
8916 | migrate processes and the kernel refuses this due to access | |
33db1b90 | 8917 | restrictions. Thanks to this "systemd-run --scope --user …" works |
4e1dfa45 | 8918 | again in pure cgroup v2 environments when invoked from the user |
313c32c3 ZJS |
8919 | session scope. |
8920 | ||
8921 | * A new TemporaryFileSystem= setting can be used to mask out part of | |
8922 | the real file system tree with tmpfs mounts. This may be combined | |
8923 | with BindPaths= and BindReadOnlyPaths= to hide files or directories | |
8924 | not relevant to the unit, while still allowing some paths lower in | |
8925 | the tree to be accessed. | |
8926 | ||
8927 | ProtectHome=tmpfs may now be used to hide user home and runtime | |
8928 | directories from units, in a way that is mostly equivalent to | |
8929 | "TemporaryFileSystem=/home /run/user /root". | |
8930 | ||
8931 | * Non-service units are now started with KeyringMode=shared by default. | |
8932 | This means that mount and swapon and other mount tools have access | |
8933 | to keys in the main keyring. | |
8934 | ||
8935 | * /sys/fs/bpf is now mounted automatically. | |
8936 | ||
8937 | * QNX virtualization is now detected by systemd-detect-virt and may | |
8938 | be used in ConditionVirtualization=. | |
8939 | ||
8940 | * IPAccounting= may now be enabled also for slice units. | |
8941 | ||
8942 | * A new -Dsplit-bin= build configuration switch may be used to specify | |
8943 | whether bin and sbin directories are merged, or if they should be | |
8944 | included separately in $PATH and various listings of executable | |
8945 | directories. The build configuration scripts will try to autodetect | |
8946 | the proper values of -Dsplit-usr= and -Dsplit-bin= based on build | |
8947 | system, but distributions are encouraged to configure this | |
8948 | explicitly. | |
8949 | ||
8950 | * A new -Dok-color= build configuration switch may be used to change | |
8951 | the colour of "OK" status messages. | |
8952 | ||
8953 | * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with | |
8954 | PrivateNetwork=yes was buggy in previous versions of systemd. This | |
8955 | means that after the upgrade and daemon-reexec, any such units must | |
8956 | be restarted. | |
8957 | ||
8958 | * INCOMPATIBILITY: as announced in the NEWS for 237, systemd-tmpfiles | |
8959 | will not exclude read-only files owned by root from cleanup. | |
8960 | ||
c657bff1 ZJS |
8961 | Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet, |
8962 | Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337, | |
8963 | Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo | |
8964 | de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel | |
8965 | Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny | |
8966 | Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib, | |
8967 | Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer, | |
8968 | Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld, | |
8969 | Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas | |
8970 | Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt, | |
8971 | MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren, | |
8972 | Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert | |
8973 | Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon | |
8974 | Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain | |
8975 | Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe, | |
8976 | Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić) | |
8977 | ||
8978 | — Warsaw, 2018-03-05 | |
8979 | ||
82c8e3e6 | 8980 | CHANGES WITH 237: |
2b0c59ba MP |
8981 | |
8982 | * Some keyboards come with a zoom see-saw or rocker which until now got | |
8983 | mapped to the Linux "zoomin/out" keys in hwdb. However, these | |
8984 | keycodes are not recognized by any major desktop. They now produce | |
8985 | Up/Down key events so that they can be used for scrolling. | |
8986 | ||
49e87292 LP |
8987 | * INCOMPATIBILITY: systemd-tmpfiles' "f" lines changed behaviour |
8988 | slightly: previously, if an argument was specified for lines of this | |
8989 | type (i.e. the right-most column was set) this string was appended to | |
8990 | existing files each time systemd-tmpfiles was run. This behaviour was | |
8991 | different from what the documentation said, and not particularly | |
8992 | useful, as repeated systemd-tmpfiles invocations would not be | |
8993 | idempotent and grow such files without bounds. With this release | |
15c5594b ZJS |
8994 | behaviour has been altered to match what the documentation says: |
8995 | lines of this type only have an effect if the indicated files don't | |
8996 | exist yet, and only then the argument string is written to the file. | |
49e87292 | 8997 | |
82c8e3e6 LP |
8998 | * FUTURE INCOMPATIBILITY: In systemd v238 we intend to slightly change |
8999 | systemd-tmpfiles behaviour: previously, read-only files owned by root | |
9000 | were always excluded from the file "aging" algorithm (i.e. the | |
9001 | automatic clean-up of directories like /tmp based on | |
9002 | atime/mtime/ctime). We intend to drop this restriction, and age files | |
9003 | by default even when owned by root and read-only. This behaviour was | |
9004 | inherited from older tools, but there have been requests to remove | |
9005 | it, and it's not obvious why this restriction was made in the first | |
28423d9a | 9006 | place. Please speak up now, if you are aware of software that requires |
82c8e3e6 LP |
9007 | this behaviour, otherwise we'll remove the restriction in v238. |
9008 | ||
95894b91 LP |
9009 | * A new environment variable $SYSTEMD_OFFLINE is now understood by |
9010 | systemctl. It takes a boolean argument. If on, systemctl assumes it | |
9011 | operates on an "offline" OS tree, and will not attempt to talk to the | |
9012 | service manager. Previously, this mode was implicitly enabled if a | |
9013 | chroot() environment was detected, and this new environment variable | |
9014 | now provides explicit control. | |
9015 | ||
1a0cd2c7 ZJS |
9016 | * .path and .socket units may now be created transiently, too. |
9017 | Previously only service, mount, automount and timer units were | |
95894b91 LP |
9018 | supported as transient units. The systemd-run tool has been updated |
9019 | to expose this new functionality, you may hence use it now to bind | |
9020 | arbitrary commands to path or socket activation on-the-fly from the | |
1a0cd2c7 ZJS |
9021 | command line. Moreover, almost all properties are now exposed for the |
9022 | unit types that already supported transient operation. | |
95894b91 LP |
9023 | |
9024 | * The systemd-mount command gained support for a new --owner= parameter | |
9025 | which takes a user name, which is then resolved and included in uid= | |
9026 | and gid= mount options string of the file system to mount. | |
9027 | ||
9028 | * A new unit condition ConditionControlGroupController= has been added | |
9029 | that checks whether a specific cgroup controller is available. | |
9030 | ||
9031 | * Unit files, udev's .link files, and systemd-networkd's .netdev and | |
9032 | .network files all gained support for a new condition | |
9033 | ConditionKernelVersion= for checking against specific kernel | |
9034 | versions. | |
9035 | ||
9036 | * In systemd-networkd, the [IPVLAN] section in .netdev files gained | |
6cddc792 | 9037 | support for configuring device flags in the Flags= setting. In the |
95894b91 | 9038 | same files, the [Tunnel] section gained support for configuring |
33db1b90 | 9039 | AllowLocalRemote=. The [Route] section in .network files gained |
95894b91 LP |
9040 | support for configuring InitialCongestionWindow=, |
9041 | InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now | |
9042 | understands RapidCommit=. | |
9043 | ||
9044 | * systemd-networkd's DHCPv6 support gained support for Prefix | |
9045 | Delegation. | |
9046 | ||
9047 | * sd-bus gained support for a new "watch-bind" feature. When this | |
9048 | feature is enabled, an sd_bus connection may be set up to connect to | |
9049 | an AF_UNIX socket in the file system as soon as it is created. This | |
9050 | functionality is useful for writing early-boot services that | |
9051 | automatically connect to the system bus as soon as it is started, | |
9052 | without ugly time-based polling. systemd-networkd and | |
9053 | systemd-resolved have been updated to make use of this | |
9054 | functionality. busctl exposes this functionality in a new | |
9055 | --watch-bind= command line switch. | |
9056 | ||
9057 | * sd-bus will now optionally synthesize a local "Connected" signal as | |
9058 | soon as a D-Bus connection is set up fully. This message mirrors the | |
9059 | already existing "Disconnected" signal which is synthesized when the | |
9060 | connection is terminated. This signal is generally useful but | |
9061 | particularly handy in combination with the "watch-bind" feature | |
9062 | described above. Synthesizing of this message has to be requested | |
9063 | explicitly through the new API call sd_bus_set_connected_signal(). In | |
9064 | addition a new call sd_bus_is_ready() has been added that checks | |
caf2a2d8 | 9065 | whether a connection is fully set up (i.e. between the "Connected" and |
95894b91 LP |
9066 | "Disconnected" signals). |
9067 | ||
9068 | * sd-bus gained two new calls sd_bus_request_name_async() and | |
9069 | sd_bus_release_name_async() for asynchronously registering bus | |
9070 | names. Similar, there is now sd_bus_add_match_async() for installing | |
9071 | a signal match asynchronously. All of systemd's own services have | |
9072 | been updated to make use of these calls. Doing these operations | |
9073 | asynchronously has two benefits: it reduces the risk of deadlocks in | |
9074 | case of cyclic dependencies between bus services, and it speeds up | |
9075 | service initialization since synchronization points for bus | |
9076 | round-trips are removed. | |
9077 | ||
9078 | * sd-bus gained two new calls sd_bus_match_signal() and | |
9079 | sd_bus_match_signal_async(), which are similar to sd_bus_add_match() | |
9080 | and sd_bus_add_match_async() but instead of taking a D-Bus match | |
9081 | string take match fields as normal function parameters. | |
9082 | ||
9083 | * sd-bus gained two new calls sd_bus_set_sender() and | |
9084 | sd_bus_message_set_sender() for setting the sender name of outgoing | |
9085 | messages (either for all outgoing messages or for just one specific | |
9086 | one). These calls are only useful in direct connections as on | |
9087 | brokered connections the broker fills in the sender anyway, | |
9088 | overwriting whatever the client filled in. | |
9089 | ||
9090 | * sd-event gained a new pseudo-handle that may be specified on all API | |
9091 | calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When | |
9092 | used this refers to the default event loop object of the calling | |
9093 | thread. Note however that this does not implicitly allocate one — | |
6cddc792 CR |
9094 | which has to be done prior by using sd_event_default(). Similarly |
9095 | sd-bus gained three new pseudo-handles SD_BUS_DEFAULT, | |
95894b91 LP |
9096 | SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer |
9097 | to the default bus of the specified type of the calling thread. Here | |
9098 | too this does not implicitly allocate bus connection objects, this | |
9099 | has to be done prior with sd_bus_default() and friends. | |
9100 | ||
9101 | * sd-event gained a new call pair | |
6cddc792 CR |
9102 | sd_event_source_{get|set}_io_fd_own(). This may be used to request |
9103 | automatic closure of the file descriptor an IO event source watches | |
95894b91 LP |
9104 | when the event source is destroyed. |
9105 | ||
9106 | * systemd-networkd gained support for natively configuring WireGuard | |
9107 | connections. | |
9108 | ||
6cddc792 CR |
9109 | * In previous versions systemd synthesized user records both for the |
9110 | "nobody" (UID 65534) and "root" (UID 0) users in nss-systemd and | |
9111 | internally. In order to simplify distribution-wide renames of the | |
95894b91 LP |
9112 | "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a |
9113 | new transitional flag file has been added: if | |
9114 | /etc/systemd/dont-synthesize-nobody exists synthesizing of the 65534 | |
9115 | user and group record within the systemd codebase is disabled. | |
9116 | ||
9117 | * systemd-notify gained a new --uid= option for selecting the source | |
9118 | user/UID to use for notification messages sent to the service | |
9119 | manager. | |
9120 | ||
31751f7e | 9121 | * journalctl gained a new --grep= option to list only entries in which |
e6501af8 ZJS |
9122 | the message matches a certain pattern. By default matching is case |
9123 | insensitive if the pattern is lowercase, and case sensitive | |
9124 | otherwise. Option --case-sensitive=yes|no can be used to override | |
9125 | this an specify case sensitivity or case insensitivity. | |
9126 | ||
56a29112 | 9127 | * There's now a "systemd-analyze service-watchdogs" command for printing |
508058c9 | 9128 | the current state of the service runtime watchdog, and optionally |
56a29112 | 9129 | enabling or disabling the per-service watchdogs system-wide if given a |
508058c9 LP |
9130 | boolean argument (i.e. the concept you configure in WatchdogSec=), for |
9131 | debugging purposes. There's also a kernel command line option | |
56a29112 | 9132 | systemd.service_watchdogs= for controlling the same. |
508058c9 LP |
9133 | |
9134 | * Two new "log-level" and "log-target" options for systemd-analyze were | |
bc99dac5 | 9135 | added that merge the now deprecated get-log-level, set-log-level and |
508058c9 LP |
9136 | get-log-target, set-log-target pairs. The deprecated options are still |
9137 | understood for backwards compatibility. The two new options print the | |
9138 | current value when no arguments are given, and set them when a | |
56a29112 | 9139 | level/target is given as an argument. |
95894b91 | 9140 | |
508058c9 LP |
9141 | * sysusers.d's "u" lines now optionally accept both a UID and a GID |
9142 | specification, separated by a ":" character, in order to create users | |
9143 | where UID and GID do not match. | |
9144 | ||
95894b91 | 9145 | Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov, |
508058c9 LP |
9146 | Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman |
9147 | Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton | |
9148 | Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov, | |
9149 | Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui, | |
9150 | Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian | |
9151 | Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander | |
9152 | Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen, | |
9153 | Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg | |
9154 | Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering, | |
9155 | Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt, | |
9156 | Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy, | |
9157 | Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał | |
9158 | Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf | |
9159 | Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer, | |
9160 | Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer, | |
9161 | Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani, | |
9162 | Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz | |
9163 | Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary | |
9164 | Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян | |
9165 | Палаузов | |
9166 | ||
9167 | — Brno, 2018-01-28 | |
2b0c59ba | 9168 | |
a1b2c92d | 9169 | CHANGES WITH 236: |
195b943d | 9170 | |
89780840 ZJS |
9171 | * The modprobe.d/ drop-in for the bonding.ko kernel module introduced |
9172 | in v235 has been extended to also set the dummy.ko module option | |
9173 | numdummies=0, preventing the kernel from automatically creating | |
9174 | dummy0. All dummy interfaces must now be explicitly created. | |
195b943d | 9175 | |
3925496a LP |
9176 | * Unknown '%' specifiers in configuration files are now rejected. This |
9177 | applies to units and tmpfiles.d configuration. Any percent characters | |
9178 | that are followed by a letter or digit that are not supposed to be | |
9179 | interpreted as the beginning of a specifier should be escaped by | |
9180 | doubling ("%%"). (So "size=5%" is still accepted, as well as | |
9181 | "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not | |
9182 | valid specifiers today.) | |
751223fe | 9183 | |
e6b2d948 | 9184 | * systemd-resolved now maintains a new dynamic |
89780840 ZJS |
9185 | /run/systemd/resolve/stub-resolv.conf compatibility file. It is |
9186 | recommended to make /etc/resolv.conf a symlink to it. This file | |
9187 | points at the systemd-resolved stub DNS 127.0.0.53 resolver and | |
9188 | includes dynamically acquired search domains, achieving more correct | |
9189 | DNS resolution by software that bypasses local DNS APIs such as NSS. | |
e6b2d948 | 9190 | |
67eb5b38 LP |
9191 | * The "uaccess" udev tag has been dropped from /dev/kvm and |
9192 | /dev/dri/renderD*. These devices now have the 0666 permissions by | |
9193 | default (but this may be changed at build-time). /dev/dri/renderD* | |
9194 | will now be owned by the "render" group along with /dev/kfd. | |
9195 | ||
89780840 ZJS |
9196 | * "DynamicUser=yes" has been enabled for systemd-timesyncd.service, |
9197 | systemd-journal-gatewayd.service and | |
9198 | systemd-journal-upload.service. This means "nss-systemd" must be | |
9199 | enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these | |
9200 | services are resolved properly. | |
67eb5b38 | 9201 | |
3925496a LP |
9202 | * In /etc/fstab two new mount options are now understood: |
9203 | x-systemd.makefs and x-systemd.growfs. The former has the effect that | |
9204 | the configured file system is formatted before it is mounted, the | |
9205 | latter that the file system is resized to the full block device size | |
9206 | after it is mounted (i.e. if the file system is smaller than the | |
9207 | partition it resides on, it's grown). This is similar to the fsck | |
9208 | logic in /etc/fstab, and pulls in systemd-makefs@.service and | |
9209 | systemd-growfs@.service as necessary, similar to | |
9210 | systemd-fsck@.service. Resizing is currently only supported on ext4 | |
9211 | and btrfs. | |
9212 | ||
67eb5b38 LP |
9213 | * In systemd-networkd, the IPv6 RA logic now optionally may announce |
9214 | DNS server and domain information. | |
9215 | ||
9216 | * Support for the LUKS2 on-disk format for encrypted partitions has | |
9217 | been added. This requires libcryptsetup2 during compilation and | |
9218 | runtime. | |
9219 | ||
89780840 | 9220 | * The systemd --user instance will now signal "readiness" when its |
67eb5b38 LP |
9221 | basic.target unit has been reached, instead of when the run queue ran |
9222 | empty for the first time. | |
9223 | ||
8ea2dcb0 ZJS |
9224 | * Tmpfiles.d with user configuration are now also supported. |
9225 | systemd-tmpfiles gained a new --user switch, and snippets placed in | |
9226 | ~/.config/user-tmpfiles.d/ and corresponding directories will be | |
9227 | executed by systemd-tmpfiles --user running in the new | |
9228 | systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service | |
9229 | running in the user session. | |
9230 | ||
9231 | * Unit files and tmpfiles.d snippets learnt three new % specifiers: | |
9232 | %S resolves to the top-level state directory (/var/lib for the system | |
9233 | instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the | |
9234 | top-level cache directory (/var/cache for the system instance, | |
9235 | $XDG_CACHE_HOME for the user instance), %L resolves to the top-level | |
9236 | logs directory (/var/log for the system instance, | |
67eb5b38 | 9237 | $XDG_CONFIG_HOME/log/ for the user instance). This matches the |
8ea2dcb0 | 9238 | existing %t specifier, that resolves to the top-level runtime |
67eb5b38 LP |
9239 | directory (/run for the system instance, and $XDG_RUNTIME_DIR for the |
9240 | user instance). | |
9241 | ||
9242 | * journalctl learnt a new parameter --output-fields= for limiting the | |
9243 | set of journal fields to output in verbose and JSON output modes. | |
9244 | ||
9245 | * systemd-timesyncd's configuration file gained a new option | |
89780840 ZJS |
9246 | RootDistanceMaxSec= for setting the maximum root distance of servers |
9247 | it'll use, as well as the new options PollIntervalMinSec= and | |
9248 | PollIntervalMaxSec= to tweak the minimum and maximum poll interval. | |
67eb5b38 LP |
9249 | |
9250 | * bootctl gained a new command "list" for listing all available boot | |
89780840 | 9251 | menu items on systems that follow the boot loader specification. |
67eb5b38 LP |
9252 | |
9253 | * systemctl gained a new --dry-run switch that shows what would be done | |
9254 | instead of doing it, and is currently supported by the shutdown and | |
9255 | sleep verbs. | |
9256 | ||
e9ad86d5 | 9257 | * ConditionSecurity= can now detect the TOMOYO security module. |
67eb5b38 LP |
9258 | |
9259 | * Unit file [Install] sections are now also respected in unit drop-in | |
89780840 | 9260 | files. This is intended to be used by drop-ins under /usr/lib/. |
67eb5b38 | 9261 | |
89780840 | 9262 | * systemd-firstboot may now also set the initial keyboard mapping. |
67eb5b38 | 9263 | |
89780840 ZJS |
9264 | * Udev "changed" events for devices which are exposed as systemd |
9265 | .device units are now propagated to units specified in | |
9266 | ReloadPropagatedFrom= as reload requests. | |
67eb5b38 | 9267 | |
89780840 ZJS |
9268 | * If a udev device has a SYSTEMD_WANTS= property containing a systemd |
9269 | unit template name (i.e. a name in the form of 'foobar@.service', | |
9270 | without the instance component between the '@' and - the '.'), then | |
9271 | the escaped sysfs path of the device is automatically used as the | |
9272 | instance. | |
67eb5b38 LP |
9273 | |
9274 | * SystemCallFilter= in unit files has been extended so that an "errno" | |
9275 | can be specified individually for each system call. Example: | |
9276 | SystemCallFilter=~uname:EILSEQ. | |
9277 | ||
9278 | * The cgroup delegation logic has been substantially updated. Delegate= | |
9279 | now optionally takes a list of controllers (instead of a boolean, as | |
9280 | before), which lists the controllers to delegate at least. | |
9281 | ||
89780840 | 9282 | * The networkd DHCPv6 client now implements the FQDN option (RFC 4704). |
67eb5b38 | 9283 | |
89780840 ZJS |
9284 | * A new LogLevelMax= setting configures the maximum log level any |
9285 | process of the service may log at (i.e. anything with a lesser | |
9286 | priority than what is specified is automatically dropped). A new | |
9287 | LogExtraFields= setting allows configuration of additional journal | |
9288 | fields to attach to all log records generated by any of the unit's | |
9289 | processes. | |
67eb5b38 | 9290 | |
89780840 ZJS |
9291 | * New StandardInputData= and StandardInputText= settings along with the |
9292 | new option StandardInput=data may be used to configure textual or | |
9293 | binary data that shall be passed to the executed service process via | |
9294 | standard input, encoded in-line in the unit file. | |
67eb5b38 LP |
9295 | |
9296 | * StandardInput=, StandardOutput= and StandardError= may now be used to | |
9297 | connect stdin/stdout/stderr of executed processes directly with a | |
9298 | file or AF_UNIX socket in the file system, using the new "file:" option. | |
9299 | ||
89780840 ZJS |
9300 | * A new unit file option CollectMode= has been added, that allows |
9301 | tweaking the garbage collection logic for units. It may be used to | |
9302 | tell systemd to garbage collect units that have failed automatically | |
9303 | (normally it only GCs units that exited successfully). systemd-run | |
9304 | and systemd-mount expose this new functionality with a new -G option. | |
9305 | ||
67eb5b38 LP |
9306 | * "machinectl bind" may now be used to bind mount non-directories |
9307 | (i.e. regularfiles, devices, fifos, sockets). | |
9308 | ||
9309 | * systemd-analyze gained a new verb "calendar" for validating and | |
9310 | testing calendar time specifications to use for OnCalendar= in timer | |
9311 | units. Besides validating the expression it will calculate the next | |
9312 | time the specified expression would elapse. | |
9313 | ||
9314 | * In addition to the pre-existing FailureAction= unit file setting | |
89780840 ZJS |
9315 | there's now SuccessAction=, for configuring a shutdown action to |
9316 | execute when a unit completes successfully. This is useful in | |
9317 | particular inside containers that shall terminate after some workload | |
9318 | has been completed. Also, both options are now supported for all unit | |
9319 | types, not just services. | |
67eb5b38 LP |
9320 | |
9321 | * networkds's IP rule support gained two new options | |
dd014eeb | 9322 | IncomingInterface= and OutgoingInterface= for configuring the incoming |
67eb5b38 LP |
9323 | and outgoing interfaces of configured rules. systemd-networkd also |
9324 | gained support for "vxcan" network devices. | |
9325 | ||
9326 | * networkd gained a new setting RequiredForOnline=, taking a | |
9327 | boolean. If set, systemd-wait-online will take it into consideration | |
9328 | when determining that the system is up, otherwise it will ignore the | |
9329 | interface for this purpose. | |
9330 | ||
9331 | * The sd_notify() protocol gained support for a new operation: with | |
9332 | FDSTOREREMOVE=1 file descriptors may be removed from the per-service | |
9333 | store again, ahead of POLLHUP or POLLERR when they are removed | |
9334 | anyway. | |
9335 | ||
f09eb768 LP |
9336 | * A new document doc/UIDS-GIDS.md has been added to the source tree, |
9337 | that documents the UID/GID range and assignment assumptions and | |
3925496a LP |
9338 | requirements of systemd. |
9339 | ||
9340 | * The watchdog device PID 1 will ping may now be configured through the | |
9341 | WatchdogDevice= configuration file setting, or by setting the | |
7c52d523 | 9342 | systemd.watchdog_service= kernel command line option. |
3925496a LP |
9343 | |
9344 | * systemd-resolved's gained support for registering DNS-SD services on | |
9345 | the local network using MulticastDNS. Services may either be | |
9346 | registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or | |
9347 | the same dir below /run, /usr/lib), or through its D-Bus API. | |
9348 | ||
a327431b DB |
9349 | * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond |
9350 | extend the effective start, runtime, and stop time. The service must | |
9351 | continue to send EXTEND_TIMEOUT_USEC within the period specified to | |
9352 | prevent the service manager from making the service as timedout. | |
9353 | ||
ea2a3c9e LP |
9354 | * systemd-resolved's DNSSEC support gained support for RFC 8080 |
9355 | (Ed25519 keys and signatures). | |
9356 | ||
a1b2c92d LP |
9357 | * The systemd-resolve command line tool gained a new set of options |
9358 | --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=, | |
9359 | --set-nta= and --revert to configure per-interface DNS configuration | |
9360 | dynamically during runtime. It's useful for pushing DNS information | |
9361 | into systemd-resolved from DNS hook scripts that various interface | |
9362 | managing software supports (such as pppd). | |
9363 | ||
9364 | * systemd-nspawn gained a new --network-namespace-path= command line | |
9365 | option, which may be used to make a container join an existing | |
9366 | network namespace, by specifying a path to a "netns" file. | |
9367 | ||
3925496a LP |
9368 | Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini, |
9369 | Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten | |
9370 | Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin | |
ea2a3c9e | 9371 | Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri |
a1b2c92d LP |
9372 | John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny |
9373 | Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de | |
9374 | Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty, | |
9375 | Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef | |
9376 | Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars | |
9377 | Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel, | |
9378 | Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz | |
9379 | Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson, | |
9380 | Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala, | |
9381 | Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal | |
9382 | Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf | |
9383 | Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer, | |
9384 | Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran | |
9385 | Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon | |
9386 | Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant | |
9387 | Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem | |
9388 | Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich, | |
9389 | Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew | |
9390 | Jędrzejewski-Szmek, Zeal Jagannatha | |
67eb5b38 | 9391 | |
ea2a3c9e | 9392 | — Berlin, 2017-12-14 |
3754abc5 | 9393 | |
582faeb4 DJL |
9394 | CHANGES WITH 235: |
9395 | ||
2bcbffd6 LP |
9396 | * INCOMPATIBILITY: systemd-logind.service and other long-running |
9397 | services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP | |
9398 | communication with the outside. This generally improves security of | |
9399 | the system, and is in almost all cases a safe and good choice, as | |
23d37367 | 9400 | these services do not and should not provide any network-facing |
2bcbffd6 LP |
9401 | functionality. However, systemd-logind uses the glibc NSS API to |
9402 | query the user database. This creates problems on systems where NSS | |
9403 | is set up to directly consult network services for user database | |
9404 | lookups. In particular, this creates incompatibilities with the | |
9405 | "nss-nis" module, which attempts to directly contact the NIS/YP | |
9406 | network servers it is configured for, and will now consistently | |
9407 | fail. In such cases, it is possible to turn off IP sandboxing for | |
9408 | systemd-logind.service (set IPAddressDeny= in its [Service] section | |
9409 | to the empty string, via a .d/ unit file drop-in). Downstream | |
9410 | distributions might want to update their nss-nis packaging to include | |
9411 | such a drop-in snippet, accordingly, to hide this incompatibility | |
9412 | from the user. Another option is to make use of glibc's nscd service | |
9413 | to proxy such network requests through a privilege-separated, minimal | |
9414 | local caching daemon, or to switch to more modern technologies such | |
9415 | sssd, whose NSS hook-ups generally do not involve direct network | |
9416 | access. In general, we think it's definitely time to question the | |
9417 | implementation choices of nss-nis, i.e. whether it's a good idea | |
9418 | today to embed a network-facing loadable module into all local | |
9419 | processes that need to query the user database, including the most | |
9420 | trivial and benign ones, such as "ls". For more details about | |
9421 | IPAddressDeny= see below. | |
9422 | ||
fccf5419 LP |
9423 | * A new modprobe.d drop-in is now shipped by default that sets the |
9424 | bonding module option max_bonds=0. This overrides the kernel default, | |
9425 | to avoid conflicts and ambiguity as to whether or not bond0 should be | |
9426 | managed by systemd-networkd or not. This resolves multiple issues | |
9427 | with bond0 properties not being applied, when bond0 is configured | |
9428 | with systemd-networkd. Distributors may choose to not package this, | |
9429 | however in that case users will be prevented from correctly managing | |
9430 | bond0 interface using systemd-networkd. | |
582faeb4 | 9431 | |
ef5a8cb1 | 9432 | * systemd-analyze gained new verbs "get-log-level" and "get-log-target" |
21723f53 ZJS |
9433 | which print the logging level and target of the system manager. They |
9434 | complement the existing "set-log-level" and "set-log-target" verbs | |
9435 | used to change those values. | |
ef5a8cb1 | 9436 | |
fccf5419 LP |
9437 | * journald.conf gained a new boolean setting ReadKMsg= which defaults |
9438 | to on. If turned off kernel log messages will not be read by | |
21723f53 ZJS |
9439 | systemd-journald or included in the logs. It also gained a new |
9440 | setting LineMax= for configuring the maximum line length in | |
9441 | STDOUT/STDERR log streams. The new default for this value is 48K, up | |
9442 | from the previous hardcoded 2048. | |
fccf5419 | 9443 | |
21723f53 ZJS |
9444 | * A new unit setting RuntimeDirectoryPreserve= has been added, which |
9445 | allows more detailed control of what to do with a runtime directory | |
9446 | configured with RuntimeDirectory= (i.e. a directory below /run or | |
9447 | $XDG_RUNTIME_DIR) after a unit is stopped. | |
fccf5419 LP |
9448 | |
9449 | * The RuntimeDirectory= setting for units gained support for creating | |
9450 | deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just | |
9451 | one top-level directory. | |
9452 | ||
9453 | * Units gained new options StateDirectory=, CacheDirectory=, | |
9454 | LogsDirectory= and ConfigurationDirectory= which are closely related | |
9455 | to RuntimeDirectory= but manage per-service directories below | |
21723f53 | 9456 | /var/lib, /var/cache, /var/log and /etc. By making use of them it is |
fccf5419 LP |
9457 | possible to write unit files which when activated automatically gain |
9458 | properly owned service specific directories in these locations, thus | |
9459 | making unit files self-contained and increasing compatibility with | |
9460 | stateless systems and factory reset where /etc or /var are | |
9461 | unpopulated at boot. Matching these new settings there's also | |
9462 | StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=, | |
9463 | ConfigurationDirectoryMode= for configuring the access mode of these | |
75dfbbac LP |
9464 | directories. These settings are particularly useful in combination |
9465 | with DynamicUser=yes as they provide secure, properly-owned, | |
9466 | writable, and stateful locations for storage, excluded from the | |
9467 | sandbox that such services live in otherwise. | |
fccf5419 LP |
9468 | |
9469 | * Automake support has been removed from this release. systemd is now | |
9470 | Meson-only. | |
9471 | ||
9472 | * systemd-journald will now aggressively cache client metadata during | |
9473 | runtime, speeding up log write performance under pressure. This comes | |
9474 | at a small price though: as much of the metadata is read | |
9475 | asynchronously from /proc/ (and isn't implicitly attached to log | |
9476 | datagrams by the kernel, like UID/GID/PID/SELinux are) this means the | |
9477 | metadata stored alongside a log entry might be slightly | |
9478 | out-of-date. Previously it could only be slightly newer than the log | |
9479 | message. The time window is small however, and given that the kernel | |
9480 | is unlikely to be improved anytime soon in this regard, this appears | |
9481 | acceptable to us. | |
9482 | ||
9483 | * nss-myhostname/systemd-resolved will now by default synthesize an | |
9484 | A/AAAA resource record for the "_gateway" hostname, pointing to the | |
9485 | current default IP gateway. Previously it did that for the "gateway" | |
9486 | name, hampering adoption, as some distributions wanted to leave that | |
38b38500 | 9487 | hostname open for local use. The old behaviour may still be |
fccf5419 LP |
9488 | requested at build time. |
9489 | ||
9490 | * systemd-networkd's [Address] section in .network files gained a new | |
9491 | Scope= setting for configuring the IP address scope. The [Network] | |
9492 | section gained a new boolean setting ConfigureWithoutCarrier= that | |
9493 | tells systemd-networkd to ignore link sensing when configuring the | |
9494 | device. The [DHCP] section gained a new Anonymize= boolean option for | |
9495 | turning on a number of options suggested in RFC 7844. A new | |
9496 | [RoutingPolicyRule] section has been added for configuring the IP | |
9497 | routing policy. The [Route] section has gained support for a new | |
9498 | Type= setting which permits configuring | |
9499 | blackhole/unreachable/prohibit routes. | |
9500 | ||
9501 | * The [VRF] section in .netdev files gained a new Table= setting for | |
9502 | configuring the routing table to use. The [Tunnel] section gained a | |
9503 | new Independent= boolean field for configuring tunnels independent of | |
9504 | an underlying network interface. The [Bridge] section gained a new | |
9505 | GroupForwardMask= option for configuration of propagation of link | |
9506 | local frames between bridge ports. | |
9507 | ||
9508 | * The WakeOnLan= setting in .link files gained support for a number of | |
9509 | new modes. A new TCP6SegmentationOffload= setting has been added for | |
9510 | configuring TCP/IPv6 hardware segmentation offload. | |
9511 | ||
9512 | * The IPv6 RA sender implementation may now optionally send out RDNSS | |
21723f53 | 9513 | and RDNSSL records to supply DNS configuration to peers. |
fccf5419 LP |
9514 | |
9515 | * systemd-nspawn gained support for a new --system-call-filter= command | |
21723f53 ZJS |
9516 | line option for adding and removing entries in the default system |
9517 | call filter it applies. Moreover systemd-nspawn has been changed to | |
6b000af4 | 9518 | implement a system call allow list instead of a deny list. |
fccf5419 LP |
9519 | |
9520 | * systemd-run gained support for a new --pipe command line option. If | |
9521 | used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run | |
9522 | are directly passed on to the activated transient service | |
21723f53 ZJS |
9523 | executable. This allows invoking arbitrary processes as systemd |
9524 | services (for example to take benefit of dependency management, | |
9525 | accounting management, resource management or log management that is | |
9526 | done automatically for services) — while still allowing them to be | |
fccf5419 LP |
9527 | integrated in a classic UNIX shell pipeline. |
9528 | ||
9529 | * When a service sends RELOAD=1 via sd_notify() and reload propagation | |
9530 | using ReloadPropagationTo= is configured, a reload is now propagated | |
9531 | to configured units. (Previously this was only done on explicitly | |
9532 | requested reloads, using "systemctl reload" or an equivalent | |
9533 | command.) | |
9534 | ||
9535 | * For each service unit a restart counter is now kept: it is increased | |
9536 | each time the service is restarted due to Restart=, and may be | |
9537 | queried using "systemctl show -p NRestarts …". | |
9538 | ||
44898c53 LP |
9539 | * New system call filter groups @aio, @sync, @chown, @setuid, @memlock, |
9540 | @signal and @timer have been added, for usage with SystemCallFilter= | |
fccf5419 LP |
9541 | in unit files and the new --system-call-filter= command line option |
9542 | of systemd-nspawn (see above). | |
9543 | ||
9544 | * ExecStart= lines in unit files gained two new modifiers: when a | |
9545 | command line is prefixed with "!" the command will be executed as | |
9546 | configured, except for the credentials applied by | |
9547 | setuid()/setgid()/setgroups(). It is very similar to the pre-existing | |
9548 | "+", but does still apply namespacing options unlike "+". There's | |
9549 | also "!!" now, which is mostly identical, but becomes a NOP on | |
9550 | systems that support ambient capabilities. This is useful to write | |
9551 | unit files that work with ambient capabilities where possible but | |
9552 | automatically fall back to traditional privilege dropping mechanisms | |
9553 | on systems where this is not supported. | |
9554 | ||
9555 | * ListenNetlink= settings in socket units now support RDMA netlink | |
9556 | sockets. | |
9557 | ||
9558 | * A new unit file setting LockPersonality= has been added which permits | |
9559 | locking down the chosen execution domain ("personality") of a service | |
9560 | during runtime. | |
9561 | ||
9562 | * A new special target "getty-pre.target" has been added, which is | |
9563 | ordered before all text logins, and may be used to order services | |
21723f53 | 9564 | before textual logins acquire access to the console. |
fccf5419 LP |
9565 | |
9566 | * systemd will now attempt to load the virtio-rng.ko kernel module very | |
9567 | early on if a VM environment supporting this is detected. This should | |
9568 | improve entropy during early boot in virtualized environments. | |
9569 | ||
9570 | * A _netdev option is now supported in /etc/crypttab that operates in a | |
9571 | similar way as the same option in /etc/fstab: it permits configuring | |
21723f53 ZJS |
9572 | encrypted devices that need to be ordered after the network is up. |
9573 | Following this logic, two new special targets | |
fccf5419 | 9574 | remote-cryptsetup-pre.target and remote-cryptsetup.target have been |
21723f53 ZJS |
9575 | added that are to cryptsetup.target what remote-fs.target and |
9576 | remote-fs-pre.target are to local-fs.target. | |
fccf5419 LP |
9577 | |
9578 | * Service units gained a new UnsetEnvironment= setting which permits | |
21723f53 ZJS |
9579 | unsetting specific environment variables for services that are |
9580 | normally passed to it (for example in order to mask out locale | |
fccf5419 LP |
9581 | settings for specific services that can't deal with it). |
9582 | ||
9583 | * Units acquired a new boolean option IPAccounting=. When turned on, IP | |
9584 | traffic accounting (packet count as well as byte count) is done for | |
9585 | the service, and shown as part of "systemctl status" or "systemd-run | |
9586 | --wait". | |
9587 | ||
9588 | * Service units acquired two new options IPAddressAllow= and | |
9589 | IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks, | |
9590 | for configuring a simple IP access control list for all sockets of | |
9591 | the unit. These options are available also on .slice and .socket | |
9592 | units, permitting flexible access list configuration for individual | |
9593 | services as well as groups of services (as defined by a slice unit), | |
9594 | including system-wide. Note that IP ACLs configured this way are | |
9595 | enforced on every single IPv4 and IPv6 socket created by any process | |
9596 | of the service unit, and apply to ingress as well as egress traffic. | |
9597 | ||
21723f53 | 9598 | * If CPUAccounting= or IPAccounting= is turned on for a unit a new |
608f70e6 | 9599 | structured log message is generated each time the unit is stopped, |
fccf5419 LP |
9600 | containing information about the consumed resources of this |
9601 | invocation. | |
9602 | ||
9603 | * A new setting KeyringMode= has been added to unit files, which may be | |
9604 | used to control how the kernel keyring is set up for executed | |
9605 | processes. | |
9606 | ||
e06fafb2 LP |
9607 | * "systemctl poweroff", "systemctl reboot", "systemctl halt", |
9608 | "systemctl kexec" and "systemctl exit" are now always asynchronous in | |
9609 | behaviour (that is: these commands return immediately after the | |
21723f53 ZJS |
9610 | operation was enqueued instead of waiting for the operation to |
9611 | complete). Previously, "systemctl poweroff" and "systemctl reboot" | |
e06fafb2 LP |
9612 | were asynchronous on systems using systemd-logind (i.e. almost |
9613 | always, and like they were on sysvinit), and the other three commands | |
9614 | were unconditionally synchronous. With this release this is cleaned | |
9615 | up, and callers will see the same asynchronous behaviour on all | |
9616 | systems for all five operations. | |
9617 | ||
9618 | * systemd-logind gained new Halt() and CanHalt() bus calls for halting | |
9619 | the system. | |
9620 | ||
fccf5419 LP |
9621 | * .timer units now accept calendar specifications in other timezones |
9622 | than UTC or the local timezone. | |
9623 | ||
f6e64b78 | 9624 | * The tmpfiles snippet var.conf has been changed to create |
21723f53 ZJS |
9625 | /var/log/btmp with access mode 0660 instead of 0600. It was owned by |
9626 | the "utmp" group already, and it appears to be generally understood | |
9627 | that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp | |
9628 | databases. Previously this was implemented correctly for all these | |
9629 | databases excepts btmp, which has been opened up like this now | |
9630 | too. Note that while the other databases are world-readable | |
9631 | (i.e. 0644), btmp is not and remains more restrictive. | |
f6e64b78 | 9632 | |
d55b0463 LP |
9633 | * The systemd-resolve tool gained a new --reset-server-features |
9634 | switch. When invoked like this systemd-resolved will forget | |
9635 | everything it learnt about the features supported by the configured | |
9636 | upstream DNS servers, and restarts the feature probing logic on the | |
cf84484a LP |
9637 | next resolver look-up for them at the highest feature level |
9638 | again. | |
9639 | ||
9640 | * The status dump systemd-resolved sends to the logs upon receiving | |
9641 | SIGUSR1 now also includes information about all DNS servers it is | |
9642 | configured to use, and the features levels it probed for them. | |
d55b0463 | 9643 | |
fccf5419 LP |
9644 | Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander |
9645 | Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar | |
76451c1d LP |
9646 | Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles |
9647 | Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel | |
9648 | Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John | |
9649 | Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov, | |
9650 | ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui, | |
9651 | Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov, | |
9652 | Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen, | |
9653 | John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg | |
9654 | Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud, | |
9655 | Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas | |
9656 | Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin | |
9657 | Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert, | |
9658 | Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell | |
9659 | Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo, | |
9660 | Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom | |
9661 | Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid, | |
9662 | Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang | |
9663 | Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek | |
fccf5419 | 9664 | |
c1719d8b | 9665 | — Berlin, 2017-10-06 |
fccf5419 | 9666 | |
4b4da299 LP |
9667 | CHANGES WITH 234: |
9668 | ||
9669 | * Meson is now supported as build system in addition to Automake. It is | |
9670 | our plan to remove Automake in one of our next releases, so that | |
9671 | Meson becomes our exclusive build system. Hence, please start using | |
9672 | the Meson build system in your downstream packaging. There's plenty | |
9673 | of documentation around how to use Meson, the extremely brief | |
9674 | summary: | |
9675 | ||
9676 | ./autogen.sh && ./configure && make && sudo make install | |
9677 | ||
9678 | becomes: | |
9679 | ||
9680 | meson build && ninja -C build && sudo ninja -C build install | |
9681 | ||
9682 | * Unit files gained support for a new JobRunningTimeoutUSec= setting, | |
9683 | which permits configuring a timeout on the time a job is | |
9684 | running. This is particularly useful for setting timeouts on jobs for | |
9685 | .device units. | |
9686 | ||
9687 | * Unit files gained two new options ConditionUser= and ConditionGroup= | |
9688 | for conditionalizing units based on the identity of the user/group | |
9689 | running a systemd user instance. | |
9690 | ||
9691 | * systemd-networkd now understands a new FlowLabel= setting in the | |
9692 | [VXLAN] section of .network files, as well as a Priority= in | |
9693 | [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN] | |
9694 | and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also | |
9695 | gained support for configuration of GENEVE links, and IPv6 address | |
9696 | labels. The [Network] section gained the new IPv6ProxyNDP= setting. | |
9697 | ||
9f09a95a | 9698 | * .link files now understand a new Port= setting. |
4b4da299 LP |
9699 | |
9700 | * systemd-networkd's DHCP support gained support for DHCP option 119 | |
9701 | (domain search list). | |
9702 | ||
9703 | * systemd-networkd gained support for serving IPv6 address ranges using | |
bc99dac5 | 9704 | the Router Advertisement protocol. The new .network configuration |
4b4da299 LP |
9705 | section [IPv6Prefix] may be used to configure the ranges to |
9706 | serve. This is implemented based on a new, minimal, native server | |
9707 | implementation of RA. | |
9708 | ||
9709 | * journalctl's --output= switch gained support for a new parameter | |
9710 | "short-iso-precise" for a mode where timestamps are shown as precise | |
9711 | ISO date values. | |
9712 | ||
9713 | * systemd-udevd's "net_id" builtin may now generate stable network | |
9714 | interface names from IBM PowerVM VIO devices as well as ACPI platform | |
9715 | devices. | |
9716 | ||
9717 | * MulticastDNS support in systemd-resolved may now be explicitly | |
9718 | enabled/disabled using the new MulticastDNS= configuration file | |
9719 | option. | |
9720 | ||
9721 | * systemd-resolved may now optionally use libidn2 instead of the libidn | |
7f7ab228 ZJS |
9722 | for processing internationalized domain names. Support for libidn2 |
9723 | should be considered experimental and should not be enabled by | |
9724 | default yet. | |
4b4da299 LP |
9725 | |
9726 | * "machinectl pull-tar" and related call may now do verification of | |
9727 | downloaded images using SUSE-style .sha256 checksum files in addition | |
9728 | to the already existing support for validating using Ubuntu-style | |
9729 | SHA256SUMS files. | |
9730 | ||
9731 | * sd-bus gained support for a new sd_bus_message_appendv() call which | |
9732 | is va_list equivalent of sd_bus_message_append(). | |
9733 | ||
9734 | * sd-boot gained support for validating images using SHIM/MOK. | |
9735 | ||
9736 | * The SMACK code learnt support for "onlycap". | |
9737 | ||
9738 | * systemd-mount --umount is now much smarter in figuring out how to | |
9739 | properly unmount a device given its mount or device path. | |
5486a31d ZJS |
9740 | |
9741 | * The code to call libnss_dns as a fallback from libnss_resolve when | |
9742 | the communication with systemd-resolved fails was removed. This | |
9743 | fallback was redundant and interfered with the [!UNAVAIL=return] | |
9744 | suffix. See nss-resolve(8) for the recommended configuration. | |
9745 | ||
9f09a95a ZJS |
9746 | * systemd-logind may now be restarted without losing state. It stores |
9747 | the file descriptors for devices it manages in the system manager | |
38d93385 | 9748 | using the FDSTORE= mechanism. Please note that further changes in |
9f09a95a ZJS |
9749 | other components may be required to make use of this (for example |
9750 | Xorg has code to listen for stops of systemd-logind and terminate | |
9751 | itself when logind is stopped or restarted, in order to avoid using | |
9752 | stale file descriptors for graphical devices, which is now | |
9753 | counterproductive and must be reverted in order for restarts of | |
9754 | systemd-logind to be safe. See | |
9755 | https://cgit.freedesktop.org/xorg/xserver/commit/?id=dc48bd653c7e101.) | |
9756 | ||
d271c5d3 | 9757 | * All kernel-install plugins are called with the environment variable |
9d8813b3 | 9758 | KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by |
d271c5d3 ZJS |
9759 | /etc/machine-id. If the machine ID could not be determined, |
9760 | $KERNEL_INSTALL_MACHINE_ID will be empty. Plugins should not put | |
9761 | anything in the entry directory (passed as the second argument) if | |
5238e957 | 9762 | $KERNEL_INSTALL_MACHINE_ID is empty. For backwards compatibility, a |
d271c5d3 ZJS |
9763 | temporary directory is passed as the entry directory and removed |
9764 | after all the plugins exit. | |
9d8813b3 | 9765 | |
357376d0 DDM |
9766 | * If KERNEL_INSTALL_MACHINE_ID is set in /etc/machine-info, kernel-install |
9767 | will now use its value as the machine ID instead of the machine ID | |
9768 | from /etc/machine-id. If KERNEL_INSTALL_MACHINE_ID isn't set in | |
9769 | /etc/machine-info and no machine ID is set in /etc/machine-id, | |
9770 | kernel-install will try to store the current machine ID there as | |
9771 | KERNEL_INSTALL_MACHINE_ID. If there is no machine ID, kernel-install | |
9772 | will generate a new UUID, store it in /etc/machine-info as | |
9773 | KERNEL_INSTALL_MACHINE_ID and use it as the machine ID. | |
9774 | ||
184d2c15 | 9775 | Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander |
ac172e52 LP |
9776 | Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir |
9777 | Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert, | |
9778 | Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb, | |
9779 | Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake, | |
184d2c15 LP |
9780 | Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide |
9781 | Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John | |
9782 | Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin, | |
9783 | Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary | |
ac172e52 LP |
9784 | Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede, |
9785 | hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan | |
9786 | Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason | |
9787 | Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg | |
9788 | Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow, | |
9789 | Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili, | |
9790 | Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas, | |
9791 | Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala, | |
9792 | Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin, | |
9793 | Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal | |
9794 | Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis, | |
9795 | Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik | |
9796 | Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr | |
9797 | Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes, | |
9798 | Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan | |
9799 | Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas | |
184d2c15 LP |
9800 | H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom |
9801 | Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog, | |
ac172e52 LP |
9802 | userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu, |
9803 | Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан | |
9804 | Георгиевски | |
4b4da299 | 9805 | |
ac172e52 | 9806 | — Berlin, 2017-07-12 |
4b4da299 | 9807 | |
a2b53448 | 9808 | CHANGES WITH 233: |
d08ee7cb | 9809 | |
23eb30b3 ZJS |
9810 | * The "hybrid" control group mode has been modified to improve |
9811 | compatibility with "legacy" cgroups-v1 setups. Specifically, the | |
9812 | "hybrid" setup of /sys/fs/cgroup is now pretty much identical to | |
9813 | "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named | |
9814 | cgroups-v1 hierarchy), the only externally visible change being that | |
9815 | the cgroups-v2 hierarchy is also mounted, to | |
9816 | /sys/fs/cgroup/unified. This should provide a large degree of | |
9817 | compatibility with "legacy" cgroups-v1, while taking benefit of the | |
9818 | better management capabilities of cgroups-v2. | |
9819 | ||
9820 | * The default control group setup mode may be selected both a boot-time | |
9821 | via a set of kernel command line parameters (specifically: | |
9822 | systemd.unified_cgroup_hierarchy= and | |
9823 | systemd.legacy_systemd_cgroup_controller=), as well as a compile-time | |
9824 | default selected on the configure command line | |
9825 | (--with-default-hierarchy=). The upstream default is "hybrid" | |
9826 | (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but | |
9827 | this will change in a future systemd version to be "unified" (pure | |
9828 | cgroups-v2 mode). The third option for the compile time option is | |
9829 | "legacy", to enter pure cgroups-v1 mode. We recommend downstream | |
9830 | distributions to default to "hybrid" mode for release distributions, | |
9831 | starting with v233. We recommend "unified" for development | |
9832 | distributions (specifically: distributions such as Fedora's rawhide) | |
9833 | as that's where things are headed in the long run. Use "legacy" for | |
9834 | greatest stability and compatibility only. | |
9835 | ||
9836 | * Note one current limitation of "unified" and "hybrid" control group | |
9837 | setup modes: the kernel currently does not permit the systemd --user | |
9838 | instance (i.e. unprivileged code) to migrate processes between two | |
9839 | disconnected cgroup subtrees, even if both are managed and owned by | |
9840 | the user. This effectively means "systemd-run --user --scope" doesn't | |
9841 | work when invoked from outside of any "systemd --user" service or | |
9842 | scope. Specifically, it is not supported from session scopes. We are | |
9843 | working on fixing this in a future systemd version. (See #3388 for | |
9844 | further details about this.) | |
9845 | ||
fb7c4eff MG |
9846 | * DBus policy files are now installed into /usr rather than /etc. Make |
9847 | sure your system has dbus >= 1.9.18 running before upgrading to this | |
9848 | version, or override the install path with --with-dbuspolicydir= . | |
9849 | ||
23eb30b3 ZJS |
9850 | * All python scripts shipped with systemd (specifically: the various |
9851 | tests written in Python) now require Python 3. | |
9852 | ||
d60c5270 | 9853 | * systemd unit tests can now run standalone (without the source or |
4dfe64f8 ZJS |
9854 | build directories), and can be installed into /usr/lib/systemd/tests/ |
9855 | with 'make install-tests'. | |
9856 | ||
23eb30b3 ZJS |
9857 | * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH, |
9858 | CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the | |
9859 | kernel. | |
9860 | ||
9861 | * Support for the %c, %r, %R specifiers in unit files has been | |
9862 | removed. Specifiers are not supposed to be dependent on configuration | |
9863 | in the unit file itself (so that they resolve the same regardless | |
9864 | where used in the unit files), but these specifiers were influenced | |
9865 | by the Slice= option. | |
9866 | ||
5cfc0a84 LP |
9867 | * The shell invoked by debug-shell.service now defaults to /bin/sh in |
9868 | all cases. If distributions want to use a different shell for this | |
9869 | purpose (for example Fedora's /sbin/sushell) they need to specify | |
9870 | this explicitly at configure time using --with-debug-shell=. | |
9871 | ||
2bcc3309 FB |
9872 | * The confirmation spawn prompt has been reworked to offer the |
9873 | following choices: | |
9874 | ||
b0eb2944 | 9875 | (c)ontinue, proceed without asking anymore |
dd6f9ac0 | 9876 | (D)ump, show the state of the unit |
2bcc3309 | 9877 | (f)ail, don't execute the command and pretend it failed |
d172b175 | 9878 | (h)elp |
eedf223a | 9879 | (i)nfo, show a short summary of the unit |
56fde33a | 9880 | (j)obs, show jobs that are in progress |
2bcc3309 FB |
9881 | (s)kip, don't execute the command and pretend it succeeded |
9882 | (y)es, execute the command | |
9883 | ||
9884 | The 'n' choice for the confirmation spawn prompt has been removed, | |
9885 | because its meaning was confusing. | |
9886 | ||
d08ee7cb LP |
9887 | The prompt may now also be redirected to an alternative console by |
9888 | specifying the console as parameter to systemd.confirm_spawn=. | |
9889 | ||
8e458bfe JW |
9890 | * Services of Type=notify require a READY=1 notification to be sent |
9891 | during startup. If no such message is sent, the service now fails, | |
9892 | even if the main process exited with a successful exit code. | |
9893 | ||
85266f9b LP |
9894 | * Services that fail to start up correctly now always have their |
9895 | ExecStopPost= commands executed. Previously, they'd enter "failed" | |
9896 | state directly, without executing these commands. | |
9897 | ||
baf32786 MP |
9898 | * The option MulticastDNS= of network configuration files has acquired |
9899 | an actual implementation. With MulticastDNS=yes a host can resolve | |
23eb30b3 | 9900 | names of remote hosts and reply to mDNS A and AAAA requests. |
fa8b4499 | 9901 | |
631b676b LP |
9902 | * When units are about to be started an additional check is now done to |
9903 | ensure that all dependencies of type BindsTo= (when used in | |
9904 | combination with After=) have been started. | |
9905 | ||
d08ee7cb LP |
9906 | * systemd-analyze gained a new verb "syscall-filter" which shows which |
9907 | system call groups are defined for the SystemCallFilter= unit file | |
23eb30b3 | 9908 | setting, and which system calls they contain. |
d08ee7cb LP |
9909 | |
9910 | * A new system call filter group "@filesystem" has been added, | |
23eb30b3 | 9911 | consisting of various file system related system calls. Group |
d08ee7cb | 9912 | "@reboot" has been added, covering reboot, kexec and shutdown related |
23eb30b3 | 9913 | calls. Finally, group "@swap" has been added covering swap |
d08ee7cb LP |
9914 | configuration related calls. |
9915 | ||
9916 | * A new unit file option RestrictNamespaces= has been added that may be | |
9917 | used to restrict access to the various process namespace types the | |
9918 | Linux kernel provides. Specifically, it may be used to take away the | |
23eb30b3 ZJS |
9919 | right for a service unit to create additional file system, network, |
9920 | user, and other namespaces. This sandboxing option is particularly | |
9921 | relevant due to the high amount of recently discovered namespacing | |
9922 | related vulnerabilities in the kernel. | |
d08ee7cb | 9923 | |
23eb30b3 ZJS |
9924 | * systemd-udev's .link files gained support for a new AutoNegotiation= |
9925 | setting for configuring Ethernet auto-negotiation. | |
d08ee7cb LP |
9926 | |
9927 | * systemd-networkd's .network files gained support for a new | |
9928 | ListenPort= setting in the [DHCP] section to explicitly configure the | |
9929 | UDP client port the DHCP client shall listen on. | |
9930 | ||
23eb30b3 ZJS |
9931 | * .network files gained a new Unmanaged= boolean setting for explicitly |
9932 | excluding one or more interfaces from management by systemd-networkd. | |
9933 | ||
9934 | * The systemd-networkd ProxyARP= option has been renamed to | |
9935 | IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been | |
9936 | renamed to ReduceARPProxy=. The old names continue to be available | |
9937 | for compatibility. | |
9938 | ||
9939 | * systemd-networkd gained support for configuring IPv6 Proxy NDP | |
9940 | addresses via the new IPv6ProxyNDPAddress= .network file setting. | |
9941 | ||
9942 | * systemd-networkd's bonding device support gained support for two new | |
9943 | configuration options ActiveSlave= and PrimarySlave=. | |
9944 | ||
9945 | * The various options in the [Match] section of .network files gained | |
9946 | support for negative matching. | |
9947 | ||
d08ee7cb LP |
9948 | * New systemd-specific mount options are now understood in /etc/fstab: |
9949 | ||
9950 | x-systemd.mount-timeout= may be used to configure the maximum | |
9951 | permitted runtime of the mount command. | |
9952 | ||
9953 | x-systemd.device-bound may be set to bind a mount point to its | |
9954 | backing device unit, in order to automatically remove a mount point | |
9955 | if its backing device is unplugged. This option may also be | |
9956 | configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property | |
9957 | on the block device, which is now automatically set for all CDROM | |
9958 | drives, so that mounted CDs are automatically unmounted when they are | |
9959 | removed from the drive. | |
9960 | ||
23eb30b3 ZJS |
9961 | x-systemd.after= and x-systemd.before= may be used to explicitly |
9962 | order a mount after or before another unit or mount point. | |
d08ee7cb LP |
9963 | |
9964 | * Enqueued start jobs for device units are now automatically garbage | |
9965 | collected if there are no jobs waiting for them anymore. | |
9966 | ||
23eb30b3 ZJS |
9967 | * systemctl list-jobs gained two new switches: with --after, for every |
9968 | queued job the jobs it's waiting for are shown; with --before the | |
9969 | jobs which it's blocking are shown. | |
d08ee7cb LP |
9970 | |
9971 | * systemd-nspawn gained support for ephemeral boots from disk images | |
9972 | (or in other words: --ephemeral and --image= may now be | |
9973 | combined). Moreover, ephemeral boots are now supported for normal | |
9974 | directories, even if the backing file system is not btrfs. Of course, | |
baf32786 MP |
9975 | if the file system does not support file system snapshots or |
9976 | reflinks, the initial copy operation will be relatively expensive, but | |
9977 | this should still be suitable for many use cases. | |
d08ee7cb LP |
9978 | |
9979 | * Calendar time specifications in .timer units now support | |
9980 | specifications relative to the end of a month by using "~" instead of | |
9981 | "-" as separator between month and day. For example, "*-02~03" means | |
23eb30b3 | 9982 | "the third last day in February". In addition a new syntax for |
d08ee7cb LP |
9983 | repeated events has been added using the "/" character. For example, |
9984 | "9..17/2:00" means "every two hours from 9am to 5pm". | |
9985 | ||
9986 | * systemd-socket-proxyd gained a new parameter --connections-max= for | |
9987 | configuring the maximum number of concurrent connections. | |
9988 | ||
23eb30b3 ZJS |
9989 | * sd-id128 gained a new API for generating unique IDs for the host in a |
9990 | way that does not leak the machine ID. Specifically, | |
d08ee7cb | 9991 | sd_id128_get_machine_app_specific() derives an ID based on the |
387f6955 | 9992 | machine ID in a well-defined, non-reversible, stable way. This is |
d08ee7cb LP |
9993 | useful whenever an identifier for the host is needed but where the |
9994 | identifier shall not be useful to identify the system beyond the | |
9995 | scope of the application itself. (Internally this uses HMAC-SHA256 as | |
9996 | keyed hash function using the machine ID as input.) | |
9997 | ||
9998 | * NotifyAccess= gained a new supported value "exec". When set | |
9999 | notifications are accepted from all processes systemd itself invoked, | |
10000 | including all control processes. | |
10001 | ||
10002 | * .nspawn files gained support for defining overlay mounts using the | |
10003 | Overlay= and OverlayReadOnly= options. Previously this functionality | |
10004 | was only available on the systemd-nspawn command line. | |
10005 | ||
10006 | * systemd-nspawn's --bind= and --overlay= options gained support for | |
10007 | bind/overlay mounts whose source lies within the container tree by | |
10008 | prefixing the source path with "+". | |
10009 | ||
10010 | * systemd-nspawn's --bind= and --overlay= options gained support for | |
10011 | automatically allocating a temporary source directory in /var/tmp | |
10012 | that is removed when the container dies. Specifically, if the source | |
10013 | directory is specified as empty string this mechanism is selected. An | |
10014 | example usage is --overlay=+/var::/var, which creates an overlay | |
86b52a39 | 10015 | mount based on the original /var contained in the image, overlaid |
d08ee7cb LP |
10016 | with a temporary directory in the host's /var/tmp. This way changes |
10017 | to /var are automatically flushed when the container shuts down. | |
10018 | ||
baf32786 MP |
10019 | * systemd-nspawn --image= option does now permit raw file system block |
10020 | devices (in addition to images containing partition tables, as | |
10021 | before). | |
d08ee7cb LP |
10022 | |
10023 | * The disk image dissection logic in systemd-nspawn gained support for | |
10024 | automatically setting up LUKS encrypted as well as Verity protected | |
10025 | partitions. When a container is booted from an encrypted image the | |
10026 | passphrase is queried at start-up time. When a container with Verity | |
10027 | data is started, the root hash is search in a ".roothash" file | |
10028 | accompanying the disk image (alternatively, pass the root hash via | |
10029 | the new --root-hash= command line option). | |
10030 | ||
10031 | * A new tool /usr/lib/systemd/systemd-dissect has been added that may | |
10032 | be used to dissect disk images the same way as systemd-nspawn does | |
10033 | it, following the Bootable Partition Specification. It may even be | |
10034 | used to mount disk images with complex partition setups (including | |
10035 | LUKS and Verity partitions) to a local host directory, in order to | |
10036 | inspect them. This tool is not considered public API (yet), and is | |
10037 | thus not installed into /usr/bin. Please do not rely on its | |
3b31c466 | 10038 | existence, since it might go away or be changed in later systemd |
d08ee7cb LP |
10039 | versions. |
10040 | ||
10041 | * A new generator "systemd-verity-generator" has been added, similar in | |
baf32786 | 10042 | style to "systemd-cryptsetup-generator", permitting automatic setup of |
d08ee7cb LP |
10043 | Verity root partitions when systemd boots up. In order to make use of |
10044 | this your partition setup should follow the Discoverable Partitions | |
10045 | Specification, and the GPT partition ID of the root file system | |
da890466 | 10046 | partition should be identical to the upper 128-bit of the Verity root |
d08ee7cb | 10047 | hash. The GPT partition ID of the Verity partition protecting it |
da890466 | 10048 | should be the lower 128-bit of the Verity root hash. If the partition |
d08ee7cb LP |
10049 | image follows this model it is sufficient to specify a single |
10050 | "roothash=" kernel command line argument to both configure which root | |
10051 | image and verity partition to use as well as the root hash for | |
10052 | it. Note that systemd-nspawn's Verity support follows the same | |
10053 | semantics, meaning that disk images with proper Verity data in place | |
10054 | may be booted in containers with systemd-nspawn as well as on | |
10055 | physical systems via the verity generator. Also note that the "mkosi" | |
10056 | tool available at https://github.com/systemd/mkosi has been updated | |
10057 | to generate Verity protected disk images following this scheme. In | |
10058 | fact, it has been updated to generate disk images that optionally | |
10059 | implement a complete UEFI SecureBoot trust chain, involving a signed | |
10060 | kernel and initrd image that incorporates such a root hash as well as | |
10061 | a Verity-enabled root partition. | |
10062 | ||
d08ee7cb LP |
10063 | * The hardware database (hwdb) udev supports has been updated to carry |
10064 | accelerometer quirks. | |
10065 | ||
10066 | * All system services are now run with a fresh kernel keyring set up | |
10067 | for them. The invocation ID is stored by default in it, thus | |
10068 | providing a safe, non-overridable way to determine the invocation | |
10069 | ID of each service. | |
10070 | ||
10071 | * Service unit files gained new BindPaths= and BindReadOnlyPaths= | |
10072 | options for bind mounting arbitrary paths in a service-specific | |
10073 | way. When these options are used, arbitrary host or service files and | |
10074 | directories may be mounted to arbitrary locations in the service's | |
10075 | view. | |
10076 | ||
10077 | * Documentation has been added that lists all of systemd's low-level | |
10078 | environment variables: | |
10079 | ||
a8a27374 | 10080 | https://github.com/systemd/systemd/blob/master/docs/ENVIRONMENT.md |
d08ee7cb LP |
10081 | |
10082 | * sd-daemon gained a new API sd_is_socket_sockaddr() for determining | |
10083 | whether a specific socket file descriptor matches a specified socket | |
10084 | address. | |
10085 | ||
10086 | * systemd-firstboot has been updated to check for the | |
10087 | systemd.firstboot= kernel command line option. It accepts a boolean | |
10088 | and when set to false the first boot questions are skipped. | |
10089 | ||
d08ee7cb | 10090 | * systemd-fstab-generator has been updated to check for the |
23eb30b3 ZJS |
10091 | systemd.volatile= kernel command line option, which either takes an |
10092 | optional boolean parameter or the special value "state". If used the | |
10093 | system may be booted in a "volatile" boot mode. Specifically, | |
10094 | "systemd.volatile" is used, the root directory will be mounted as | |
d08ee7cb | 10095 | tmpfs, and only /usr is mounted from the actual root file system. If |
23eb30b3 ZJS |
10096 | "systemd.volatile=state" is used, the root directory will be mounted |
10097 | as usual, but /var is mounted as tmpfs. This concept provides similar | |
d08ee7cb LP |
10098 | functionality as systemd-nspawn's --volatile= option, but provides it |
10099 | on physical boots. Use this option for implementing stateless | |
10100 | systems, or testing systems with all state and/or configuration reset | |
10101 | to the defaults. (Note though that many distributions are not | |
23eb30b3 | 10102 | prepared to boot up without a populated /etc or /var, though.) |
d08ee7cb LP |
10103 | |
10104 | * systemd-gpt-auto-generator gained support for LUKS encrypted root | |
10105 | partitions. Previously it only supported LUKS encrypted partitions | |
10106 | for all other uses, except for the root partition itself. | |
10107 | ||
10108 | * Socket units gained support for listening on AF_VSOCK sockets for | |
10109 | communication in virtualized QEMU environments. | |
10110 | ||
10111 | * The "configure" script gained a new option --with-fallback-hostname= | |
10112 | for specifying the fallback hostname to use if none is configured in | |
10113 | /etc/hostname. For example, by specifying | |
10114 | --with-fallback-hostname=fedora it is possible to default to a | |
23eb30b3 | 10115 | hostname of "fedora" on pristine installations. |
d08ee7cb LP |
10116 | |
10117 | * systemd-cgls gained support for a new --unit= switch for listing only | |
10118 | the control groups of a specific unit. Similar --user-unit= has been | |
10119 | added for listing only the control groups of a specific user unit. | |
10120 | ||
10121 | * systemd-mount gained a new --umount switch for unmounting a mount or | |
10122 | automount point (and all mount/automount points below it). | |
10123 | ||
10124 | * systemd will now refuse full configuration reloads (via systemctl | |
10125 | daemon-reload and related calls) unless at least 16MiB of free space | |
10126 | are available in /run. This is a safety precaution in order to ensure | |
10127 | that generators can safely operate after the reload completed. | |
10128 | ||
10129 | * A new unit file option RootImage= has been added, which has a similar | |
10130 | effect as RootDirectory= but mounts the service's root directory from | |
10131 | a disk image instead of plain directory. This logic reuses the same | |
10132 | image dissection and mount logic that systemd-nspawn already uses, | |
10133 | and hence supports any disk images systemd-nspawn supports, including | |
10134 | those following the Discoverable Partition Specification, as well as | |
10135 | Verity enabled images. This option enables systemd to run system | |
10136 | services directly off disk images acting as resource bundles, | |
10137 | possibly even including full integrity data. | |
10138 | ||
10139 | * A new MountAPIVFS= unit file option has been added, taking a boolean | |
baf32786 | 10140 | argument. If enabled /proc, /sys and /dev (collectively called the |
d08ee7cb LP |
10141 | "API VFS") will be mounted for the service. This is only relevant if |
10142 | RootDirectory= or RootImage= is used for the service, as these mounts | |
10143 | are of course in place in the host mount namespace anyway. | |
10144 | ||
10145 | * systemd-nspawn gained support for a new --pivot-root= switch. If | |
10146 | specified the root directory within the container image is pivoted to | |
10147 | the specified mount point, while the original root disk is moved to a | |
10148 | different place. This option enables booting of ostree images | |
10149 | directly with systemd-nspawn. | |
10150 | ||
d08ee7cb | 10151 | * The systemd build scripts will no longer complain if the NTP server |
23eb30b3 | 10152 | addresses are not changed from the defaults. Google now supports |
d08ee7cb LP |
10153 | these NTP servers officially. We still recommend downstreams to |
10154 | properly register an NTP pool with the NTP pool project though. | |
10155 | ||
c1ec34d1 | 10156 | * coredumpctl gained a new "--reverse" option for printing the list |
d08ee7cb LP |
10157 | of coredumps in reverse order. |
10158 | ||
23eb30b3 ZJS |
10159 | * coredumpctl will now show additional information about truncated and |
10160 | inaccessible coredumps, as well as coredumps that are still being | |
10161 | processed. It also gained a new --quiet switch for suppressing | |
10162 | additional informational message in its output. | |
10163 | ||
10164 | * coredumpctl gained support for only showing coredumps newer and/or | |
10165 | older than specific timestamps, using the new --since= and --until= | |
10166 | options, reminiscent of journalctl's options by the same name. | |
10167 | ||
d08ee7cb | 10168 | * The systemd-coredump logic has been improved so that it may be reused |
23eb30b3 | 10169 | to collect backtraces in non-compiled languages, for example in |
d08ee7cb LP |
10170 | scripting languages such as Python. |
10171 | ||
10172 | * machinectl will now show the UID shift of local containers, if user | |
10173 | namespacing is enabled for them. | |
10174 | ||
baf32786 | 10175 | * systemd will now optionally run "environment generator" binaries at |
d08ee7cb LP |
10176 | configuration load time. They may be used to add environment |
10177 | variables to the environment block passed to services invoked. One | |
baf32786 | 10178 | user environment generator is shipped by default that sets up |
23eb30b3 ZJS |
10179 | environment variables based on files dropped into /etc/environment.d |
10180 | and ~/.config/environment.d/. | |
d08ee7cb | 10181 | |
a2b53448 LP |
10182 | * systemd-resolved now includes the new, recently published 2017 DNSSEC |
10183 | root key (KSK). | |
10184 | ||
a2b53448 LP |
10185 | * hostnamed has been updated to report a new chassis type of |
10186 | "convertible" to cover "foldable" laptops that can both act as a | |
10187 | tablet and as a laptop, such as various Lenovo Yoga devices. | |
10188 | ||
d08ee7cb LP |
10189 | Contributions from: Adrián López, Alexander Galanin, Alexander |
10190 | Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch | |
10191 | Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric | |
10192 | Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri, | |
10193 | Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner, | |
10194 | David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry | |
10195 | Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly, | |
10196 | Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn | |
10197 | Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter, | |
a2b53448 LP |
10198 | Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede, |
10199 | Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan | |
10200 | Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen, | |
10201 | Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart | |
10202 | Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de | |
10203 | Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry, | |
10204 | Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de | |
10205 | Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal | |
10206 | Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak, | |
10207 | Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip | |
10208 | Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin | |
10209 | Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx, | |
10210 | Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi, | |
10211 | Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève, | |
10212 | Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor | |
10213 | Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar | |
10214 | Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb, | |
10215 | Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun, | |
10216 | YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр | |
10217 | Тихонов | |
10218 | ||
10219 | — Berlin, 2017-03-01 | |
d08ee7cb | 10220 | |
54b24597 | 10221 | CHANGES WITH 232: |
76153ad4 | 10222 | |
05f426d2 LP |
10223 | * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and |
10224 | RestrictAddressFamilies= enabled. These sandboxing options should | |
10225 | generally be compatible with the various external udev call-out | |
10226 | binaries we are aware of, however there may be exceptions, in | |
10227 | particular when exotic languages for these call-outs are used. In | |
10228 | this case, consider turning off these settings locally. | |
10229 | ||
4ffe2479 ZJS |
10230 | * The new RemoveIPC= option can be used to remove IPC objects owned by |
10231 | the user or group of a service when that service exits. | |
10232 | ||
6fa44114 | 10233 | * The new ProtectKernelModules= option can be used to disable explicit |
4c37970d LP |
10234 | load and unload operations of kernel modules by a service. In |
10235 | addition access to /usr/lib/modules is removed if this option is set. | |
6fa44114 | 10236 | |
4a77c53d ZJS |
10237 | * ProtectSystem= option gained a new value "strict", which causes the |
10238 | whole file system tree with the exception of /dev, /proc, and /sys, | |
10239 | to be remounted read-only for a service. | |
10240 | ||
e49e2c25 | 10241 | * The new ProtectKernelTunables= option can be used to disable |
4a77c53d ZJS |
10242 | modification of configuration files in /sys and /proc by a service. |
10243 | Various directories and files are remounted read-only, so access is | |
10244 | restricted even if the file permissions would allow it. | |
10245 | ||
6fa44114 | 10246 | * The new ProtectControlGroups= option can be used to disable write |
4a77c53d ZJS |
10247 | access by a service to /sys/fs/cgroup. |
10248 | ||
10249 | * Various systemd services have been hardened with | |
10250 | ProtectKernelTunables=yes, ProtectControlGroups=yes, | |
10251 | RestrictAddressFamilies=. | |
10252 | ||
4ffe2479 ZJS |
10253 | * Support for dynamically creating users for the lifetime of a service |
10254 | has been added. If DynamicUser=yes is specified, user and group IDs | |
1d3a473b | 10255 | will be allocated from the range 61184…65519 for the lifetime of the |
4ffe2479 ZJS |
10256 | service. They can be resolved using the new nss-systemd.so NSS |
10257 | module. The module must be enabled in /etc/nsswitch.conf. Services | |
10258 | started in this way have PrivateTmp= and RemoveIPC= enabled, so that | |
10259 | any resources allocated by the service will be cleaned up when the | |
4a77c53d ZJS |
10260 | service exits. They also have ProtectHome=read-only and |
10261 | ProtectSystem=strict enabled, so they are not able to make any | |
10262 | permanent modifications to the system. | |
4ffe2479 | 10263 | |
171ae2cd | 10264 | * The nss-systemd module also always resolves root and nobody, making |
4ffe2479 | 10265 | it possible to have no /etc/passwd or /etc/group files in minimal |
171ae2cd | 10266 | container or chroot environments. |
4ffe2479 ZJS |
10267 | |
10268 | * Services may be started with their own user namespace using the new | |
171ae2cd LP |
10269 | boolean PrivateUsers= option. Only root, nobody, and the uid/gid |
10270 | under which the service is running are mapped. All other users are | |
10271 | mapped to nobody. | |
4ffe2479 ZJS |
10272 | |
10273 | * Support for the cgroup namespace has been added to systemd-nspawn. If | |
10274 | supported by kernel, the container system started by systemd-nspawn | |
10275 | will have its own view of the cgroup hierarchy. This new behaviour | |
10276 | can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable. | |
10277 | ||
10278 | * The new MemorySwapMax= option can be used to limit the maximum swap | |
10279 | usage under the unified cgroup hierarchy. | |
10280 | ||
10281 | * Support for the CPU controller in the unified cgroup hierarchy has | |
10282 | been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting= | |
10283 | options. This controller requires out-of-tree patches for the kernel | |
10284 | and the support is provisional. | |
10285 | ||
171ae2cd LP |
10286 | * Mount and automount units may now be created transiently |
10287 | (i.e. dynamically at runtime via the bus API, instead of requiring | |
10288 | unit files in the file system). | |
10289 | ||
10290 | * systemd-mount is a new tool which may mount file systems – much like | |
10291 | mount(8), optionally pulling in additional dependencies through | |
10292 | transient .mount and .automount units. For example, this tool | |
10293 | automatically runs fsck on a backing block device before mounting, | |
10294 | and allows the automount logic to be used dynamically from the | |
10295 | command line for establishing mount points. This tool is particularly | |
10296 | useful when dealing with removable media, as it will ensure fsck is | |
10297 | run – if necessary – before the first access and that the file system | |
10298 | is quickly unmounted after each access by utilizing the automount | |
10299 | logic. This maximizes the chance that the file system on the | |
10300 | removable media stays in a clean state, and if it isn't in a clean | |
10301 | state is fixed automatically. | |
4ffe2479 ZJS |
10302 | |
10303 | * LazyUnmount=yes option for mount units has been added to expose the | |
10304 | umount --lazy option. Similarly, ForceUnmount=yes exposes the --force | |
10305 | option. | |
10306 | ||
10307 | * /efi will be used as the mount point of the EFI boot partition, if | |
10308 | the directory is present, and the mount point was not configured | |
10309 | through other means (e.g. fstab). If /efi directory does not exist, | |
10310 | /boot will be used as before. This makes it easier to automatically | |
10311 | mount the EFI partition on systems where /boot is used for something | |
10312 | else. | |
10313 | ||
171ae2cd LP |
10314 | * When operating on GPT disk images for containers, systemd-nspawn will |
10315 | now mount the ESP to /boot or /efi according to the same rules as PID | |
10316 | 1 running on a host. This allows tools like "bootctl" to operate | |
10317 | correctly within such containers, in order to make container images | |
10318 | bootable on physical systems. | |
10319 | ||
4a77c53d | 10320 | * disk/by-id and disk/by-path symlinks are now created for NVMe drives. |
4ffe2479 ZJS |
10321 | |
10322 | * Two new user session targets have been added to support running | |
10323 | graphical sessions under the systemd --user instance: | |
10324 | graphical-session.target and graphical-session-pre.target. See | |
10325 | systemd.special(7) for a description of how those targets should be | |
10326 | used. | |
10327 | ||
10328 | * The vconsole initialization code has been significantly reworked to | |
d4c08299 | 10329 | use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better |
4ffe2479 ZJS |
10330 | support unicode keymaps. Font and keymap configuration will now be |
10331 | copied to all allocated virtual consoles. | |
10332 | ||
05ecf467 | 10333 | * FreeBSD's bhyve virtualization is now detected. |
4ffe2479 | 10334 | |
d4c08299 | 10335 | * Information recorded in the journal for core dumps now includes the |
4ffe2479 ZJS |
10336 | contents of /proc/mountinfo and the command line of the process at |
10337 | the top of the process hierarchy (which is usually the init process | |
10338 | of the container). | |
10339 | ||
171ae2cd | 10340 | * systemd-journal-gatewayd learned the --directory= option to serve |
4ffe2479 ZJS |
10341 | files from the specified location. |
10342 | ||
10343 | * journalctl --root=… can be used to peruse the journal in the | |
10344 | /var/log/ directories inside of a container tree. This is similar to | |
10345 | the existing --machine= option, but does not require the container to | |
10346 | be active. | |
10347 | ||
10348 | * The hardware database has been extended to support | |
10349 | ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify | |
10350 | trackball devices. | |
10351 | ||
10352 | MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to | |
10353 | specify the click rate for mice which include a horizontal wheel with | |
10354 | a click rate that is different than the one for the vertical wheel. | |
10355 | ||
10356 | * systemd-run gained a new --wait option that makes service execution | |
171ae2cd LP |
10357 | synchronous. (Specifically, the command will not return until the |
10358 | specified service binary exited.) | |
4ffe2479 | 10359 | |
171ae2cd | 10360 | * systemctl gained a new --wait option that causes the start command to |
4a77c53d ZJS |
10361 | wait until the units being started have terminated again. |
10362 | ||
171ae2cd | 10363 | * A new journal output mode "short-full" has been added which displays |
4ffe2479 | 10364 | timestamps with abbreviated English day names and adds a timezone |
171ae2cd LP |
10365 | suffix. Those timestamps include more information than the default |
10366 | "short" output mode, and can be passed directly to journalctl's | |
10367 | --since= and --until= options. | |
4ffe2479 ZJS |
10368 | |
10369 | * /etc/resolv.conf will be bind-mounted into containers started by | |
10370 | systemd-nspawn, if possible, so any changes to resolv.conf contents | |
10371 | are automatically propagated to the container. | |
10372 | ||
10373 | * The number of instances for socket-activated services originating | |
171ae2cd LP |
10374 | from a single IP address can be limited with |
10375 | MaxConnectionsPerSource=, extending the existing setting of | |
10376 | MaxConnections=. | |
4ffe2479 | 10377 | |
4a77c53d ZJS |
10378 | * systemd-networkd gained support for vcan ("Virtual CAN") interface |
10379 | configuration. | |
10380 | ||
10381 | * .netdev and .network configuration can now be extended through | |
10382 | drop-ins. | |
10383 | ||
4ffe2479 ZJS |
10384 | * UDP Segmentation Offload, TCP Segmentation Offload, Generic |
10385 | Segmentation Offload, Generic Receive Offload, Large Receive Offload | |
10386 | can be enabled and disabled using the new UDPSegmentationOffload=, | |
10387 | TCPSegmentationOffload=, GenericSegmentationOffload=, | |
10388 | GenericReceiveOffload=, LargeReceiveOffload= options in the | |
10389 | [Link] section of .link files. | |
10390 | ||
171ae2cd LP |
10391 | * The Spanning Tree Protocol, Priority, Aging Time, and the Default |
10392 | Port VLAN ID can be configured for bridge devices using the new STP=, | |
10393 | Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge] | |
10394 | section of .netdev files. | |
4ffe2479 | 10395 | |
171ae2cd | 10396 | * The route table to which routes received over DHCP or RA should be |
4a77c53d ZJS |
10397 | added can be configured with the new RouteTable= option in the [DHCP] |
10398 | and [IPv6AcceptRA] sections of .network files. | |
10399 | ||
171ae2cd | 10400 | * The Address Resolution Protocol can be disabled on links managed by |
4ffe2479 ZJS |
10401 | systemd-networkd using the ARP=no setting in the [Link] section of |
10402 | .network files. | |
10403 | ||
171ae2cd LP |
10404 | * New environment variables $SERVICE_RESULT, $EXIT_CODE and |
10405 | $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and | |
10406 | encode information about the result and exit codes of the current | |
10407 | service runtime cycle. | |
4ffe2479 | 10408 | |
4a77c53d | 10409 | * systemd-sysctl will now configure kernel parameters in the order |
1f4f4cf7 | 10410 | they occur in the configuration files. This matches what sysctl |
4a77c53d ZJS |
10411 | has been traditionally doing. |
10412 | ||
10413 | * kernel-install "plugins" that are executed to perform various | |
10414 | tasks after a new kernel is added and before an old one is removed | |
10415 | can now return a special value to terminate the procedure and | |
10416 | prevent any later plugins from running. | |
10417 | ||
76153ad4 | 10418 | * Journald's SplitMode=login setting has been deprecated. It has been |
d4c08299 | 10419 | removed from documentation, and its use is discouraged. In a future |
76153ad4 ZJS |
10420 | release it will be completely removed, and made equivalent to current |
10421 | default of SplitMode=uid. | |
10422 | ||
4a77c53d ZJS |
10423 | * Storage=both option setting in /etc/systemd/coredump.conf has been |
10424 | removed. With fast LZ4 compression storing the core dump twice is not | |
10425 | useful. | |
10426 | ||
4ffe2479 ZJS |
10427 | * The --share-system systemd-nspawn option has been replaced with an |
10428 | (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of | |
10429 | this functionality is discouraged. In addition the variables | |
10430 | $SYSTEMD_NSPAWN_SHARE_NS_IPC, $SYSTEMD_NSPAWN_SHARE_NS_PID, | |
10431 | $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of | |
10432 | individual namespaces. | |
10433 | ||
171ae2cd LP |
10434 | * "machinectl list" now shows the IP address of running containers in |
10435 | the output, as well as OS release information. | |
10436 | ||
10437 | * "loginctl list" now shows the TTY of each session in the output. | |
10438 | ||
10439 | * sd-bus gained new API calls sd_bus_track_set_recursive(), | |
10440 | sd_bus_track_get_recursive(), sd_bus_track_count_name(), | |
10441 | sd_bus_track_count_sender(). They permit usage of sd_bus_track peer | |
10442 | tracking objects in a "recursive" mode, where a single client can be | |
10443 | counted multiple times, if it takes multiple references. | |
10444 | ||
10445 | * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and | |
bc99dac5 | 10446 | sd_bus_get_exit_on_disconnect(). They may be used to make a |
171ae2cd LP |
10447 | process using sd-bus automatically exit if the bus connection is |
10448 | severed. | |
10449 | ||
10450 | * Bus clients of the service manager may now "pin" loaded units into | |
10451 | memory, by taking an explicit reference on them. This is useful to | |
10452 | ensure the client can retrieve runtime data about the service even | |
10453 | after the service completed execution. Taking such a reference is | |
10454 | available only for privileged clients and should be helpful to watch | |
10455 | running services in a race-free manner, and in particular collect | |
10456 | information about exit statuses and results. | |
10457 | ||
4c37970d LP |
10458 | * The nss-resolve module has been changed to strictly return UNAVAIL |
10459 | when communication via D-Bus with resolved failed, and NOTFOUND when | |
10460 | a lookup completed but was negative. This means it is now possible to | |
10461 | neatly configure fallbacks using nsswitch.conf result checking | |
10462 | expressions. Taking benefit of this, the new recommended | |
10463 | configuration line for the "hosts" entry in /etc/nsswitch.conf is: | |
10464 | ||
10465 | hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname | |
10466 | ||
10467 | * A new setting CtrlAltDelBurstAction= has been added to | |
10468 | /etc/systemd/system.conf which may be used to configure the precise | |
10469 | behaviour if the user on the console presses Ctrl-Alt-Del more often | |
10470 | than 7 times in 2s. Previously this would unconditionally result in | |
10471 | an expedited, immediate reboot. With this new setting the precise | |
10472 | operation may be configured in more detail, and also turned off | |
10473 | entirely. | |
10474 | ||
10475 | * In .netdev files two new settings RemoteChecksumTx= and | |
10476 | RemoteChecksumRx= are now understood that permit configuring the | |
10477 | remote checksumming logic for VXLAN networks. | |
10478 | ||
10479 | * The service manager learnt a new "invocation ID" concept for invoked | |
10480 | services. Each runtime cycle of a service will get a new invocation | |
da890466 | 10481 | ID (a 128-bit random UUID) assigned that identifies the current |
4c37970d LP |
10482 | run of the service uniquely and globally. A new invocation ID |
10483 | is generated each time a service starts up. The journal will store | |
10484 | the invocation ID of a service along with any logged messages, thus | |
10485 | making the invocation ID useful for matching the online runtime of a | |
10486 | service with the offline log data it generated in a safe way without | |
10487 | relying on synchronized timestamps. In many ways this new service | |
10488 | invocation ID concept is similar to the kernel's boot ID concept that | |
10489 | uniquely and globally identifies the runtime of each boot. The | |
10490 | invocation ID of a service is passed to the service itself via an | |
10491 | environment variable ($INVOCATION_ID). A new bus call | |
10492 | GetUnitByInvocationID() has been added that is similar to GetUnit() | |
10493 | but instead of retrieving the bus path for a unit by its name | |
10494 | retrieves it by its invocation ID. The returned path is valid only as | |
10495 | long as the passed invocation ID is current. | |
10496 | ||
10497 | * systemd-resolved gained a new "DNSStubListener" setting in | |
10498 | resolved.conf. It either takes a boolean value or the special values | |
10499 | "udp" and "tcp", and configures whether to enable the stub DNS | |
10500 | listener on 127.0.0.53:53. | |
10501 | ||
10502 | * IP addresses configured via networkd may now carry additional | |
10503 | configuration settings supported by the kernel. New options include: | |
10504 | HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=, | |
10505 | PrefixRoute=, AutoJoin=. | |
10506 | ||
10507 | * The PAM configuration fragment file for "user@.service" shipped with | |
10508 | systemd (i.e. the --user instance of systemd) has been stripped to | |
10509 | the minimum necessary to make the system boot. Previously, it | |
10510 | contained Fedora-specific stanzas that did not apply to other | |
10511 | distributions. It is expected that downstream distributions add | |
10512 | additional configuration lines, matching their needs to this file, | |
10513 | using it only as rough template of what systemd itself needs. Note | |
10514 | that this reduced fragment does not even include an invocation of | |
10515 | pam_limits which most distributions probably want to add, even though | |
10516 | systemd itself does not need it. (There's also the new build time | |
10517 | option --with-pamconfdir=no to disable installation of the PAM | |
10518 | fragment entirely.) | |
10519 | ||
10520 | * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO | |
10521 | capability is now also dropped from its set (in addition to | |
10522 | CAP_SYS_MKNOD as before). | |
10523 | ||
10524 | * In service unit files it is now possible to connect a specific named | |
10525 | file descriptor with stdin/stdout/stdout of an executed service. The | |
10526 | name may be specified in matching .socket units using the | |
10527 | FileDescriptorName= setting. | |
10528 | ||
10529 | * A number of journal settings may now be configured on the kernel | |
10530 | command line. Specifically, the following options are now understood: | |
10531 | systemd.journald.max_level_console=, | |
10532 | systemd.journald.max_level_store=, | |
10533 | systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=, | |
10534 | systemd.journald.max_level_wall=. | |
10535 | ||
10536 | * "systemctl is-enabled --full" will now show by which symlinks a unit | |
10537 | file is enabled in the unit dependency tree. | |
10538 | ||
b4eed568 LP |
10539 | * Support for VeraCrypt encrypted partitions has been added to the |
10540 | "cryptsetup" logic and /etc/crypttab. | |
10541 | ||
10542 | * systemd-detect-virt gained support for a new --private-users switch | |
10543 | that checks whether the invoking processes are running inside a user | |
10544 | namespace. Similar, a new special value "private-users" for the | |
10545 | existing ConditionVirtualization= setting has been added, permitting | |
10546 | skipping of specific units in user namespace environments. | |
10547 | ||
07393b6e LP |
10548 | Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John, |
10549 | Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin | |
10550 | Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner, | |
10551 | Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez, | |
10552 | Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick, | |
10553 | Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg, | |
10554 | Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric | |
10555 | Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang, | |
10556 | Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke, | |
10557 | Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan | |
10558 | Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker, | |
10559 | Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski, | |
10560 | Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering, | |
10561 | Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš | |
10562 | Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou, | |
10563 | Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej | |
10564 | Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy, | |
10565 | Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike | |
10566 | Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer, | |
10567 | Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny | |
10568 | Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant | |
10569 | Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit, | |
10570 | Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut | |
10571 | Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann | |
10572 | E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew | |
10573 | Jędrzejewski-Szmek, Zeal Jagannatha | |
10574 | ||
54b24597 | 10575 | — Santa Fe, 2016-11-03 |
07393b6e | 10576 | |
5cd118ba MP |
10577 | CHANGES WITH 231: |
10578 | ||
fcd30826 LP |
10579 | * In service units the various ExecXYZ= settings have been extended |
10580 | with an additional special character as first argument of the | |
43eb109a | 10581 | assigned value: if the character '+' is used the specified command |
fcd30826 LP |
10582 | line it will be run with full privileges, regardless of User=, |
10583 | Group=, CapabilityBoundingSet= and similar options. The effect is | |
10584 | similar to the existing PermissionsStartOnly= option, but allows | |
10585 | configuration of this concept for each executed command line | |
10586 | independently. | |
10587 | ||
10588 | * Services may now alter the service watchdog timeout at runtime by | |
10589 | sending a WATCHDOG_USEC= message via sd_notify(). | |
10590 | ||
10591 | * MemoryLimit= and related unit settings now optionally take percentage | |
10592 | specifications. The percentage is taken relative to the amount of | |
10593 | physical memory in the system (or in case of containers, the assigned | |
10594 | amount of memory). This allows scaling service resources neatly with | |
771de3f5 | 10595 | the amount of RAM available on the system. Similarly, systemd-logind's |
fcd30826 LP |
10596 | RuntimeDirectorySize= option now also optionally takes percentage |
10597 | values. | |
10598 | ||
10599 | * In similar fashion TasksMax= takes percentage values now, too. The | |
10600 | value is taken relative to the configured maximum number of processes | |
10601 | on the system. The per-service task maximum has been changed to 15% | |
10602 | using this functionality. (Effectively this is an increase of 512 → | |
10603 | 4915 for service units, given the kernel's default pid_max setting.) | |
10604 | ||
10605 | * Calendar time specifications in .timer units now understand a ".." | |
10606 | syntax for time ranges. Example: "4..7:10" may now be used for | |
10607 | defining a timer that is triggered at 4:10am, 5:10am, 6:10am and | |
10608 | 7:10am every day. | |
10609 | ||
10610 | * The InaccessableDirectories=, ReadOnlyDirectories= and | |
10611 | ReadWriteDirectories= unit file settings have been renamed to | |
10612 | InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be | |
10613 | applied to all kinds of file nodes, and not just directories, with | |
10614 | the exception of symlinks. Specifically these settings may now be | |
10615 | used on block and character device nodes, UNIX sockets and FIFOS as | |
10616 | well as regular files. The old names of these settings remain | |
10617 | available for compatibility. | |
10618 | ||
10619 | * systemd will now log about all service processes it kills forcibly | |
10620 | (using SIGKILL) because they remained after the clean shutdown phase | |
10621 | of the service completed. This should help identifying services that | |
10622 | shut down uncleanly. Moreover if KillUserProcesses= is enabled in | |
10623 | systemd-logind's configuration a similar log message is generated for | |
10624 | processes killed at the end of each session due to this setting. | |
10625 | ||
10626 | * systemd will now set the $JOURNAL_STREAM environment variable for all | |
10627 | services whose stdout/stderr are connected to the Journal (which | |
10628 | effectively means by default: all services). The variable contains | |
10629 | the device and inode number of the file descriptor used for | |
10630 | stdout/stderr. This may be used by invoked programs to detect whether | |
10631 | their stdout/stderr is connected to the Journal, in which case they | |
10632 | can switch over to direct Journal communication, thus being able to | |
10633 | pass extended, structured metadata along with their log messages. As | |
10634 | one example, this is now used by glib's logging primitives. | |
10635 | ||
10636 | * When using systemd's default tmp.mount unit for /tmp, the mount point | |
10637 | will now be established with the "nosuid" and "nodev" options. This | |
10638 | avoids privilege escalation attacks that put traps and exploits into | |
33db1b90 | 10639 | /tmp. However, this might cause problems if you e.g. put container |
5cd118ba MP |
10640 | images or overlays into /tmp; if you need this, override tmp.mount's |
10641 | "Options=" with a drop-in, or mount /tmp from /etc/fstab with your | |
10642 | desired options. | |
10643 | ||
fcd30826 | 10644 | * systemd now supports the "memory" cgroup controller also on |
4e1dfa45 | 10645 | cgroup v2. |
fcd30826 LP |
10646 | |
10647 | * The systemd-cgtop tool now optionally takes a control group path as | |
10648 | command line argument. If specified, the control group list shown is | |
10649 | limited to subgroups of that group. | |
10650 | ||
10651 | * The SystemCallFilter= unit file setting gained support for | |
10652 | pre-defined, named system call filter sets. For example | |
10653 | SystemCallFilter=@clock is now an effective way to make all clock | |
771de3f5 | 10654 | changing-related system calls unavailable to a service. A number of |
fcd30826 LP |
10655 | similar pre-defined groups are defined. Writing system call filters |
10656 | for system services is simplified substantially with this new | |
10657 | concept. Accordingly, all of systemd's own, long-running services now | |
10658 | enable system call filtering based on this, by default. | |
10659 | ||
10660 | * A new service setting MemoryDenyWriteExecute= has been added, taking | |
10661 | a boolean value. If turned on, a service may no longer create memory | |
10662 | mappings that are writable and executable at the same time. This | |
10663 | enhances security for services where this is enabled as it becomes | |
10664 | harder to dynamically write and then execute memory in exploited | |
10665 | service processes. This option has been enabled for all of systemd's | |
10666 | own long-running services. | |
10667 | ||
10668 | * A new RestrictRealtime= service setting has been added, taking a | |
10669 | boolean argument. If set the service's processes may no longer | |
10670 | acquire realtime scheduling. This improves security as realtime | |
10671 | scheduling may otherwise be used to easily freeze the system. | |
10672 | ||
10673 | * systemd-nspawn gained a new switch --notify-ready= taking a boolean | |
10674 | value. This may be used for requesting that the system manager inside | |
10675 | of the container reports start-up completion to nspawn which then | |
10676 | propagates this notification further to the service manager | |
10677 | supervising nspawn itself. A related option NotifyReady= in .nspawn | |
10678 | files has been added too. This functionality allows ordering of the | |
10679 | start-up of multiple containers using the usual systemd ordering | |
10680 | primitives. | |
10681 | ||
10682 | * machinectl gained a new command "stop" that is an alias for | |
10683 | "terminate". | |
10684 | ||
10685 | * systemd-resolved gained support for contacting DNS servers on | |
10686 | link-local IPv6 addresses. | |
10687 | ||
10688 | * If systemd-resolved receives the SIGUSR2 signal it will now flush all | |
10689 | its caches. A method call for requesting the same operation has been | |
10690 | added to the bus API too, and is made available via "systemd-resolve | |
10691 | --flush-caches". | |
10692 | ||
771de3f5 | 10693 | * systemd-resolve gained a new --status switch. If passed a brief |
fcd30826 LP |
10694 | summary of the used DNS configuration with per-interface information |
10695 | is shown. | |
10696 | ||
10697 | * resolved.conf gained a new Cache= boolean option, defaulting to | |
10698 | on. If turned off local DNS caching is disabled. This comes with a | |
10699 | performance penalty in particular when DNSSEC is enabled. Note that | |
771de3f5 | 10700 | resolved disables its internal caching implicitly anyway, when the |
fcd30826 LP |
10701 | configured DNS server is on a host-local IP address such as ::1 or |
10702 | 127.0.0.1, thus automatically avoiding double local caching. | |
10703 | ||
10704 | * systemd-resolved now listens on the local IP address 127.0.0.53:53 | |
10705 | for DNS requests. This improves compatibility with local programs | |
10706 | that do not use the libc NSS or systemd-resolved's bus APIs for name | |
10707 | resolution. This minimal DNS service is only available to local | |
10708 | programs and does not implement the full DNS protocol, but enough to | |
10709 | cover local DNS clients. A new, static resolv.conf file, listing just | |
10710 | this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is | |
10711 | now recommended to make /etc/resolv.conf a symlink to this file in | |
10712 | order to route all DNS lookups to systemd-resolved, regardless if | |
10713 | done via NSS, the bus API or raw DNS packets. Note that this local | |
10714 | DNS service is not as fully featured as the libc NSS or | |
10715 | systemd-resolved's bus APIs. For example, as unicast DNS cannot be | |
10716 | used to deliver link-local address information (as this implies | |
10717 | sending a local interface index along), LLMNR/mDNS support via this | |
10718 | interface is severely restricted. It is thus strongly recommended for | |
10719 | all applications to use the libc NSS API or native systemd-resolved | |
10720 | bus API instead. | |
10721 | ||
10722 | * systemd-networkd's bridge support learned a new setting | |
10723 | VLANFiltering= for controlling VLAN filtering. Moreover a new section | |
10724 | in .network files has been added for configuring VLAN bridging in | |
10725 | more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN]. | |
10726 | ||
10727 | * systemd-networkd's IPv6 Router Advertisement code now makes use of | |
10728 | the DNSSL and RDNSS options. This means IPv6 DNS configuration may | |
10729 | now be acquired without relying on DHCPv6. Two new options | |
10730 | UseDomains= and UseDNS= have been added to configure this behaviour. | |
10731 | ||
10732 | * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been | |
10733 | renamed IPv6AcceptRA=, without altering its behaviour. The old | |
10734 | setting name remains available for compatibility reasons. | |
10735 | ||
10736 | * The systemd-networkd VTI/VTI6 tunneling support gained new options | |
10737 | Key=, InputKey= and OutputKey=. | |
10738 | ||
10739 | * systemd-networkd gained support for VRF ("Virtual Routing Function") | |
10740 | interface configuration. | |
10741 | ||
10742 | * "systemctl edit" may now be used to create new unit files by | |
10743 | specifying the --force switch. | |
10744 | ||
10745 | * sd-event gained a new function sd_event_get_iteration() for | |
10746 | requesting the current iteration counter of the event loop. It starts | |
10747 | at zero and is increased by one with each event loop iteration. | |
10748 | ||
43a569a1 ZJS |
10749 | * A new rpm macro %systemd_ordering is provided by the macros.systemd |
10750 | file. It can be used in lieu of %systemd_requires in packages which | |
10751 | don't use any systemd functionality and are intended to be installed | |
10752 | in minimal containers without systemd present. This macro provides | |
ce830873 | 10753 | ordering dependencies to ensure that if the package is installed in |
43a569a1 ZJS |
10754 | the same rpm transaction as systemd, systemd will be installed before |
10755 | the scriptlets for the package are executed, allowing unit presets | |
10756 | to be handled. | |
10757 | ||
10758 | New macros %_systemdgeneratordir and %_systemdusergeneratordir have | |
10759 | been added to simplify packaging of generators. | |
10760 | ||
10761 | * The os-release file gained VERSION_CODENAME field for the | |
10762 | distribution nickname (e.g. VERSION_CODENAME=woody). | |
10763 | ||
10764 | * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1 | |
10765 | can be set to disable parsing of metadata and the creation | |
10766 | of persistent symlinks for that device. | |
10767 | ||
0f1da52b LP |
10768 | * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess" |
10769 | to make them available to logged-in users has been reverted. | |
10770 | ||
10771 | * Much of the common code of the various systemd components is now | |
10772 | built into an internal shared library libsystemd-shared-231.so | |
10773 | (incorporating the systemd version number in the name, to be updated | |
10774 | with future releases) that the components link to. This should | |
10775 | decrease systemd footprint both in memory during runtime and on | |
10776 | disk. Note that the shared library is not for public use, and is | |
ead6bd25 | 10777 | neither API nor ABI stable, but is likely to change with every new |
1ecbf32f ZJS |
10778 | released update. Packagers need to make sure that binaries |
10779 | linking to libsystemd-shared.so are updated in step with the | |
10780 | library. | |
43a569a1 | 10781 | |
fcd30826 LP |
10782 | * Configuration for "mkosi" is now part of the systemd |
10783 | repository. mkosi is a tool to easily build legacy-free OS images, | |
10784 | and is available on github: https://github.com/systemd/mkosi. If | |
10785 | "mkosi" is invoked in the build tree a new raw OS image is generated | |
10786 | incorporating the systemd sources currently being worked on and a | |
10787 | clean, fresh distribution installation. The generated OS image may be | |
ce830873 | 10788 | booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical |
fcd30826 LP |
10789 | UEFI PC. This functionality is particularly useful to easily test |
10790 | local changes made to systemd in a pristine, defined environment. See | |
f09eb768 | 10791 | doc/HACKING for details. |
ceeddf79 | 10792 | |
4ffe2479 ZJS |
10793 | * configure learned the --with-support-url= option to specify the |
10794 | distribution's bugtracker. | |
10795 | ||
38b383d9 LP |
10796 | Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor |
10797 | Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika | |
10798 | Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar | |
10799 | Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse, | |
10800 | Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David | |
10801 | Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias | |
10802 | Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler, | |
10803 | Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan | |
10804 | Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke | |
10805 | Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart | |
10806 | Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel | |
10807 | Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov, | |
10808 | Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz, | |
10809 | Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran, | |
10810 | Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier, | |
10811 | Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas | |
771de3f5 ZJS |
10812 | Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan, |
10813 | Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič, | |
38b383d9 | 10814 | WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek |
5cd118ba | 10815 | |
38b383d9 | 10816 | — Berlin, 2016-07-25 |
5cd118ba | 10817 | |
46e40fab | 10818 | CHANGES WITH 230: |
7f6e8043 | 10819 | |
61ecb465 LP |
10820 | * DNSSEC is now turned on by default in systemd-resolved (in |
10821 | "allow-downgrade" mode), but may be turned off during compile time by | |
10822 | passing "--with-default-dnssec=no" to "configure" (and of course, | |
10823 | during runtime with DNSSEC= in resolved.conf). We recommend | |
10824 | downstreams to leave this on at least during development cycles and | |
10825 | report any issues with the DNSSEC logic upstream. We are very | |
10826 | interested in collecting feedback about the DNSSEC validator and its | |
10827 | limitations in the wild. Note however, that DNSSEC support is | |
10828 | probably nothing downstreams should turn on in stable distros just | |
96d49011 | 10829 | yet, as it might create incompatibilities with a few DNS servers and |
61ecb465 LP |
10830 | networks. We tried hard to make sure we downgrade to non-DNSSEC mode |
10831 | automatically whenever we detect such incompatible setups, but there | |
10832 | might be systems we do not cover yet. Hence: please help us testing | |
10833 | the DNSSEC code, leave this on where you can, report back, but then | |
10834 | again don't consider turning this on in your stable, LTS or | |
e40a326c LP |
10835 | production release just yet. (Note that you have to enable |
10836 | nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved | |
38b38500 | 10837 | and its DNSSEC mode for hostname resolution from local |
e40a326c | 10838 | applications.) |
61ecb465 | 10839 | |
96515dbf | 10840 | * systemd-resolve conveniently resolves DANE records with the --tlsa |
e40a326c | 10841 | option and OPENPGPKEY records with the --openpgp option. It also |
e75690c3 | 10842 | supports dumping raw DNS record data via the new --raw= switch. |
96515dbf | 10843 | |
97e5530c ZJS |
10844 | * systemd-logind will now by default terminate user processes that are |
10845 | part of the user session scope unit (session-XX.scope) when the user | |
977f2bea | 10846 | logs out. This behavior is controlled by the KillUserProcesses= |
e40a326c LP |
10847 | setting in logind.conf, and the previous default of "no" is now |
10848 | changed to "yes". This means that user sessions will be properly | |
10849 | cleaned up after, but additional steps are necessary to allow | |
10850 | intentionally long-running processes to survive logout. | |
97e5530c ZJS |
10851 | |
10852 | While the user is logged in at least once, user@.service is running, | |
10853 | and any service that should survive the end of any individual login | |
10854 | session can be started at a user service or scope using systemd-run. | |
e40a326c | 10855 | systemd-run(1) man page has been extended with an example which shows |
8951eaec | 10856 | how to run screen in a scope unit underneath user@.service. The same |
e40a326c | 10857 | command works for tmux. |
97e5530c ZJS |
10858 | |
10859 | After the user logs out of all sessions, user@.service will be | |
10860 | terminated too, by default, unless the user has "lingering" enabled. | |
10861 | To effectively allow users to run long-term tasks even if they are | |
152199f2 ZJS |
10862 | logged out, lingering must be enabled for them. See loginctl(1) for |
10863 | details. The default polkit policy was modified to allow users to | |
10864 | set lingering for themselves without authentication. | |
7f6e8043 | 10865 | |
95365a57 | 10866 | Previous defaults can be restored at compile time by the |
e40a326c | 10867 | --without-kill-user-processes option to "configure". |
7f6e8043 | 10868 | |
e75690c3 ZJS |
10869 | * systemd-logind gained new configuration settings SessionsMax= and |
10870 | InhibitorsMax=, both with a default of 8192. It will not register new | |
188d3082 | 10871 | user sessions or inhibitors above this limit. |
e75690c3 ZJS |
10872 | |
10873 | * systemd-logind will now reload configuration on SIGHUP. | |
10874 | ||
96515dbf | 10875 | * The unified cgroup hierarchy added in Linux 4.5 is now supported. |
e40a326c | 10876 | Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to |
25b0e6cb LP |
10877 | enable. Also, support for the "io" cgroup controller in the unified |
10878 | hierarchy has been added, so that the "memory", "pids" and "io" are | |
10879 | now the controllers that are supported on the unified hierarchy. | |
e40a326c | 10880 | |
96515dbf ZJS |
10881 | WARNING: it is not possible to use previous systemd versions with |
10882 | systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it | |
10883 | is necessary to also update systemd in the initramfs if using the | |
e40a326c | 10884 | unified hierarchy. An updated SELinux policy is also required. |
96515dbf | 10885 | |
e40a326c LP |
10886 | * LLDP support has been extended, and both passive (receive-only) and |
10887 | active (sender) modes are supported. Passive mode ("routers-only") is | |
8951eaec ZJS |
10888 | enabled by default in systemd-networkd. Active LLDP mode is enabled |
10889 | by default for containers on the internal network. The "networkctl | |
e40a326c LP |
10890 | lldp" command may be used to list information gathered. "networkctl |
10891 | status" will also show basic LLDP information on connected peers now. | |
96515dbf | 10892 | |
e40a326c LP |
10893 | * The IAID and DUID unique identifier sent in DHCP requests may now be |
10894 | configured for the system and each .network file managed by | |
e75690c3 ZJS |
10895 | systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options. |
10896 | ||
10897 | * systemd-networkd gained support for configuring proxy ARP support for | |
10898 | each interface, via the ProxyArp= setting in .network files. It also | |
10899 | gained support for configuring the multicast querier feature of | |
10900 | bridge devices, via the new MulticastQuerier= setting in .netdev | |
10901 | files. Similarly, snooping on the IGMP traffic can be controlled | |
10902 | via the new setting MulticastSnooping=. | |
10903 | ||
10904 | A new setting PreferredLifetime= has been added for addresses | |
10905 | configured in .network file to configure the lifetime intended for an | |
10906 | address. | |
10907 | ||
10908 | The systemd-networkd DHCP server gained the option EmitRouter=, which | |
10909 | defaults to yes, to configure whether the DHCP Option 3 (Router) | |
10910 | should be emitted. | |
96515dbf | 10911 | |
e40a326c | 10912 | * The testing tool /usr/lib/systemd/systemd-activate is renamed to |
97e5530c ZJS |
10913 | systemd-socket-activate and installed into /usr/bin. It is now fully |
10914 | supported. | |
10915 | ||
e40a326c LP |
10916 | * systemd-journald now uses separate threads to flush changes to disk |
10917 | when closing journal files, thus reducing impact of slow disk I/O on | |
10918 | logging performance. | |
96515dbf | 10919 | |
e75690c3 ZJS |
10920 | * The sd-journal API gained two new calls |
10921 | sd_journal_open_directory_fd() and sd_journal_open_files_fd() which | |
10922 | can be used to open journal files using file descriptors instead of | |
10923 | file or directory paths. sd_journal_open_container() has been | |
10924 | deprecated, sd_journal_open_directory_fd() should be used instead | |
10925 | with the flag SD_JOURNAL_OS_ROOT. | |
10926 | ||
10927 | * journalctl learned a new output mode "-o short-unix" that outputs log | |
10928 | lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970 | |
10929 | UTC). It also gained support for a new --no-hostname setting to | |
10930 | suppress the hostname column in the family of "short" output modes. | |
10931 | ||
e40a326c LP |
10932 | * systemd-ask-password now optionally skips printing of the password to |
10933 | stdout with --no-output which can be useful in scripts. | |
96515dbf ZJS |
10934 | |
10935 | * Framebuffer devices (/dev/fb*) and 3D printers and scanners | |
10936 | (devices tagged with ID_MAKER_TOOL) are now tagged with | |
10937 | "uaccess" and are available to logged in users. | |
10938 | ||
e75690c3 | 10939 | * The DeviceAllow= unit setting now supports specifiers (with "%"). |
e40a326c LP |
10940 | |
10941 | * "systemctl show" gained a new --value switch, which allows print a | |
10942 | only the contents of a specific unit property, without also printing | |
8951eaec ZJS |
10943 | the property's name. Similar support was added to "show*" verbs |
10944 | of loginctl and machinectl that output "key=value" lists. | |
e40a326c | 10945 | |
e75690c3 ZJS |
10946 | * A new unit type "generated" was added for files dynamically generated |
10947 | by generator tools. Similarly, a new unit type "transient" is used | |
10948 | for unit files created using the runtime API. "systemctl enable" will | |
10949 | refuse to operate on such files. | |
10950 | ||
e40a326c LP |
10951 | * A new command "systemctl revert" has been added that may be used to |
10952 | revert to the vendor version of a unit file, in case local changes | |
10953 | have been made by adding drop-ins or overriding the unit file. | |
10954 | ||
10955 | * "machinectl clean" gained a new verb to automatically remove all or | |
10956 | just hidden container images. | |
10957 | ||
e40a326c LP |
10958 | * systemd-tmpfiles gained support for a new line type "e" for emptying |
10959 | directories, if they exist, without creating them if they don't. | |
10960 | ||
e40a326c LP |
10961 | * systemd-nspawn gained support for automatically patching the UID/GIDs |
10962 | of the owners and the ACLs of all files and directories in a | |
10963 | container tree to match the UID/GID user namespacing range selected | |
10964 | for the container invocation. This mode is enabled via the new | |
24597ee0 ZJS |
10965 | --private-users-chown switch. It also gained support for |
10966 | automatically choosing a free, previously unused UID/GID range when | |
10967 | starting a container, via the new --private-users=pick setting (which | |
10968 | implies --private-users-chown). Together, these options for the first | |
10969 | time make user namespacing for nspawn containers fully automatic and | |
10970 | thus deployable. The systemd-nspawn@.service template unit file has | |
10971 | been changed to use this functionality by default. | |
e40a326c | 10972 | |
25b0e6cb LP |
10973 | * systemd-nspawn gained a new --network-zone= switch, that allows |
10974 | creating ad-hoc virtual Ethernet links between multiple containers, | |
10975 | that only exist as long as at least one container referencing them is | |
10976 | running. This allows easy connecting of multiple containers with a | |
10977 | common link that implements an Ethernet broadcast domain. Each of | |
10978 | these network "zones" may be named relatively freely by the user, and | |
10979 | may be referenced by any number of containers, but each container may | |
10980 | only reference one of these "zones". On the lower level, this is | |
10981 | implemented by an automatically managed bridge network interface for | |
10982 | each zone, that is created when the first container referencing its | |
10983 | zone is created and removed when the last one referencing its zone | |
10984 | terminates. | |
10985 | ||
e40a326c | 10986 | * The default start timeout may now be configured on the kernel command |
8951eaec ZJS |
10987 | line via systemd.default_timeout_start_sec=. It was already |
10988 | configurable via the DefaultTimeoutStartSec= option in | |
10989 | /etc/systemd/system.conf. | |
e40a326c | 10990 | |
030bd839 | 10991 | * Socket units gained a new TriggerLimitIntervalSec= and |
e40a326c LP |
10992 | TriggerLimitBurst= setting to configure a limit on the activation |
10993 | rate of the socket unit. | |
10994 | ||
10995 | * The LimitNICE= setting now optionally takes normal UNIX nice values | |
10996 | in addition to the raw integer limit value. If the specified | |
1d3a473b | 10997 | parameter is prefixed with "+" or "-" and is in the range -20…19 the |
e40a326c LP |
10998 | value is understood as UNIX nice value. If not prefixed like this it |
10999 | is understood as raw RLIMIT_NICE limit. | |
11000 | ||
999a43f8 LP |
11001 | * Note that the effect of the PrivateDevices= unit file setting changed |
11002 | slightly with this release: the per-device /dev file system will be | |
11003 | mounted read-only from this version on, and will have "noexec" | |
188d3082 | 11004 | set. This (minor) change of behavior might cause some (exceptional) |
999a43f8 LP |
11005 | legacy software to break, when PrivateDevices=yes is set for its |
11006 | service. Please leave PrivateDevices= off if you run into problems | |
11007 | with this. | |
11008 | ||
e75690c3 ZJS |
11009 | * systemd-bootchart has been split out to a separate repository: |
11010 | https://github.com/systemd/systemd-bootchart | |
11011 | ||
11012 | * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be | |
11013 | merged into the kernel in its current form. | |
11014 | ||
11015 | * The compatibility libraries libsystemd-daemon.so, | |
11016 | libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so | |
11017 | which have been deprecated since systemd-209 have been removed along | |
11018 | with the corresponding pkg-config files. All symbols provided by | |
11019 | those libraries are provided by libsystemd.so. | |
11020 | ||
11021 | * The Capabilities= unit file setting has been removed (it is ignored | |
11022 | for backwards compatibility). AmbientCapabilities= and | |
11023 | CapabilityBoundingSet= should be used instead. | |
11024 | ||
4f9020fa DR |
11025 | * A new special target has been added, initrd-root-device.target, |
11026 | which creates a synchronization point for dependencies of the root | |
11027 | device in early userspace. Initramfs builders must ensure that this | |
11028 | target is now included in early userspace. | |
11029 | ||
e75690c3 ZJS |
11030 | Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov, |
11031 | Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin | |
11032 | Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens | |
11033 | Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh, | |
11034 | Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David | |
11035 | R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny | |
11036 | Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck | |
11037 | Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik | |
77ff6022 CG |
11038 | Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo |
11039 | Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth, | |
e75690c3 ZJS |
11040 | John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos |
11041 | Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir | |
11042 | Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt, | |
77ff6022 CG |
11043 | Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný, |
11044 | Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin, | |
11045 | mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween, | |
e75690c3 ZJS |
11046 | Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern, |
11047 | Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert | |
11048 | Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan | |
11049 | Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain | |
11050 | Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller, | |
11051 | Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen, | |
46e40fab ZJS |
11052 | Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso, |
11053 | Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev, | |
11054 | Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew | |
11055 | Jędrzejewski-Szmek | |
e40a326c | 11056 | |
46e40fab | 11057 | — Fairfax, 2016-05-21 |
96515dbf | 11058 | |
61f32bff MP |
11059 | CHANGES WITH 229: |
11060 | ||
d5f8b295 LP |
11061 | * The systemd-resolved DNS resolver service has gained a substantial |
11062 | set of new features, most prominently it may now act as a DNSSEC | |
11063 | validating stub resolver. DNSSEC mode is currently turned off by | |
ed5f8840 ZJS |
11064 | default, but is expected to be turned on by default in one of the |
11065 | next releases. For now, we invite everybody to test the DNSSEC logic | |
11066 | by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The | |
11067 | service also gained a full set of D-Bus interfaces, including calls | |
11068 | to configure DNS and DNSSEC settings per link (for use by external | |
11069 | network management software). systemd-resolved and systemd-networkd | |
11070 | now distinguish between "search" and "routing" domains. The former | |
11071 | are used to qualify single-label names, the latter are used purely | |
11072 | for routing lookups within certain domains to specific links. | |
11073 | resolved now also synthesizes RRs for all entries from /etc/hosts. | |
d5f8b295 LP |
11074 | |
11075 | * The systemd-resolve tool (which is a client utility for | |
ed5f8840 ZJS |
11076 | systemd-resolved) has been improved considerably and is now fully |
11077 | supported and documented. Hence it has moved from /usr/lib/systemd to | |
11078 | /usr/bin. | |
d5f8b295 LP |
11079 | |
11080 | * /dev/disk/by-path/ symlink support has been (re-)added for virtio | |
11081 | devices. | |
11082 | ||
a7c723c0 LP |
11083 | * The coredump collection logic has been reworked: when a coredump is |
11084 | collected it is now written to disk, compressed and processed | |
11085 | (including stacktrace extraction) from a new instantiated service | |
11086 | systemd-coredump@.service, instead of directly from the | |
11087 | /proc/sys/kernel/core_pattern hook we provide. This is beneficial as | |
11088 | processing large coredumps can take up a substantial amount of | |
11089 | resources and time, and this previously happened entirely outside of | |
11090 | systemd's service supervision. With the new logic the core_pattern | |
11091 | hook only does minimal metadata collection before passing off control | |
11092 | to the new instantiated service, which is configured with a time | |
11093 | limit, a nice level and other settings to minimize negative impact on | |
11094 | the rest of the system. Also note that the new logic will honour the | |
11095 | RLIMIT_CORE setting of the crashed process, which now allows users | |
11096 | and processes to turn off coredumping for their processes by setting | |
11097 | this limit. | |
11098 | ||
11099 | * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1 | |
11100 | and all forked processes by default. Previously, PID 1 would leave | |
11101 | the setting at "0" for all processes, as set by the kernel. Note that | |
11102 | the resource limit traditionally has no effect on the generated | |
11103 | coredumps on the system if the /proc/sys/kernel/core_pattern hook | |
11104 | logic is used. Since the limit is now honoured (see above) its | |
11105 | default has been changed so that the coredumping logic is enabled by | |
11106 | default for all processes, while allowing specific opt-out. | |
11107 | ||
11108 | * When the stacktrace is extracted from processes of system users, this | |
11109 | is now done as "systemd-coredump" user, in order to sandbox this | |
11110 | potentially security sensitive parsing operation. (Note that when | |
11111 | processing coredumps of normal users this is done under the user ID | |
11112 | of process that crashed, as before.) Packagers should take notice | |
11113 | that it is now necessary to create the "systemd-coredump" system user | |
11114 | and group at package installation time. | |
11115 | ||
d5f8b295 LP |
11116 | * The systemd-activate socket activation testing tool gained support |
11117 | for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram | |
11118 | and --seqpacket switches. It also has been extended to support both | |
11119 | new-style and inetd-style file descriptor passing. Use the new | |
11120 | --inetd switch to request inetd-style file descriptor passing. | |
11121 | ||
8968aea0 MP |
11122 | * Most systemd tools now honor a new $SYSTEMD_COLORS environment |
11123 | variable, which takes a boolean value. If set to false, ANSI color | |
d5f8b295 LP |
11124 | output is disabled in the tools even when run on a terminal that |
11125 | supports it. | |
11126 | ||
11127 | * The VXLAN support in networkd now supports two new settings | |
11128 | DestinationPort= and PortRange=. | |
11129 | ||
11130 | * A new systemd.machine_id= kernel command line switch has been added, | |
11131 | that may be used to set the machine ID in /etc/machine-id if it is | |
11132 | not initialized yet. This command line option has no effect if the | |
11133 | file is already initialized. | |
11134 | ||
11135 | * systemd-nspawn gained a new --as-pid2 switch that invokes any | |
11136 | specified command line as PID 2 rather than PID 1 in the | |
ed5f8840 ZJS |
11137 | container. In this mode PID 1 is a minimal stub init process that |
11138 | implements the special POSIX and Linux semantics of PID 1 regarding | |
11139 | signal and child process management. Note that this stub init process | |
11140 | is implemented in nspawn itself and requires no support from the | |
11141 | container image. This new logic is useful to support running | |
11142 | arbitrary commands in the container, as normal processes are | |
d5f8b295 LP |
11143 | generally not prepared to run as PID 1. |
11144 | ||
11145 | * systemd-nspawn gained a new --chdir= switch for setting the current | |
11146 | working directory for the process started in the container. | |
11147 | ||
ed5f8840 ZJS |
11148 | * "journalctl /dev/sda" will now output all kernel log messages for |
11149 | specified device from the current boot, in addition to all devices | |
11150 | that are parents of it. This should make log output about devices | |
11151 | pretty useful, as long as kernel drivers attach enough metadata to | |
11152 | the log messages. (The usual SATA drivers do.) | |
d5f8b295 LP |
11153 | |
11154 | * The sd-journal API gained two new calls | |
11155 | sd_journal_has_runtime_files() and sd_journal_has_persistent_files() | |
11156 | that report whether log data from /run or /var has been found. | |
11157 | ||
11158 | * journalctl gained a new switch "--fields" that prints all journal | |
33db1b90 | 11159 | record field names currently in use in the journal. This is backed |
d5f8b295 LP |
11160 | by two new sd-journal API calls sd_journal_enumerate_fields() and |
11161 | sd_journal_restart_fields(). | |
11162 | ||
11163 | * Most configurable timeouts in systemd now expect an argument of | |
8968aea0 MP |
11164 | "infinity" to turn them off, instead of "0" as before. The semantics |
11165 | from now on is that a timeout of "0" means "now", and "infinity" | |
11166 | means "never". To maintain backwards compatibility, "0" continues to | |
11167 | turn off previously existing timeout settings. | |
d5f8b295 LP |
11168 | |
11169 | * "systemctl reload-or-try-restart" has been renamed to "systemctl | |
8968aea0 MP |
11170 | try-reload-or-restart" to clarify what it actually does: the "try" |
11171 | logic applies to both reloading and restarting, not just restarting. | |
11172 | The old name continues to be accepted for compatibility. | |
11173 | ||
11174 | * On boot-up, when PID 1 detects that the system clock is behind the | |
11175 | release date of the systemd version in use, the clock is now set | |
11176 | to the latter. Previously, this was already done in timesyncd, in order | |
d5f8b295 LP |
11177 | to avoid running with clocks set to the various clock epochs such as |
11178 | 1902, 1938 or 1970. With this change the logic is now done in PID 1 | |
11179 | in addition to timesyncd during early boot-up, so that it is enforced | |
11180 | before the first process is spawned by systemd. Note that the logic | |
11181 | in timesyncd remains, as it is more comprehensive and ensures | |
ed5f8840 | 11182 | clock monotonicity by maintaining a persistent timestamp file in |
d5f8b295 LP |
11183 | /var. Since /var is generally not available in earliest boot or the |
11184 | initrd, this part of the logic remains in timesyncd, and is not done | |
11185 | by PID 1. | |
11186 | ||
50f48ad3 DM |
11187 | * Support for tweaking details in net_cls.class_id through the |
11188 | NetClass= configuration directive has been removed, as the kernel | |
11189 | people have decided to deprecate that controller in cgroup v2. | |
11190 | Userspace tools such as nftables are moving over to setting rules | |
11191 | that are specific to the full cgroup path of a task, which obsoletes | |
11192 | these controllers anyway. The NetClass= directive is kept around for | |
11193 | legacy compatibility reasons. For a more in-depth description of the | |
11194 | kernel change, please refer to the respective upstream commit: | |
11195 | ||
11196 | https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671 | |
11197 | ||
d5f8b295 | 11198 | * A new service setting RuntimeMaxSec= has been added that may be used |
8968aea0 | 11199 | to specify a maximum runtime for a service. If the timeout is hit, the |
d5f8b295 LP |
11200 | service is terminated and put into a failure state. |
11201 | ||
8968aea0 MP |
11202 | * A new service setting AmbientCapabilities= has been added. It allows |
11203 | configuration of additional Linux process capabilities that are | |
11204 | passed to the activated processes. This is only available on very | |
d5f8b295 LP |
11205 | recent kernels. |
11206 | ||
11207 | * The process resource limit settings in service units may now be used | |
11208 | to configure hard and soft limits individually. | |
11209 | ||
8968aea0 | 11210 | * The various libsystemd APIs such as sd-bus or sd-event now publicly |
ed5f8840 ZJS |
11211 | expose support for gcc's __attribute__((cleanup())) C extension. |
11212 | Specifically, for many object destructor functions alternative | |
11213 | versions have been added that have names suffixed with "p" and take a | |
11214 | pointer to a pointer to the object to destroy, instead of just a | |
11215 | pointer to the object itself. This is useful because these destructor | |
11216 | functions may be used directly as parameters to the cleanup | |
11217 | construct. Internally, systemd has been a heavy user of this GCC | |
11218 | extension for a long time, and with this change similar support is | |
11219 | now available to consumers of the library outside of systemd. Note | |
8968aea0 | 11220 | that by using this extension in your sources compatibility with old |
ed5f8840 ZJS |
11221 | and strictly ANSI compatible C compilers is lost. However, all gcc or |
11222 | LLVM versions of recent years support this extension. | |
d5f8b295 LP |
11223 | |
11224 | * Timer units gained support for a new setting RandomizedDelaySec= that | |
8968aea0 MP |
11225 | allows configuring some additional randomized delay to the configured |
11226 | time. This is useful to spread out timer events to avoid load peaks in | |
11227 | clusters or larger setups. | |
d5f8b295 LP |
11228 | |
11229 | * Calendar time specifications now support sub-second accuracy. | |
11230 | ||
11231 | * Socket units now support listening on SCTP and UDP-lite protocol | |
11232 | sockets. | |
11233 | ||
11234 | * The sd-event API now comes with a full set of man pages. | |
11235 | ||
11236 | * Older versions of systemd contained experimental support for | |
11237 | compressing journal files and coredumps with the LZ4 compressor that | |
11238 | was not compatible with the lz4 binary (due to API limitations of the | |
11239 | lz4 library). This support has been removed; only support for files | |
11240 | compatible with the lz4 binary remains. This LZ4 logic is now | |
11241 | officially supported and no longer considered experimental. | |
11242 | ||
11243 | * The dkr image import logic has been removed again from importd. dkr's | |
11244 | micro-services focus doesn't fit into the machine image focus of | |
11245 | importd, and quickly got out of date with the upstream dkr API. | |
11246 | ||
11247 | * Creation of the /run/lock/lockdev/ directory was dropped from | |
11248 | tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have | |
61f32bff MP |
11249 | been available for many years. If you still need this, you need to |
11250 | create your own tmpfiles.d config file with: | |
d5f8b295 LP |
11251 | |
11252 | d /run/lock/lockdev 0775 root lock - | |
61f32bff | 11253 | |
dd95b381 LP |
11254 | * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction= |
11255 | and RebootArgument= have been moved from the [Service] section of | |
11256 | unit files to [Unit], and they are now supported on all unit types, | |
11257 | not just service units. Of course, systemd will continue to | |
11258 | understand these settings also at the old location, in order to | |
11259 | maintain compatibility. | |
11260 | ||
3545ab35 LP |
11261 | Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander |
11262 | Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov, | |
11263 | Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler, | |
11264 | Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan | |
11265 | Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack, | |
11266 | David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman, | |
11267 | Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen, | |
11268 | Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen, | |
11269 | Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub | |
11270 | Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek, | |
11271 | Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos, | |
11272 | lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel | |
11273 | Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer, | |
11274 | Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils | |
11275 | Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz, | |
11276 | Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant | |
11277 | Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel | |
11278 | Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito | |
11279 | Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek | |
11280 | ||
ccddd104 | 11281 | — Berlin, 2016-02-11 |
61f32bff | 11282 | |
a11c7ea5 LP |
11283 | CHANGES WITH 228: |
11284 | ||
a11c7ea5 LP |
11285 | * A number of properties previously only settable in unit |
11286 | files are now also available as properties to set when | |
11287 | creating transient units programmatically via the bus, as it | |
11288 | is exposed with systemd-run's --property= | |
11289 | setting. Specifically, these are: SyslogIdentifier=, | |
11290 | SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=, | |
11291 | EnvironmentFile=, ReadWriteDirectories=, | |
11292 | ReadOnlyDirectories=, InaccessibleDirectories=, | |
11293 | ProtectSystem=, ProtectHome=, RuntimeDirectory=. | |
11294 | ||
28c85daf LP |
11295 | * When creating transient services via the bus API it is now |
11296 | possible to pass in a set of file descriptors to use as | |
11297 | STDIN/STDOUT/STDERR for the invoked process. | |
a11c7ea5 | 11298 | |
f1f8a5a5 LP |
11299 | * Slice units may now be created transiently via the bus APIs, |
11300 | similar to the way service and scope units may already be | |
11301 | created transiently. | |
11302 | ||
a11c7ea5 LP |
11303 | * Wherever systemd expects a calendar timestamp specification |
11304 | (like in journalctl's --since= and --until= switches) UTC | |
11305 | timestamps are now supported. Timestamps suffixed with "UTC" | |
11306 | are now considered to be in Universal Time Coordinated | |
11307 | instead of the local timezone. Also, timestamps may now | |
815bb5bd | 11308 | optionally be specified with sub-second accuracy. Both of |
a11c7ea5 LP |
11309 | these additions also apply to recurring calendar event |
11310 | specification, such as OnCalendar= in timer units. | |
11311 | ||
28c85daf LP |
11312 | * journalctl gained a new "--sync" switch that asks the |
11313 | journal daemon to write all so far unwritten log messages to | |
11314 | disk and sync the files, before returning. | |
11315 | ||
a11c7ea5 LP |
11316 | * systemd-tmpfiles learned two new line types "q" and "Q" that |
11317 | operate like "v", but also set up a basic btrfs quota | |
11318 | hierarchy when used on a btrfs file system with quota | |
11319 | enabled. | |
11320 | ||
f1f8a5a5 LP |
11321 | * tmpfiles' "v", "q" and "Q" will now create a plain directory |
11322 | instead of a subvolume (even on a btrfs file system) if the | |
11323 | root directory is a plain directory, and not a | |
11324 | subvolume. This should simplify things with certain chroot() | |
11325 | environments which are not aware of the concept of btrfs | |
11326 | subvolumes. | |
11327 | ||
a11c7ea5 LP |
11328 | * systemd-detect-virt gained a new --chroot switch to detect |
11329 | whether execution takes place in a chroot() environment. | |
11330 | ||
28c85daf | 11331 | * CPUAffinity= now takes CPU index ranges in addition to |
a11c7ea5 LP |
11332 | individual indexes. |
11333 | ||
28c85daf | 11334 | * The various memory-related resource limit settings (such as |
1d3a473b | 11335 | LimitAS=) now understand the usual K, M, G, … suffixes to |
28c85daf | 11336 | the base of 1024 (IEC). Similar, the time-related resource |
1d3a473b ZJS |
11337 | limit settings understand the usual min, h, day, … suffixes |
11338 | now. | |
28c85daf | 11339 | |
f1f8a5a5 LP |
11340 | * There's a new system.conf setting DefaultTasksMax= to |
11341 | control the default TasksMax= setting for services and | |
11342 | scopes running on the system. (TasksMax= is the primary | |
11343 | setting that exposes the "pids" cgroup controller on systemd | |
11344 | and was introduced in the previous systemd release.) The | |
11345 | setting now defaults to 512, which means services that are | |
11346 | not explicitly configured otherwise will only be able to | |
11347 | create 512 processes or threads at maximum, from this | |
11348 | version on. Note that this means that thread- or | |
11349 | process-heavy services might need to be reconfigured to set | |
11350 | TasksMax= to a higher value. It is sufficient to set | |
11351 | TasksMax= in these specific unit files to a higher value, or | |
11352 | even "infinity". Similar, there's now a logind.conf setting | |
11353 | UserTasksMax= that defaults to 4096 and limits the total | |
11354 | number of processes or tasks each user may own | |
11355 | concurrently. nspawn containers also have the TasksMax= | |
11356 | value set by default now, to 8192. Note that all of this | |
11357 | only has an effect if the "pids" cgroup controller is | |
11358 | enabled in the kernel. The general benefit of these changes | |
11359 | should be a more robust and safer system, that provides a | |
11360 | certain amount of per-service fork() bomb protection. | |
11361 | ||
28c85daf LP |
11362 | * systemd-nspawn gained the new --network-veth-extra= switch |
11363 | to define additional and arbitrarily-named virtual Ethernet | |
11364 | links between the host and the container. | |
11365 | ||
11366 | * A new service execution setting PassEnvironment= has been | |
11367 | added that allows importing select environment variables | |
11368 | from PID1's environment block into the environment block of | |
11369 | the service. | |
11370 | ||
ddb4b0d3 | 11371 | * Timer units gained support for a new RemainAfterElapse= |
595bfe7d | 11372 | setting which takes a boolean argument. It defaults to on, |
ddb4b0d3 LP |
11373 | exposing behaviour unchanged to previous releases. If set to |
11374 | off, timer units are unloaded after they elapsed if they | |
11375 | cannot elapse again. This is particularly useful for | |
11376 | transient timer units, which shall not stay around longer | |
11377 | than until they first elapse. | |
11378 | ||
a11c7ea5 | 11379 | * systemd will now bump the net.unix.max_dgram_qlen to 512 by |
28c85daf LP |
11380 | default now (the kernel default is 16). This is beneficial |
11381 | for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it | |
a11c7ea5 LP |
11382 | allows substantially larger numbers of queued |
11383 | datagrams. This should increase the capability of systemd to | |
11384 | parallelize boot-up, as logging and sd_notify() are unlikely | |
11385 | to stall execution anymore. If you need to change the value | |
11386 | from the new defaults, use the usual sysctl.d/ snippets. | |
11387 | ||
28c85daf LP |
11388 | * The compression framing format used by the journal or |
11389 | coredump processing has changed to be in line with what the | |
11390 | official LZ4 tools generate. LZ4 compression support in | |
11391 | systemd was considered unsupported previously, as the format | |
11392 | was not compatible with the normal tools. With this release | |
11393 | this has changed now, and it is hence safe for downstream | |
11394 | distributions to turn it on. While not compressing as well | |
815bb5bd | 11395 | as the XZ, LZ4 is substantially faster, which makes |
28c85daf LP |
11396 | it a good default choice for the compression logic in the |
11397 | journal and in coredump handling. | |
a11c7ea5 | 11398 | |
28c85daf LP |
11399 | * Any reference to /etc/mtab has been dropped from |
11400 | systemd. The file has been obsolete since a while, but | |
11401 | systemd refused to work on systems where it was incorrectly | |
815bb5bd | 11402 | set up (it should be a symlink or non-existent). Please make |
28c85daf LP |
11403 | sure to update to util-linux 2.27.1 or newer in conjunction |
11404 | with this systemd release, which also drops any reference to | |
11405 | /etc/mtab. If you maintain a distribution make sure that no | |
11406 | software you package still references it, as this is a | |
11407 | likely source of bugs. There's also a glibc bug pending, | |
11408 | asking for removal of any reference to this obsolete file: | |
11409 | ||
11410 | https://sourceware.org/bugzilla/show_bug.cgi?id=19108 | |
a11c7ea5 | 11411 | |
d5bd92bb LP |
11412 | Note that only util-linux versions built with |
11413 | --enable-libmount-force-mountinfo are supported. | |
11414 | ||
a11c7ea5 LP |
11415 | * Support for the ".snapshot" unit type has been removed. This |
11416 | feature turned out to be little useful and little used, and | |
11417 | has now been removed from the core and from systemctl. | |
11418 | ||
b9e2f7eb LP |
11419 | * The dependency types RequiresOverridable= and |
11420 | RequisiteOverridable= have been removed from systemd. They | |
11421 | have been used only very sparingly to our knowledge and | |
11422 | other options that provide a similar effect (such as | |
11423 | systemctl --mode=ignore-dependencies) are much more useful | |
11424 | and commonly used. Moreover, they were only half-way | |
11425 | implemented as the option to control behaviour regarding | |
11426 | these dependencies was never added to systemctl. By removing | |
11427 | these dependency types the execution engine becomes a bit | |
11428 | simpler. Unit files that use these dependencies should be | |
11429 | changed to use the non-Overridable dependency types | |
11430 | instead. In fact, when parsing unit files with these | |
11431 | options, that's what systemd will automatically convert them | |
11432 | too, but it will also warn, asking users to fix the unit | |
11433 | files accordingly. Removal of these dependency types should | |
11434 | only affect a negligible number of unit files in the wild. | |
11435 | ||
11436 | * Behaviour of networkd's IPForward= option changed | |
11437 | (again). It will no longer maintain a per-interface setting, | |
11438 | but propagate one way from interfaces where this is enabled | |
11439 | to the global kernel setting. The global setting will be | |
11440 | enabled when requested by a network that is set up, but | |
11441 | never be disabled again. This change was made to make sure | |
11442 | IPv4 and IPv6 behaviour regarding packet forwarding is | |
11443 | similar (as the Linux IPv6 stack does not support | |
11444 | per-interface control of this setting) and to minimize | |
11445 | surprises. | |
11446 | ||
28c85daf LP |
11447 | * In unit files the behaviour of %u, %U, %h, %s has |
11448 | changed. These specifiers will now unconditionally resolve | |
11449 | to the various user database fields of the user that the | |
11450 | systemd instance is running as, instead of the user | |
11451 | configured in the specific unit via User=. Note that this | |
11452 | effectively doesn't change much, as resolving of these | |
11453 | specifiers was already turned off in the --system instance | |
11454 | of systemd, as we cannot do NSS lookups from PID 1. In the | |
11455 | --user instance of systemd these specifiers where correctly | |
11456 | resolved, but hardly made any sense, since the user instance | |
11457 | lacks privileges to do user switches anyway, and User= is | |
ce830873 | 11458 | hence useless. Moreover, even in the --user instance of |
28c85daf LP |
11459 | systemd behaviour was awkward as it would only take settings |
11460 | from User= assignment placed before the specifier into | |
11461 | account. In order to unify and simplify the logic around | |
11462 | this the specifiers will now always resolve to the | |
11463 | credentials of the user invoking the manager (which in case | |
11464 | of PID 1 is the root user). | |
11465 | ||
11466 | Contributions from: Andrew Jones, Beniamino Galvani, Boyuan | |
11467 | Yang, Daniel Machon, Daniel Mack, David Herrmann, David | |
11468 | Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin, | |
f1f8a5a5 LP |
11469 | Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo |
11470 | Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan | |
11471 | Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers, | |
11472 | Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel | |
11473 | Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark | |
11474 | Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich, | |
11475 | Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens, | |
11476 | Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer, | |
11477 | Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden, | |
11478 | Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen, | |
11479 | Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew | |
11480 | Jędrzejewski-Szmek | |
28c85daf | 11481 | |
ccddd104 | 11482 | — Berlin, 2015-11-18 |
a11c7ea5 | 11483 | |
c97e586d DM |
11484 | CHANGES WITH 227: |
11485 | ||
11486 | * systemd now depends on util-linux v2.27. More specifically, | |
11487 | the newly added mount monitor feature in libmount now | |
11488 | replaces systemd's former own implementation. | |
11489 | ||
11490 | * libmount mandates /etc/mtab not to be regular file, and | |
11491 | systemd now enforces this condition at early boot. | |
11492 | /etc/mtab has been deprecated and warned about for a very | |
11493 | long time, so systems running systemd should already have | |
11494 | stopped having this file around as anything else than a | |
11495 | symlink to /proc/self/mounts. | |
11496 | ||
33db1b90 | 11497 | * Support for the "pids" cgroup controller has been added. It |
d046fb93 | 11498 | allows accounting the number of tasks in a cgroup and |
c97e586d DM |
11499 | enforcing limits on it. This adds two new setting |
11500 | TasksAccounting= and TasksMax= to each unit, as well as a | |
6fd5517b | 11501 | global option DefaultTasksAccounting=. |
c97e586d DM |
11502 | |
11503 | * Support for the "net_cls" cgroup controller has been added. | |
fe08a30b LP |
11504 | It allows assigning a net class ID to each task in the |
11505 | cgroup, which can then be used in firewall rules and traffic | |
11506 | shaping configurations. Note that the kernel netfilter net | |
11507 | class code does not currently work reliably for ingress | |
11508 | packets on unestablished sockets. | |
c97e586d DM |
11509 | |
11510 | This adds a new config directive called NetClass= to CGroup | |
6fd5517b | 11511 | enabled units. Allowed values are positive numbers for fixed |
c97e586d DM |
11512 | assignments and "auto" for picking a free value |
11513 | automatically. | |
11514 | ||
21d86c61 DM |
11515 | * 'systemctl is-system-running' now returns 'offline' if the |
11516 | system is not booted with systemd. This command can now be | |
11517 | used as a substitute for 'systemd-notify --booted'. | |
11518 | ||
11519 | * Watchdog timeouts have been increased to 3 minutes for all | |
11520 | in-tree service files. Apparently, disk IO issues are more | |
11521 | frequent than we hoped, and user reported >1 minute waiting | |
11522 | for disk IO. | |
11523 | ||
11524 | * 'machine-id-commit' functionality has been merged into | |
11525 | 'machine-id-setup --commit'. The separate binary has been | |
11526 | removed. | |
11527 | ||
d046fb93 LP |
11528 | * The WorkingDirectory= directive in unit files may now be set |
11529 | to the special value '~'. In this case, the working | |
11530 | directory is set to the home directory of the user | |
11531 | configured in User=. | |
21d86c61 | 11532 | |
fe08a30b LP |
11533 | * "machinectl shell" will now open the shell in the home |
11534 | directory of the selected user by default. | |
11535 | ||
21d86c61 | 11536 | * The CrashChVT= configuration file setting is renamed to |
d046fb93 LP |
11537 | CrashChangeVT=, following our usual logic of not |
11538 | abbreviating unnecessarily. The old directive is still | |
11539 | supported for compat reasons. Also, this directive now takes | |
11540 | an integer value between 1 and 63, or a boolean value. The | |
11541 | formerly supported '-1' value for disabling stays around for | |
11542 | compat reasons. | |
21d86c61 | 11543 | |
fe08a30b | 11544 | * The PrivateTmp=, PrivateDevices=, PrivateNetwork=, |
8b5f9d15 | 11545 | NoNewPrivileges=, TTYPath=, WorkingDirectory= and |
fe08a30b LP |
11546 | RootDirectory= properties can now be set for transient |
11547 | units. | |
11548 | ||
11549 | * The systemd-analyze tool gained a new "set-log-target" verb | |
11550 | to change the logging target the system manager logs to | |
11551 | dynamically during runtime. This is similar to how | |
11552 | "systemd-analyze set-log-level" already changes the log | |
11553 | level. | |
11554 | ||
11555 | * In nspawn /sys is now mounted as tmpfs, with only a selected | |
11556 | set of subdirectories mounted in from the real sysfs. This | |
11557 | enhances security slightly, and is useful for ensuring user | |
11558 | namespaces work correctly. | |
11559 | ||
11560 | * Support for USB FunctionFS activation has been added. This | |
11561 | allows implementation of USB gadget services that are | |
11562 | activated as soon as they are requested, so that they don't | |
595bfe7d | 11563 | have to run continuously, similar to classic socket |
fe08a30b LP |
11564 | activation. |
11565 | ||
11566 | * The "systemctl exit" command now optionally takes an | |
11567 | additional parameter that sets the exit code to return from | |
11568 | the systemd manager when exiting. This is only relevant when | |
11569 | running the systemd user instance, or when running the | |
11570 | system instance in a container. | |
11571 | ||
11572 | * sd-bus gained the new API calls sd_bus_path_encode_many() | |
11573 | and sd_bus_path_decode_many() that allow easy encoding and | |
11574 | decoding of multiple identifier strings inside a D-Bus | |
11575 | object path. Another new call sd_bus_default_flush_close() | |
11576 | has been added to flush and close per-thread default | |
11577 | connections. | |
11578 | ||
11579 | * systemd-cgtop gained support for a -M/--machine= switch to | |
11580 | show the control groups within a certain container only. | |
11581 | ||
11582 | * "systemctl kill" gained support for an optional --fail | |
11583 | switch. If specified the requested operation will fail of no | |
11584 | processes have been killed, because the unit had no | |
11585 | processes attached, or similar. | |
11586 | ||
bdba9227 DM |
11587 | * A new systemd.crash_reboot=1 kernel command line option has |
11588 | been added that triggers a reboot after crashing. This can | |
11589 | also be set through CrashReboot= in systemd.conf. | |
11590 | ||
11591 | * The RuntimeDirectory= setting now understands unit | |
11592 | specifiers like %i or %f. | |
11593 | ||
ce830873 | 11594 | * A new (still internal) library API sd-ipv4acd has been added, |
fe08a30b LP |
11595 | that implements address conflict detection for IPv4. It's |
11596 | based on code from sd-ipv4ll, and will be useful for | |
11597 | detecting DHCP address conflicts. | |
11598 | ||
bdba9227 DM |
11599 | * File descriptors passed during socket activation may now be |
11600 | named. A new API sd_listen_fds_with_names() is added to | |
33db1b90 | 11601 | access the names. The default names may be overridden, |
bdba9227 DM |
11602 | either in the .socket file using the FileDescriptorName= |
11603 | parameter, or by passing FDNAME= when storing the file | |
11604 | descriptors using sd_notify(). | |
fe08a30b | 11605 | |
d046fb93 LP |
11606 | * systemd-networkd gained support for: |
11607 | ||
0053598f | 11608 | - Setting the IPv6 Router Advertisement settings via |
edf4126f | 11609 | IPv6AcceptRouterAdvertisements= in .network files. |
d046fb93 LP |
11610 | |
11611 | - Configuring the HelloTimeSec=, MaxAgeSec= and | |
11612 | ForwardDelaySec= bridge parameters in .netdev files. | |
11613 | ||
11614 | - Configuring PreferredSource= for static routes in | |
edf4126f | 11615 | .network files. |
fe08a30b | 11616 | |
bdba9227 DM |
11617 | * The "ask-password" framework used to query for LUKS harddisk |
11618 | passwords or SSL passwords during boot gained support for | |
11619 | caching passwords in the kernel keyring, if it is | |
11620 | available. This makes sure that the user only has to type in | |
11621 | a passphrase once if there are multiple objects to unlock | |
11622 | with the same one. Previously, such password caching was | |
11623 | available only when Plymouth was used; this moves the | |
11624 | caching logic into the systemd codebase itself. The | |
11625 | "systemd-ask-password" utility gained a new --keyname= | |
11626 | switch to control which kernel keyring key to use for | |
11627 | caching a password in. This functionality is also useful for | |
11628 | enabling display managers such as gdm to automatically | |
11629 | unlock the user's GNOME keyring if its passphrase, the | |
11630 | user's password and the harddisk password are the same, if | |
11631 | gdm-autologin is used. | |
fe08a30b LP |
11632 | |
11633 | * When downloading tar or raw images using "machinectl | |
11634 | pull-tar" or "machinectl pull-raw", a matching ".nspawn" | |
11635 | file is now also downloaded, if it is available and stored | |
11636 | next to the image file. | |
c97e586d | 11637 | |
91d0d699 LP |
11638 | * Units of type ".socket" gained a new boolean setting |
11639 | Writable= which is only useful in conjunction with | |
11640 | ListenSpecial=. If true, enables opening the specified | |
11641 | special file in O_RDWR mode rather than O_RDONLY mode. | |
11642 | ||
11643 | * systemd-rfkill has been reworked to become a singleton | |
11644 | service that is activated through /dev/rfkill on each rfkill | |
11645 | state change and saves the settings to disk. This way, | |
11646 | systemd-rfkill is now compatible with devices that exist | |
11647 | only intermittendly, and even restores state if the previous | |
11648 | system shutdown was abrupt rather than clean. | |
11649 | ||
d046fb93 LP |
11650 | * The journal daemon gained support for vacuuming old journal |
11651 | files controlled by the number of files that shall remain, | |
11652 | in addition to the already existing control by size and by | |
11653 | date. This is useful as journal interleaving performance | |
6dd6a9c4 | 11654 | degrades with too many separate journal files, and allows |
d046fb93 LP |
11655 | putting an effective limit on them. The new setting defaults |
11656 | to 100, but this may be changed by setting SystemMaxFiles= | |
11657 | and RuntimeMaxFiles= in journald.conf. Also, the | |
11658 | "journalctl" tool gained the new --vacuum-files= switch to | |
11659 | manually vacuum journal files to leave only the specified | |
11660 | number of files in place. | |
c48eb61f | 11661 | |
bdba9227 DM |
11662 | * udev will now create /dev/disk/by-path links for ATA devices |
11663 | on kernels where that is supported. | |
c30f086f | 11664 | |
efce0ffe | 11665 | * Galician, Serbian, Turkish and Korean translations were added. |
c97e586d | 11666 | |
61e6771c LP |
11667 | Contributions from: Aaro Koskinen, Alban Crequy, Beniamino |
11668 | Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao | |
11669 | (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David | |
11670 | Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin, | |
11671 | Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel | |
11672 | de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner, | |
11673 | Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen, | |
11674 | Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir | |
11675 | Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart | |
11676 | Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, | |
11677 | Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl, | |
11678 | Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike | |
11679 | Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma, | |
11680 | Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer, | |
11681 | Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani, | |
11682 | Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel | |
11683 | Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich, | |
11684 | Zbigniew Jędrzejewski-Szmek, Марко М. Костић | |
11685 | ||
ccddd104 | 11686 | — Berlin, 2015-10-07 |
c97e586d | 11687 | |
c9912c5e DH |
11688 | CHANGES WITH 226: |
11689 | ||
5e8d4254 LP |
11690 | * The DHCP implementation of systemd-networkd gained a set of |
11691 | new features: | |
11692 | ||
11693 | - The DHCP server now supports emitting DNS and NTP | |
11694 | information. It may be enabled and configured via | |
11695 | EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS | |
11696 | and NTP information is enabled, but no servers are | |
11697 | configured, the corresponding uplink information (if there | |
11698 | is any) is propagated. | |
11699 | ||
11700 | - Server and client now support transmission and reception | |
11701 | of timezone information. It can be configured via the | |
11702 | newly introduced network options UseTimezone=, | |
33db1b90 | 11703 | EmitTimezone=, and Timezone=. Transmission of timezone |
5e8d4254 LP |
11704 | information is enabled between host and containers by |
11705 | default now: the container will change its local timezone | |
11706 | to what the host has set. | |
11707 | ||
11708 | - Lease timeouts can now be configured via | |
11709 | MaxLeaseTimeSec= and DefaultLeaseTimeSec=. | |
11710 | ||
11711 | - The DHCP server improved on the stability of | |
11712 | leases. Clients are more likely to get the same lease | |
11713 | information back, even if the server loses state. | |
11714 | ||
11715 | - The DHCP server supports two new configuration options to | |
11716 | control the lease address pool metrics, PoolOffset= and | |
11717 | PoolSize=. | |
11718 | ||
11719 | * The encapsulation limit of tunnels in systemd-networkd may | |
11720 | now be configured via 'EncapsulationLimit='. It allows | |
11721 | modifying the maximum additional levels of encapsulation | |
11722 | that are permitted to be prepended to a packet. | |
11723 | ||
11724 | * systemd now supports the concept of user buses replacing | |
11725 | session buses, if used with dbus-1.10 (and enabled via dbus | |
11726 | --enable-user-session). It previously only supported this on | |
11727 | kdbus-enabled systems, and this release expands this to | |
11728 | 'dbus-daemon' systems. | |
11729 | ||
11730 | * systemd-networkd now supports predictable interface names | |
11731 | for virtio devices. | |
11732 | ||
11733 | * systemd now optionally supports the new Linux kernel | |
11734 | "unified" control group hierarchy. If enabled via the kernel | |
11735 | command-line option 'systemd.unified_cgroup_hierarchy=1', | |
11736 | systemd will try to mount the unified cgroup hierarchy | |
11737 | directly on /sys/fs/cgroup. If not enabled, or not | |
11738 | available, systemd will fall back to the legacy cgroup | |
11739 | hierarchy setup, as before. Host system and containers can | |
11740 | mix and match legacy and unified hierarchies as they | |
856ca72b | 11741 | wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY |
5e8d4254 LP |
11742 | environment variable to individually select the hierarchy to |
11743 | use for executed containers. By default, nspawn will use the | |
11744 | unified hierarchy for the containers if the host uses the | |
11745 | unified hierarchy, and the legacy hierarchy otherwise. | |
11746 | Please note that at this point the unified hierarchy is an | |
11747 | experimental kernel feature and is likely to change in one | |
33db1b90 | 11748 | of the next kernel releases. Therefore, it should not be |
5e8d4254 LP |
11749 | enabled by default in downstream distributions yet. The |
11750 | minimum required kernel version for the unified hierarchy to | |
11751 | work is 4.2. Note that when the unified hierarchy is used | |
11752 | for the first time delegated access to controllers is | |
11753 | safe. Because of this systemd-nspawn containers will get | |
11754 | access to controllers now, as will systemd user | |
11755 | sessions. This means containers and user sessions may now | |
11756 | manage their own resources, partitioning up what the system | |
11757 | grants them. | |
11758 | ||
11759 | * A new special scope unit "init.scope" has been introduced | |
11760 | that encapsulates PID 1 of the system. It may be used to | |
11761 | determine resource usage and enforce resource limits on PID | |
11762 | 1 itself. PID 1 hence moved out of the root of the control | |
11763 | group tree. | |
11764 | ||
11765 | * The cgtop tool gained support for filtering out kernel | |
11766 | threads when counting tasks in a control group. Also, the | |
11767 | count of processes is now recursively summed up by | |
11768 | default. Two options -k and --recursive= have been added to | |
11769 | revert to old behaviour. The tool has also been updated to | |
11770 | work correctly in containers now. | |
11771 | ||
11772 | * systemd-nspawn's --bind= and --bind-ro= options have been | |
11773 | extended to allow creation of non-recursive bind mounts. | |
11774 | ||
c626bf1d DM |
11775 | * libsystemd gained two new calls sd_pid_get_cgroup() and |
11776 | sd_peer_get_cgroup() which return the control group path of | |
5e8d4254 LP |
11777 | a process or peer of a connected AF_UNIX socket. This |
11778 | function call is particularly useful when implementing | |
11779 | delegated subtrees support in the control group hierarchy. | |
11780 | ||
11781 | * The "sd-event" event loop API of libsystemd now supports | |
11782 | correct dequeuing of real-time signals, without losing | |
11783 | signal events. | |
11784 | ||
d35f51ea ZJS |
11785 | * When systemd requests a polkit decision when managing units it |
11786 | will now add additional fields to the request, including unit | |
11787 | name and desired operation. This enables more powerful polkit | |
11788 | policies, that make decisions depending on these parameters. | |
c9912c5e | 11789 | |
47f5a38c LP |
11790 | * nspawn learnt support for .nspawn settings files, that may |
11791 | accompany the image files or directories of containers, and | |
11792 | may contain additional settings for the container. This is | |
11793 | an alternative to configuring container parameters via the | |
11794 | nspawn command line. | |
11795 | ||
2f77decc LP |
11796 | Contributions from: Cristian Rodríguez, Daniel Mack, David |
11797 | Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe | |
11798 | Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan | |
11799 | Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel | |
11800 | Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal | |
11801 | Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin | |
11802 | Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel | |
23d08d1b | 11803 | Andersen, Tom Gundersen, Torstein Husebø |
c9912c5e | 11804 | |
ccddd104 | 11805 | — Berlin, 2015-09-08 |
c9912c5e | 11806 | |
ec5249a2 DM |
11807 | CHANGES WITH 225: |
11808 | ||
5e8d4254 LP |
11809 | * machinectl gained a new verb 'shell' which opens a fresh |
11810 | shell on the target container or the host. It is similar to | |
11811 | the existing 'login' command of machinectl, but spawns the | |
11812 | shell directly without prompting for username or | |
11813 | password. The pseudo machine '.host' now refers to the local | |
11814 | host and is used by default. Hence, 'machinectl shell' can | |
11815 | be used as replacement for 'su -' which spawns a session as | |
11816 | a fresh systemd unit in a way that is fully isolated from | |
11817 | the originating session. | |
11818 | ||
11819 | * systemd-networkd learned to cope with private-zone DHCP | |
11820 | options and allows other programs to query the values. | |
11821 | ||
11822 | * SELinux access control when enabling/disabling units is no | |
d35f51ea ZJS |
11823 | longer enforced with this release. The previous implementation |
11824 | was incorrect, and a new corrected implementation is not yet | |
11825 | available. As unit file operations are still protected via | |
11826 | polkit and D-Bus policy this is not a security problem. Yet, | |
11827 | distributions which care about optimal SELinux support should | |
11828 | probably not stabilize on this release. | |
5e8d4254 LP |
11829 | |
11830 | * sd-bus gained support for matches of type "arg0has=", that | |
11831 | test for membership of strings in string arrays sent in bus | |
11832 | messages. | |
11833 | ||
11834 | * systemd-resolved now dumps the contents of its DNS and LLMNR | |
11835 | caches to the logs on reception of the SIGUSR1 signal. This | |
11836 | is useful to debug DNS behaviour. | |
11837 | ||
11838 | * The coredumpctl tool gained a new --directory= option to | |
11839 | operate on journal files in a specific directory. | |
11840 | ||
11841 | * "systemctl reboot" and related commands gained a new | |
11842 | "--message=" option which may be used to set a free-text | |
11843 | wall message when shutting down or rebooting the | |
11844 | system. This message is also logged, which is useful for | |
11845 | figuring out the reason for a reboot or shutdown a | |
11846 | posteriori. | |
11847 | ||
11848 | * The "systemd-resolve-host" tool's -i switch now takes | |
11849 | network interface numbers as alternative to interface names. | |
11850 | ||
11851 | * A new unit file setting for services has been introduced: | |
11852 | UtmpMode= allows configuration of how precisely systemd | |
11853 | handles utmp and wtmp entries for the service if this is | |
11854 | enabled. This allows writing services that appear similar to | |
11855 | user sessions in the output of the "w", "who", "last" and | |
11856 | "lastlog" tools. | |
11857 | ||
11858 | * systemd-resolved will now locally synthesize DNS resource | |
11859 | records for the "localhost" and "gateway" domains as well as | |
11860 | the local hostname. This should ensure that clients querying | |
11861 | RRs via resolved will get similar results as those going via | |
11862 | NSS, if nss-myhostname is enabled. | |
11863 | ||
11864 | Contributions from: Alastair Hughes, Alex Crawford, Daniel | |
11865 | Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski, | |
11866 | Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan | |
11867 | Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers, | |
11868 | Kefeng Wang, Lennart Poettering, Major Hayden, Marcel | |
11869 | Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt | |
11870 | Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim, | |
11871 | Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer, | |
11872 | reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings, | |
11873 | Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe | |
11874 | Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts, | |
11875 | WaLyong Cho, Zbigniew Jędrzejewski-Szmek | |
e1439a14 | 11876 | |
ccddd104 | 11877 | — Berlin, 2015-08-27 |
ec5249a2 | 11878 | |
11811e85 DH |
11879 | CHANGES WITH 224: |
11880 | ||
10fa421c DH |
11881 | * The systemd-efi-boot-generator functionality was merged into |
11882 | systemd-gpt-auto-generator. | |
11883 | ||
5e8d4254 LP |
11884 | * systemd-networkd now supports Group Policy for vxlan |
11885 | devices. It can be enabled via the new boolean configuration | |
11886 | option called 'GroupPolicyExtension='. | |
10fa421c | 11887 | |
11811e85 DH |
11888 | Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David |
11889 | Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart | |
11890 | Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen | |
11891 | ||
ccddd104 | 11892 | — Berlin, 2015-07-31 |
11811e85 | 11893 | |
e57eaef8 DH |
11894 | CHANGES WITH 223: |
11895 | ||
11896 | * The python-systemd code has been removed from the systemd repository. | |
11897 | A new repository has been created which accommodates the code from | |
11898 | now on, and we kindly ask distributions to create a separate package | |
11899 | for this: https://github.com/systemd/python-systemd | |
11900 | ||
01608bc8 | 11901 | * The systemd daemon will now reload its main configuration |
e57eaef8 DH |
11902 | (/etc/systemd/system.conf) on daemon-reload. |
11903 | ||
11904 | * sd-dhcp now exposes vendor specific extensions via | |
11905 | sd_dhcp_lease_get_vendor_specific(). | |
11906 | ||
931618d0 DM |
11907 | * systemd-networkd gained a number of new configuration options. |
11908 | ||
11909 | - A new boolean configuration option for TAP devices called | |
37d54b93 | 11910 | 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the |
931618d0 DM |
11911 | device, thus allowing to send and receive GSO packets. |
11912 | ||
11913 | - A new tunnel configuration option called 'CopyDSCP='. | |
11914 | If enabled, the DSCP field of ip6 tunnels is copied into the | |
11915 | decapsulated packet. | |
11916 | ||
11917 | - A set of boolean bridge configuration options were added. | |
11918 | 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=', | |
11919 | and 'UnicastFlood=' are now parsed by networkd and applied to the | |
11920 | respective bridge link device via the respective IFLA_BRPORT_* | |
11921 | netlink attribute. | |
11922 | ||
11923 | - A new string configuration option to override the hostname sent | |
11924 | to a DHCP server, called 'Hostname='. If set and 'SendHostname=' | |
11925 | is true, networkd will use the configured hostname instead of the | |
11926 | system hostname when sending DHCP requests. | |
11927 | ||
11928 | - A new tunnel configuration option called 'IPv6FlowLabel='. If set, | |
11929 | networkd will configure the IPv6 flow-label of the tunnel device | |
11930 | according to RFC2460. | |
e57eaef8 | 11931 | |
f5f113f6 DH |
11932 | - The 'macvtap' virtual network devices are now supported, similar to |
11933 | the already supported 'macvlan' devices. | |
11934 | ||
e57eaef8 | 11935 | * systemd-resolved now implements RFC5452 to improve resilience against |
01608bc8 | 11936 | cache poisoning. Additionally, source port randomization is enabled |
e57eaef8 DH |
11937 | by default to further protect against DNS spoofing attacks. |
11938 | ||
11939 | * nss-mymachines now supports translating UIDs and GIDs of running | |
11940 | containers with user-namespaces enabled. If a container 'foo' | |
11941 | translates a host uid 'UID' to the container uid 'TUID', then | |
11942 | nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID' | |
11943 | (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are | |
11944 | mapped as 'vg-foo-TGID'. | |
11945 | ||
11946 | Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel | |
e4e66993 DH |
11947 | Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov, |
11948 | HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig), | |
11949 | Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers, | |
11950 | Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael | |
11951 | Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim, | |
11952 | Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo, | |
11953 | Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom | |
11954 | Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo, | |
11955 | Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek | |
11956 | ||
ccddd104 | 11957 | — Berlin, 2015-07-29 |
e57eaef8 | 11958 | |
0db83ad7 | 11959 | CHANGES WITH 222: |
5541c889 | 11960 | |
861b02eb KS |
11961 | * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules. |
11962 | There are no known issues with current sysfs, and udev does not need | |
11963 | or should be used to work around such bugs. | |
11964 | ||
11965 | * udev does no longer enable USB HID power management. Several reports | |
11966 | indicate, that some devices cannot handle that setting. | |
0db83ad7 DH |
11967 | |
11968 | * The udev accelerometer helper was removed. The functionality | |
11969 | is now fully included in iio-sensor-proxy. But this means, | |
11970 | older iio-sensor-proxy versions will no longer provide | |
11971 | accelerometer/orientation data with this systemd version. | |
11972 | Please upgrade iio-sensor-proxy to version 1.0. | |
11973 | ||
5541c889 DH |
11974 | * networkd gained a new configuration option IPv6PrivacyExtensions= |
11975 | which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions | |
11976 | for Stateless Address") on selected networks. | |
11977 | ||
9b361114 DM |
11978 | * For the sake of fewer build-time dependencies and less code in the |
11979 | main repository, the python bindings are about to be removed in the | |
11980 | next release. A new repository has been created which accommodates | |
11981 | the code from now on, and we kindly ask distributions to create a | |
11982 | separate package for this. The removal will take place in v223. | |
11983 | ||
11984 | https://github.com/systemd/python-systemd | |
11985 | ||
0db83ad7 DH |
11986 | Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera, |
11987 | Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack, | |
11988 | daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric | |
11989 | Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario, | |
5541c889 DH |
11990 | Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens |
11991 | (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering, | |
0db83ad7 DH |
11992 | Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal |
11993 | Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne, | |
2d1ca112 DH |
11994 | Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein |
11995 | Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek | |
0db83ad7 | 11996 | |
ccddd104 | 11997 | — Berlin, 2015-07-07 |
0db83ad7 | 11998 | |
0f0467e6 MP |
11999 | CHANGES WITH 221: |
12000 | ||
470e72d4 | 12001 | * The sd-bus.h and sd-event.h APIs have now been declared |
5f92d24f | 12002 | stable and have been added to the official interface of |
470e72d4 LP |
12003 | libsystemd.so. sd-bus implements an alternative D-Bus client |
12004 | library, that is relatively easy to use, very efficient and | |
12005 | supports both classic D-Bus as well as kdbus as transport | |
12006 | backend. sd-event is a generic event loop abstraction that | |
12007 | is built around Linux epoll, but adds features such as event | |
0aee49d5 | 12008 | prioritization or efficient timer handling. Both APIs are good |
470e72d4 LP |
12009 | choices for C programs looking for a bus and/or event loop |
12010 | implementation that is minimal and does not have to be | |
5f92d24f | 12011 | portable to other kernels. |
0f0467e6 | 12012 | |
470e72d4 LP |
12013 | * kdbus support is no longer compile-time optional. It is now |
12014 | always built-in. However, it can still be disabled at | |
12015 | runtime using the kdbus=0 kernel command line setting, and | |
c6551464 | 12016 | that setting may be changed to default to off, by specifying |
470e72d4 LP |
12017 | --disable-kdbus at build-time. Note though that the kernel |
12018 | command line setting has no effect if the kdbus.ko kernel | |
12019 | module is not installed, in which case kdbus is (obviously) | |
12020 | also disabled. We encourage all downstream distributions to | |
0aee49d5 | 12021 | begin testing kdbus by adding it to the kernel images in the |
470e72d4 LP |
12022 | development distributions, and leaving kdbus support in |
12023 | systemd enabled. | |
0f0467e6 | 12024 | |
470e72d4 LP |
12025 | * The minimal required util-linux version has been bumped to |
12026 | 2.26. | |
12027 | ||
12028 | * Support for chkconfig (--enable-chkconfig) was removed in | |
0aee49d5 | 12029 | favor of calling an abstraction tool |
470e72d4 LP |
12030 | /lib/systemd/systemd-sysv-install. This needs to be |
12031 | implemented for your distribution. See "SYSV INIT.D SCRIPTS" | |
12032 | in README for details. | |
12033 | ||
12034 | * If there's a systemd unit and a SysV init script for the | |
12035 | same service name, and the user executes "systemctl enable" | |
12036 | for it (or a related call), then this will now enable both | |
12037 | (or execute the related operation on both), not just the | |
12038 | unit. | |
12039 | ||
12040 | * The libudev API documentation has been converted from gtkdoc | |
12041 | into man pages. | |
12042 | ||
12043 | * gudev has been removed from the systemd tree, it is now an | |
12044 | external project. | |
12045 | ||
12046 | * The systemd-cgtop tool learnt a new --raw switch to generate | |
0aee49d5 | 12047 | "raw" (machine parsable) output. |
470e72d4 LP |
12048 | |
12049 | * networkd's IPForwarding= .network file setting learnt the | |
12050 | new setting "kernel", which ensures that networkd does not | |
12051 | change the IP forwarding sysctl from the default kernel | |
12052 | state. | |
12053 | ||
12054 | * The systemd-logind bus API now exposes a new boolean | |
12055 | property "Docked" that reports whether logind considers the | |
12056 | system "docked", i.e. connected to a docking station or not. | |
12057 | ||
12058 | Contributions from: Alex Crawford, Andreas Pokorny, Andrei | |
12059 | Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez, | |
12060 | Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann, | |
12061 | David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed | |
12062 | Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario, | |
12063 | Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek, | |
12064 | Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang, | |
12065 | Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart | |
12066 | Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario | |
12067 | Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich, | |
12068 | Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes, | |
b912e251 LP |
12069 | Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip |
12070 | Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani, | |
12071 | Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein | |
12072 | Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner | |
12073 | Fink, Zbigniew Jędrzejewski-Szmek | |
470e72d4 | 12074 | |
ccddd104 | 12075 | — Berlin, 2015-06-19 |
0f0467e6 | 12076 | |
481a0aa2 LP |
12077 | CHANGES WITH 220: |
12078 | ||
f7a73a25 DH |
12079 | * The gudev library has been extracted into a separate repository |
12080 | available at: https://git.gnome.org/browse/libgudev/ | |
12081 | It is now managed as part of the Gnome project. Distributions | |
12082 | are recommended to pass --disable-gudev to systemd and use | |
12083 | gudev from the Gnome project instead. gudev is still included | |
12084 | in systemd, for now. It will be removed soon, though. Please | |
12085 | also see the announcement-thread on systemd-devel: | |
56cadcb6 | 12086 | https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html |
f7a73a25 | 12087 | |
481a0aa2 LP |
12088 | * systemd now exposes a CPUUsageNSec= property for each |
12089 | service unit on the bus, that contains the overall consumed | |
12090 | CPU time of a service (the sum of what each process of the | |
12091 | service consumed). This value is only available if | |
12092 | CPUAccounting= is turned on for a service, and is then shown | |
12093 | in the "systemctl status" output. | |
12094 | ||
12095 | * Support for configuring alternative mappings of the old SysV | |
12096 | runlevels to systemd targets has been removed. They are now | |
29d1fcb4 | 12097 | hardcoded in a way that runlevels 2, 3, 4 all map to |
481a0aa2 LP |
12098 | multi-user.target and 5 to graphical.target (which |
12099 | previously was already the default behaviour). | |
12100 | ||
12101 | * The auto-mounter logic gained support for mount point | |
12102 | expiry, using a new TimeoutIdleSec= setting in .automount | |
12103 | units. (Also available as x-systemd.idle-timeout= in /etc/fstab). | |
12104 | ||
12105 | * The EFI System Partition (ESP) as mounted to /boot by | |
12106 | systemd-efi-boot-generator will now be unmounted | |
29d1fcb4 | 12107 | automatically after 2 minutes of not being used. This should |
481a0aa2 LP |
12108 | minimize the risk of ESP corruptions. |
12109 | ||
12110 | * New /etc/fstab options x-systemd.requires= and | |
12111 | x-systemd.requires-mounts-for= are now supported to express | |
12112 | additional dependencies for mounts. This is useful for | |
28423d9a | 12113 | journaling file systems that support external journal |
481a0aa2 LP |
12114 | devices or overlay file systems that require underlying file |
12115 | systems to be mounted. | |
12116 | ||
12117 | * systemd does not support direct live-upgrades (via systemctl | |
12118 | daemon-reexec) from versions older than v44 anymore. As no | |
12119 | distribution we are aware of shipped such old versions in a | |
12120 | stable release this should not be problematic. | |
12121 | ||
12122 | * When systemd forks off a new per-connection service instance | |
12123 | it will now set the $REMOTE_ADDR environment variable to the | |
12124 | remote IP address, and $REMOTE_PORT environment variable to | |
12125 | the remote IP port. This behaviour is similar to the | |
12126 | corresponding environment variables defined by CGI. | |
12127 | ||
12128 | * systemd-networkd gained support for uplink failure | |
12129 | detection. The BindCarrier= option allows binding interface | |
12130 | configuration dynamically to the link sense of other | |
12131 | interfaces. This is useful to achieve behaviour like in | |
12132 | network switches. | |
12133 | ||
12134 | * systemd-networkd gained support for configuring the DHCP | |
12135 | client identifier to use when requesting leases. | |
12136 | ||
12137 | * systemd-networkd now has a per-network UseNTP= option to | |
12138 | configure whether NTP server information acquired via DHCP | |
12139 | is passed on to services like systemd-timesyncd. | |
12140 | ||
12141 | * systemd-networkd gained support for vti6 tunnels. | |
12142 | ||
1579dd2c LP |
12143 | * Note that systemd-networkd manages the sysctl variable |
12144 | /proc/sys/net/ipv[46]/conf/*/forwarding for each interface | |
12145 | it is configured for since v219. The variable controls IP | |
12146 | forwarding, and is a per-interface alternative to the global | |
12147 | /proc/sys/net/ipv[46]/ip_forward. This setting is | |
12148 | configurable in the IPForward= option, which defaults to | |
12149 | "no". This means if networkd is used for an interface it is | |
12150 | no longer sufficient to set the global sysctl option to turn | |
12151 | on IP forwarding! Instead, the .network file option | |
12152 | IPForward= needs to be turned on! Note that the | |
12153 | implementation of this behaviour was broken in v219 and has | |
12154 | been fixed in v220. | |
12155 | ||
481a0aa2 LP |
12156 | * Many bonding and vxlan options are now configurable in |
12157 | systemd-networkd. | |
12158 | ||
12159 | * systemd-nspawn gained a new --property= setting to set unit | |
12160 | properties for the container scope. This is useful for | |
ce830873 | 12161 | setting resource parameters (e.g. "CPUShares=500") on |
481a0aa2 LP |
12162 | containers started from the command line. |
12163 | ||
12164 | * systemd-nspawn gained a new --private-users= switch to make | |
12165 | use of user namespacing available on recent Linux kernels. | |
12166 | ||
12167 | * systemd-nspawn may now be called as part of a shell pipeline | |
12168 | in which case the pipes used for stdin and stdout are passed | |
12169 | directly to the process invoked in the container, without | |
12170 | indirection via a pseudo tty. | |
12171 | ||
12172 | * systemd-nspawn gained a new switch to control the UNIX | |
12173 | signal to use when killing the init process of the container | |
12174 | when shutting down. | |
12175 | ||
12176 | * systemd-nspawn gained a new --overlay= switch for mounting | |
12177 | overlay file systems into the container using the new kernel | |
12178 | overlayfs support. | |
12179 | ||
12180 | * When a container image is imported via systemd-importd and | |
12181 | the host file system is not btrfs, a loopback block device | |
12182 | file is created in /var/lib/machines.raw with a btrfs file | |
12183 | system inside. It is then mounted to /var/lib/machines to | |
12184 | enable btrfs features for container management. The loopback | |
12185 | file and btrfs file system is grown as needed when container | |
12186 | images are imported via systemd-importd. | |
12187 | ||
12188 | * systemd-machined/systemd-importd gained support for btrfs | |
12189 | quota, to enforce container disk space limits on disk. This | |
12190 | is exposed in "machinectl set-limit". | |
12191 | ||
12192 | * systemd-importd now can import containers from local .tar, | |
12193 | .raw and .qcow2 images, and export them to .tar and .raw. It | |
12194 | can also import dkr v2 images now from the network (on top | |
12195 | of v1 as before). | |
12196 | ||
12197 | * systemd-importd gained support for verifying downloaded | |
12198 | images with gpg2 (previously only gpg1 was supported). | |
12199 | ||
d35f51ea ZJS |
12200 | * systemd-machined, systemd-logind, systemd: most bus calls are |
12201 | now accessible to unprivileged processes via polkit. Also, | |
12202 | systemd-logind will now allow users to kill their own sessions | |
12203 | without further privileges or authorization. | |
481a0aa2 LP |
12204 | |
12205 | * systemd-shutdownd has been removed. This service was | |
12206 | previously responsible for implementing scheduled shutdowns | |
12207 | as exposed in /usr/bin/shutdown's time parameter. This | |
12208 | functionality has now been moved into systemd-logind and is | |
12209 | accessible via a bus interface. | |
12210 | ||
12211 | * "systemctl reboot" gained a new switch --firmware-setup that | |
12212 | can be used to reboot into the EFI firmware setup, if that | |
12213 | is available. systemd-logind now exposes an API on the bus | |
12214 | to trigger such reboots, in case graphical desktop UIs want | |
12215 | to cover this functionality. | |
12216 | ||
12217 | * "systemctl enable", "systemctl disable" and "systemctl mask" | |
1579dd2c | 12218 | now support a new "--now" switch. If specified the units |
481a0aa2 LP |
12219 | that are enabled will also be started, and the ones |
12220 | disabled/masked also stopped. | |
12221 | ||
12222 | * The Gummiboot EFI boot loader tool has been merged into | |
1a2d5fbe DH |
12223 | systemd, and renamed to "systemd-boot". The bootctl tool has been |
12224 | updated to support systemd-boot. | |
481a0aa2 LP |
12225 | |
12226 | * An EFI kernel stub has been added that may be used to create | |
12227 | kernel EFI binaries that contain not only the actual kernel, | |
12228 | but also an initrd, boot splash, command line and OS release | |
12229 | information. This combined binary can then be signed as a | |
12230 | single image, so that the firmware can verify it all in one | |
1a2d5fbe | 12231 | step. systemd-boot has special support for EFI binaries created |
481a0aa2 LP |
12232 | like this and can extract OS release information from them |
12233 | and show them in the boot menu. This functionality is useful | |
12234 | to implement cryptographically verified boot schemes. | |
12235 | ||
12236 | * Optional support has been added to systemd-fsck to pass | |
12237 | fsck's progress report to an AF_UNIX socket in the file | |
12238 | system. | |
12239 | ||
6b000af4 LP |
12240 | * udev will no longer create device symlinks for all block devices by |
12241 | default. A deny list for excluding special block devices from this | |
387f6955 | 12242 | logic has been turned into an allow list that requires picking block |
6b000af4 | 12243 | devices explicitly that require device symlinks. |
481a0aa2 LP |
12244 | |
12245 | * A new (currently still internal) API sd-device.h has been | |
12246 | added to libsystemd. This modernized API is supposed to | |
12247 | replace libudev eventually. In fact, already much of libudev | |
12248 | is now just a wrapper around sd-device.h. | |
12249 | ||
12250 | * A new hwdb database for storing metadata about pointing | |
12251 | stick devices has been added. | |
12252 | ||
12253 | * systemd-tmpfiles gained support for setting file attributes | |
12254 | similar to the "chattr" tool with new 'h' and 'H' lines. | |
12255 | ||
12256 | * systemd-journald will no longer unconditionally set the | |
12257 | btrfs NOCOW flag on new journal files. This is instead done | |
12258 | with tmpfiles snippet using the new 'h' line type. This | |
12259 | allows easy disabling of this logic, by masking the | |
12260 | journal-nocow.conf tmpfiles file. | |
12261 | ||
12262 | * systemd-journald will now translate audit message types to | |
12263 | human readable identifiers when writing them to the | |
12264 | journal. This should improve readability of audit messages. | |
12265 | ||
12266 | * The LUKS logic gained support for the offset= and skip= | |
12267 | options in /etc/crypttab, as previously implemented by | |
12268 | Debian. | |
12269 | ||
12270 | * /usr/lib/os-release gained a new optional field VARIANT= for | |
12271 | distributions that support multiple variants (such as a | |
1d3a473b | 12272 | desktop edition, a server edition, …) |
481a0aa2 LP |
12273 | |
12274 | Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy, | |
12275 | Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin | |
12276 | Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel, | |
12277 | Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž | |
12278 | Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian | |
12279 | Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel | |
12280 | Mack, Daniel Mustieles, daurnimator, Davide Bettio, David | |
12281 | Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov, | |
12282 | Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke, | |
12283 | Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López | |
12284 | Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan | |
12285 | Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John | |
12286 | Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay | |
12287 | Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas | |
12288 | De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz | |
12289 | Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel | |
12290 | Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett, | |
12291 | Michael Biebl, Michael Marineau, Michael Olbrich, Michal | |
12292 | Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik | |
12293 | Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter | |
12294 | Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny | |
12295 | Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick, | |
12296 | Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker, | |
12297 | Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas | |
12298 | Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom | |
12299 | Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will | |
12300 | Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek | |
12301 | ||
ccddd104 | 12302 | — Berlin, 2015-05-22 |
481a0aa2 | 12303 | |
615aaf41 LP |
12304 | CHANGES WITH 219: |
12305 | ||
615aaf41 LP |
12306 | * Introduce a new API "sd-hwdb.h" for querying the hardware |
12307 | metadata database. With this minimal interface one can query | |
12308 | and enumerate the udev hwdb, decoupled from the old libudev | |
12309 | library. libudev's interface for this is now only a wrapper | |
12310 | around sd-hwdb. A new tool systemd-hwdb has been added to | |
12311 | interface with and update the database. | |
12312 | ||
12313 | * When any of systemd's tools copies files (for example due to | |
12314 | tmpfiles' C lines) a btrfs reflink will attempted first, | |
12315 | before bytewise copying is done. | |
12316 | ||
12317 | * systemd-nspawn gained a new --ephemeral switch. When | |
12318 | specified a btrfs snapshot is taken of the container's root | |
12319 | directory, and immediately removed when the container | |
12320 | terminates again. Thus, a container can be started whose | |
12321 | changes never alter the container's root directory, and are | |
12322 | lost on container termination. This switch can also be used | |
12323 | for starting a container off the root file system of the | |
12324 | host without affecting the host OS. This switch is only | |
12325 | available on btrfs file systems. | |
12326 | ||
12327 | * systemd-nspawn gained a new --template= switch. It takes the | |
12328 | path to a container tree to use as template for the tree | |
7edecf21 | 12329 | specified via --directory=, should that directory be |
615aaf41 LP |
12330 | missing. This allows instantiating containers dynamically, |
12331 | on first run. This switch is only available on btrfs file | |
12332 | systems. | |
12333 | ||
12334 | * When a .mount unit refers to a mount point on which multiple | |
12335 | mounts are stacked, and the .mount unit is stopped all of | |
12336 | the stacked mount points will now be unmounted until no | |
12337 | mount point remains. | |
12338 | ||
12339 | * systemd now has an explicit notion of supported and | |
12340 | unsupported unit types. Jobs enqueued for unsupported unit | |
12341 | types will now fail with an "unsupported" error code. More | |
12342 | specifically .swap, .automount and .device units are not | |
12343 | supported in containers, .busname units are not supported on | |
12344 | non-kdbus systems. .swap and .automount are also not | |
12345 | supported if their respective kernel compile time options | |
12346 | are disabled. | |
12347 | ||
12348 | * machinectl gained support for two new "copy-from" and | |
12349 | "copy-to" commands for copying files from a running | |
12350 | container to the host or vice versa. | |
12351 | ||
12352 | * machinectl gained support for a new "bind" command to bind | |
12353 | mount host directories into local containers. This is | |
12354 | currently only supported for nspawn containers. | |
12355 | ||
12356 | * networkd gained support for configuring bridge forwarding | |
12357 | database entries (fdb) from .network files. | |
12358 | ||
12359 | * A new tiny daemon "systemd-importd" has been added that can | |
12360 | download container images in tar, raw, qcow2 or dkr formats, | |
12361 | and make them available locally in /var/lib/machines, so | |
12362 | that they can run as nspawn containers. The daemon can GPG | |
12363 | verify the downloads (not supported for dkr, since it has no | |
12364 | provisions for verifying downloads). It will transparently | |
12365 | decompress bz2, xz, gzip compressed downloads if necessary, | |
12366 | and restore sparse files on disk. The daemon uses privilege | |
12367 | separation to ensure the actual download logic runs with | |
94e5ba37 | 12368 | fewer privileges than the daemon itself. machinectl has |
615aaf41 LP |
12369 | gained new commands "pull-tar", "pull-raw" and "pull-dkr" to |
12370 | make the functionality of importd available to the | |
12371 | user. With this in place the Fedora and Ubuntu "Cloud" | |
12372 | images can be downloaded and booted as containers unmodified | |
12373 | (the Fedora images lack the appropriate GPG signature files | |
12374 | currently, so they cannot be verified, but this will change | |
12375 | soon, hopefully). Note that downloading images is currently | |
12376 | only fully supported on btrfs. | |
12377 | ||
12378 | * machinectl is now able to list container images found in | |
12379 | /var/lib/machines, along with some metadata about sizes of | |
12380 | disk and similar. If the directory is located on btrfs and | |
12381 | quota is enabled, this includes quota display. A new command | |
12382 | "image-status" has been added that shows additional | |
12383 | information about images. | |
12384 | ||
12385 | * machinectl is now able to clone container images | |
12386 | efficiently, if the underlying file system (btrfs) supports | |
f59dba26 | 12387 | it, with the new "machinectl clone" command. It also |
615aaf41 LP |
12388 | gained commands for renaming and removing images, as well as |
12389 | marking them read-only or read-write (supported also on | |
12390 | legacy file systems). | |
12391 | ||
12392 | * networkd gained support for collecting LLDP network | |
12393 | announcements, from hardware that supports this. This is | |
12394 | shown in networkctl output. | |
12395 | ||
12396 | * systemd-run gained support for a new -t (--pty) switch for | |
12397 | invoking a binary on a pty whose input and output is | |
12398 | connected to the invoking terminal. This allows executing | |
12399 | processes as system services while interactively | |
12400 | communicating with them via the terminal. Most interestingly | |
12401 | this is supported across container boundaries. Invoking | |
12402 | "systemd-run -t /bin/bash" is an alternative to running a | |
12403 | full login session, the difference being that the former | |
12404 | will not register a session, nor go through the PAM session | |
12405 | setup. | |
12406 | ||
12407 | * tmpfiles gained support for a new "v" line type for creating | |
12408 | btrfs subvolumes. If the underlying file system is a legacy | |
12409 | file system, this automatically degrades to creating a | |
12410 | normal directory. Among others /var/lib/machines is now | |
12411 | created like this at boot, should it be missing. | |
12412 | ||
12413 | * The directory /var/lib/containers/ has been deprecated and | |
12414 | been replaced by /var/lib/machines. The term "machines" has | |
12415 | been used in the systemd context as generic term for both | |
12416 | VMs and containers, and hence appears more appropriate for | |
12417 | this, as the directory can also contain raw images bootable | |
12418 | via qemu/kvm. | |
12419 | ||
12420 | * systemd-nspawn when invoked with -M but without --directory= | |
12421 | or --image= is now capable of searching for the container | |
12422 | root directory, subvolume or disk image automatically, in | |
12423 | /var/lib/machines. systemd-nspawn@.service has been updated | |
12424 | to make use of this, thus allowing it to be used for raw | |
12425 | disk images, too. | |
12426 | ||
12427 | * A new machines.target unit has been introduced that is | |
12428 | supposed to group all containers/VMs invoked as services on | |
12429 | the system. systemd-nspawn@.service has been updated to | |
12430 | integrate with that. | |
12431 | ||
12432 | * machinectl gained a new "start" command, for invoking a | |
12433 | container as a service. "machinectl start foo" is mostly | |
12434 | equivalent to "systemctl start systemd-nspawn@foo.service", | |
12435 | but handles escaping in a nicer way. | |
12436 | ||
12437 | * systemd-nspawn will now mount most of the cgroupfs tree | |
12438 | read-only into each container, with the exception of the | |
12439 | container's own subtree in the name=systemd hierarchy. | |
12440 | ||
12441 | * journald now sets the special FS_NOCOW file flag for its | |
12442 | journal files. This should improve performance on btrfs, by | |
12443 | avoiding heavy fragmentation when journald's write-pattern | |
12444 | is used on COW file systems. It degrades btrfs' data | |
12445 | integrity guarantees for the files to the same levels as for | |
12446 | ext3/ext4 however. This should be OK though as journald does | |
12447 | its own data integrity checks and all its objects are | |
12448 | checksummed on disk. Also, journald should handle btrfs disk | |
12449 | full events a lot more gracefully now, by processing SIGBUS | |
12450 | errors, and not relying on fallocate() anymore. | |
12451 | ||
12452 | * When journald detects that journal files it is writing to | |
12453 | have been deleted it will immediately start new journal | |
12454 | files. | |
12455 | ||
12456 | * systemd now provides a way to store file descriptors | |
4c37970d | 12457 | per-service in PID 1. This is useful for daemons to ensure |
615aaf41 | 12458 | that fds they require are not lost during a daemon |
94e5ba37 | 12459 | restart. The fds are passed to the daemon on the next |
615aaf41 LP |
12460 | invocation in the same way socket activation fds are |
12461 | passed. This is now used by journald to ensure that the | |
12462 | various sockets connected to all the system's stdout/stderr | |
12463 | are not lost when journald is restarted. File descriptors | |
12464 | may be stored in PID 1 via the sd_pid_notify_with_fds() API, | |
12465 | an extension to sd_notify(). Note that a limit is enforced | |
12466 | on the number of fds a service can store in PID 1, and it | |
12467 | defaults to 0, so that no fds may be stored, unless this is | |
12468 | explicitly turned on. | |
12469 | ||
12470 | * The default TERM variable to use for units connected to a | |
12471 | terminal, when no other value is explicitly is set is now | |
12472 | vt220 rather than vt102. This should be fairly safe still, | |
12473 | but allows PgUp/PgDn work. | |
12474 | ||
12475 | * The /etc/crypttab option header= as known from Debian is now | |
12476 | supported. | |
12477 | ||
12478 | * "loginctl user-status" and "loginctl session-status" will | |
12479 | now show the last 10 lines of log messages of the | |
12480 | user/session following the status output. Similar, | |
12481 | "machinectl status" will show the last 10 log lines | |
12482 | associated with a virtual machine or container | |
12483 | service. (Note that this is usually not the log messages | |
12484 | done in the VM/container itself, but simply what the | |
12485 | container manager logs. For nspawn this includes all console | |
12486 | output however.) | |
12487 | ||
12488 | * "loginctl session-status" without further argument will now | |
12489 | show the status of the session of the caller. Similar, | |
12490 | "lock-session", "unlock-session", "activate", | |
12491 | "enable-linger", "disable-linger" may now be called without | |
12492 | session/user parameter in which case they apply to the | |
12493 | caller's session/user. | |
12494 | ||
12495 | * An X11 session scriptlet is now shipped that uploads | |
12496 | $DISPLAY and $XAUTHORITY into the environment of the systemd | |
12497 | --user daemon if a session begins. This should improve | |
12498 | compatibility with X11 enabled applications run as systemd | |
12499 | user services. | |
12500 | ||
12501 | * Generators are now subject to masking via /etc and /run, the | |
12502 | same way as unit files. | |
12503 | ||
12504 | * networkd .network files gained support for configuring | |
12505 | per-link IPv4/IPv6 packet forwarding as well as IPv4 | |
12506 | masquerading. This is by default turned on for veth links to | |
12507 | containers, as registered by systemd-nspawn. This means that | |
12508 | nspawn containers run with --network-veth will now get | |
12509 | automatic routed access to the host's networks without any | |
12510 | further configuration or setup, as long as networkd runs on | |
12511 | the host. | |
12512 | ||
12513 | * systemd-nspawn gained the --port= (-p) switch to expose TCP | |
12514 | or UDP posts of a container on the host. With this in place | |
12515 | it is possible to run containers with private veth links | |
12516 | (--network-veth), and have their functionality exposed on | |
12517 | the host as if their services were running directly on the | |
12518 | host. | |
12519 | ||
dd2fd155 | 12520 | * systemd-nspawn's --network-veth switch now gained a short |
615aaf41 LP |
12521 | version "-n", since with the changes above it is now truly |
12522 | useful out-of-the-box. The systemd-nspawn@.service has been | |
12523 | updated to make use of it too by default. | |
12524 | ||
12525 | * systemd-nspawn will now maintain a per-image R/W lock, to | |
12526 | ensure that the same image is not started more than once | |
12527 | writable. (It's OK to run an image multiple times | |
12528 | simultaneously in read-only mode.) | |
12529 | ||
12530 | * systemd-nspawn's --image= option is now capable of | |
12531 | dissecting and booting MBR and GPT disk images that contain | |
12532 | only a single active Linux partition. Previously it | |
12533 | supported only GPT disk images with proper GPT type | |
12534 | IDs. This allows running cloud images from major | |
12535 | distributions directly with systemd-nspawn, without | |
12536 | modification. | |
12537 | ||
12538 | * In addition to collecting mouse dpi data in the udev | |
12539 | hardware database, there's now support for collecting angle | |
12540 | information for mouse scroll wheels. The database is | |
7edecf21 | 12541 | supposed to guarantee similar scrolling behavior on mice |
615aaf41 LP |
12542 | that it knows about. There's also support for collecting |
12543 | information about Touchpad types. | |
12544 | ||
12545 | * udev's input_id built-in will now also collect touch screen | |
12546 | dimension data and attach it to probed devices. | |
12547 | ||
12548 | * /etc/os-release gained support for a Distribution Privacy | |
12549 | Policy link field. | |
12550 | ||
12551 | * networkd gained support for creating "ipvlan", "gretap", | |
12552 | "ip6gre", "ip6gretap" and "ip6tnl" network devices. | |
12553 | ||
12554 | * systemd-tmpfiles gained support for "a" lines for setting | |
12555 | ACLs on files. | |
12556 | ||
12557 | * systemd-nspawn will now mount /tmp in the container to | |
12558 | tmpfs, automatically. | |
12559 | ||
12560 | * systemd now exposes the memory.usage_in_bytes cgroup | |
12561 | attribute and shows it for each service in the "systemctl | |
12562 | status" output, if available. | |
12563 | ||
12564 | * When the user presses Ctrl-Alt-Del more than 7x within 2s an | |
12565 | immediate reboot is triggered. This useful if shutdown is | |
12566 | hung and is unable to complete, to expedite the | |
12567 | operation. Note that this kind of reboot will still unmount | |
12568 | all file systems, and hence should not result in fsck being | |
12569 | run on next reboot. | |
12570 | ||
12571 | * A .device unit for an optical block device will now be | |
12572 | considered active only when a medium is in the drive. Also, | |
12573 | mount units are now bound to their backing devices thus | |
12574 | triggering automatic unmounting when devices become | |
12575 | unavailable. With this in place systemd will now | |
12576 | automatically unmount left-over mounts when a CD-ROM is | |
5bc9ea07 | 12577 | ejected or a USB stick is yanked from the system. |
615aaf41 LP |
12578 | |
12579 | * networkd-wait-online now has support for waiting for | |
12580 | specific interfaces only (with globbing), and for giving up | |
12581 | after a configurable timeout. | |
12582 | ||
12583 | * networkd now exits when idle. It will be automatically | |
12584 | restarted as soon as interfaces show up, are removed or | |
12585 | change state. networkd will stay around as long as there is | |
12586 | at least one DHCP state machine or similar around, that keep | |
12587 | it non-idle. | |
12588 | ||
12589 | * networkd may now configure IPv6 link-local addressing in | |
12590 | addition to IPv4 link-local addressing. | |
12591 | ||
12592 | * The IPv6 "token" for use in SLAAC may now be configured for | |
12593 | each .network interface in networkd. | |
12594 | ||
12595 | * Routes configured with networkd may now be assigned a scope | |
12596 | in .network files. | |
12597 | ||
12598 | * networkd's [Match] sections now support globbing and lists | |
12599 | of multiple space-separated matches per item. | |
12600 | ||
11ea2781 | 12601 | Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser, |
d2c643c6 LP |
12602 | Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos |
12603 | Morata Castillo, Chris Atkinson, Chris J. Arges, Christian | |
12604 | Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie, | |
12605 | Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack, | |
12606 | Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald, | |
12607 | Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de | |
12608 | Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan | |
12609 | Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas | |
12610 | Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken | |
12611 | Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian, | |
12612 | Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas, | |
12613 | Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko | |
12614 | Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl, | |
12615 | Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas | |
11ea2781 LP |
12616 | Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul |
12617 | Martin, Peter Hutterer, Peter Mattern, Philippe De Swert, | |
12618 | Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny | |
12619 | Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick, | |
12620 | Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain | |
12621 | Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom | |
12622 | Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar | |
d2c643c6 LP |
12623 | Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland |
12624 | Hoffmann, Zbigniew Jędrzejewski-Szmek | |
11ea2781 | 12625 | |
ccddd104 | 12626 | — Berlin, 2015-02-16 |
11ea2781 | 12627 | |
d4f5a1f4 DH |
12628 | CHANGES WITH 218: |
12629 | ||
f9e00a9f LP |
12630 | * When querying unit file enablement status (for example via |
12631 | "systemctl is-enabled"), a new state "indirect" is now known | |
12632 | which indicates that a unit might not be enabled itself, but | |
c7683ffb | 12633 | another unit listed in its Also= setting might be. |
f9e00a9f LP |
12634 | |
12635 | * Similar to the various existing ConditionXYZ= settings for | |
b938cb90 | 12636 | units, there are now matching AssertXYZ= settings. While |
f9e00a9f LP |
12637 | failing conditions cause a unit to be skipped, but its job |
12638 | to succeed, failing assertions declared like this will cause | |
12639 | a unit start operation and its job to fail. | |
12640 | ||
12641 | * hostnamed now knows a new chassis type "embedded". | |
12642 | ||
12643 | * systemctl gained a new "edit" command. When used on a unit | |
b938cb90 | 12644 | file, this allows extending unit files with .d/ drop-in |
f9e00a9f LP |
12645 | configuration snippets or editing the full file (after |
12646 | copying it from /usr/lib to /etc). This will invoke the | |
12647 | user's editor (as configured with $EDITOR), and reload the | |
12648 | modified configuration after editing. | |
12649 | ||
12650 | * "systemctl status" now shows the suggested enablement state | |
12651 | for a unit, as declared in the (usually vendor-supplied) | |
12652 | system preset files. | |
12653 | ||
38b38500 | 12654 | * nss-myhostname will now resolve the single-label hostname |
f9e00a9f LP |
12655 | "gateway" to the locally configured default IP routing |
12656 | gateways, ordered by their metrics. This assigns a stable | |
12657 | name to the used gateways, regardless which ones are | |
12658 | currently configured. Note that the name will only be | |
12659 | resolved after all other name sources (if nss-myhostname is | |
12660 | configured properly) and should hence not negatively impact | |
38b38500 | 12661 | systems that use the single-label hostname "gateway" in |
f9e00a9f LP |
12662 | other contexts. |
12663 | ||
12664 | * systemd-inhibit now allows filtering by mode when listing | |
12665 | inhibitors. | |
12666 | ||
122676c9 | 12667 | * Scope and service units gained a new "Delegate" boolean |
b938cb90 | 12668 | property, which, when set, allows processes running inside the |
122676c9 LP |
12669 | unit to further partition resources. This is primarily |
12670 | useful for systemd user instances as well as container | |
12671 | managers. | |
f9e00a9f LP |
12672 | |
12673 | * journald will now pick up audit messages directly from | |
12674 | the kernel, and log them like any other log message. The | |
12675 | audit fields are split up and fully indexed. This means that | |
12676 | journalctl in many ways is now a (nicer!) alternative to | |
12677 | ausearch, the traditional audit client. Note that this | |
b938cb90 | 12678 | implements only a minimal audit client. If you want the |
f9e00a9f LP |
12679 | special audit modes like reboot-on-log-overflow, please use |
12680 | the traditional auditd instead, which can be used in | |
12681 | parallel to journald. | |
12682 | ||
12683 | * The ConditionSecurity= unit file option now understands the | |
12684 | special string "audit" to check whether auditing is | |
12685 | available. | |
12686 | ||
12687 | * journalctl gained two new commands --vacuum-size= and | |
12688 | --vacuum-time= to delete old journal files until the | |
a8eaaee7 | 12689 | remaining ones take up no more than the specified size on disk, |
f9e00a9f LP |
12690 | or are not older than the specified time. |
12691 | ||
12692 | * A new, native PPPoE library has been added to sd-network, | |
12693 | systemd's library of light-weight networking protocols. This | |
12694 | library will be used in a future version of networkd to | |
12695 | enable PPPoE communication without an external pppd daemon. | |
12696 | ||
12697 | * The busctl tool now understands a new "capture" verb that | |
12698 | works similar to "monitor", but writes a packet capture | |
12699 | trace to STDOUT that can be redirected to a file which is | |
12700 | compatible with libcap's capture file format. This can then | |
12701 | be loaded in Wireshark and similar tools to inspect bus | |
12702 | communication. | |
12703 | ||
12704 | * The busctl tool now understands a new "tree" verb that shows | |
12705 | the object trees of a specific service on the bus, or of all | |
12706 | services. | |
12707 | ||
12708 | * The busctl tool now understands a new "introspect" verb that | |
12709 | shows all interfaces and members of objects on the bus, | |
12710 | including their signature and values. This is particularly | |
12711 | useful to get more information about bus objects shown by | |
12712 | the new "busctl tree" command. | |
12713 | ||
12714 | * The busctl tool now understands new verbs "call", | |
12715 | "set-property" and "get-property" for invoking bus method | |
12716 | calls, setting and getting bus object properties in a | |
12717 | friendly way. | |
12718 | ||
12719 | * busctl gained a new --augment-creds= argument that controls | |
12720 | whether the tool shall augment credential information it | |
12721 | gets from the bus with data from /proc, in a possibly | |
12722 | race-ful way. | |
12723 | ||
12724 | * nspawn's --link-journal= switch gained two new values | |
12725 | "try-guest" and "try-host" that work like "guest" and | |
17c29493 | 12726 | "host", but do not fail if the host has no persistent |
28423d9a | 12727 | journaling enabled. -j is now equivalent to |
f9e00a9f LP |
12728 | --link-journal=try-guest. |
12729 | ||
12730 | * macvlan network devices created by nspawn will now have | |
12731 | stable MAC addresses. | |
12732 | ||
12733 | * A new SmackProcessLabel= unit setting has been added, which | |
12734 | controls the SMACK security label processes forked off by | |
12735 | the respective unit shall use. | |
12736 | ||
d4f5a1f4 DH |
12737 | * If compiled with --enable-xkbcommon, systemd-localed will |
12738 | verify x11 keymap settings by compiling the given keymap. It | |
12739 | will spew out warnings if the compilation fails. This | |
12740 | requires libxkbcommon to be installed. | |
12741 | ||
b938cb90 | 12742 | * When a coredump is collected, a larger number of metadata |
f9e00a9f | 12743 | fields is now collected and included in the journal records |
b938cb90 | 12744 | created for it. More specifically, control group membership, |
f9e00a9f LP |
12745 | environment variables, memory maps, working directory, |
12746 | chroot directory, /proc/$PID/status, and a list of open file | |
12747 | descriptors is now stored in the log entry. | |
12748 | ||
17c29493 | 12749 | * The udev hwdb now contains DPI information for mice. For |
f9e00a9f LP |
12750 | details see: |
12751 | ||
12752 | http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html | |
12753 | ||
12754 | * All systemd programs that read standalone configuration | |
12755 | files in /etc now also support a corresponding series of | |
997b2b43 JT |
12756 | .conf.d configuration directories in /etc/, /run/, |
12757 | /usr/local/lib/, /usr/lib/, and (if configured with | |
33db1b90 | 12758 | --enable-split-usr) /lib/. In particular, the following |
997b2b43 JT |
12759 | configuration files now have corresponding configuration |
12760 | directories: system.conf user.conf, logind.conf, | |
12761 | journald.conf, sleep.conf, bootchart.conf, coredump.conf, | |
12762 | resolved.conf, timesyncd.conf, journal-remote.conf, and | |
33db1b90 | 12763 | journal-upload.conf. Note that distributions should use the |
997b2b43 JT |
12764 | configuration directories in /usr/lib/; the directories in |
12765 | /etc/ are reserved for the system administrator. | |
12766 | ||
f9e00a9f LP |
12767 | * systemd-rfkill will no longer take the rfkill device name |
12768 | into account when storing rfkill state on disk, as the name | |
12769 | might be dynamically assigned and not stable. Instead, the | |
12770 | ID_PATH udev variable combined with the rfkill type (wlan, | |
1d3a473b | 12771 | bluetooth, …) is used. |
f9e00a9f LP |
12772 | |
12773 | * A new service systemd-machine-id-commit.service has been | |
12774 | added. When used on systems where /etc is read-only during | |
12775 | boot, and /etc/machine-id is not initialized (but an empty | |
12776 | file), this service will copy the temporary machine ID | |
12777 | created as replacement into /etc after the system is fully | |
12778 | booted up. This is useful for systems that are freshly | |
12779 | installed with a non-initialized machine ID, but should get | |
12780 | a fixed machine ID for subsequent boots. | |
12781 | ||
12782 | * networkd's .netdev files now provide a large set of | |
a8eaaee7 | 12783 | configuration parameters for VXLAN devices. Similarly, the |
f9e00a9f LP |
12784 | bridge port cost parameter is now configurable in .network |
12785 | files. There's also new support for configuring IP source | |
12786 | routing. networkd .link files gained support for a new | |
12787 | OriginalName= match that is useful to match against the | |
12788 | original interface name the kernel assigned. .network files | |
12789 | may include MTU= and MACAddress= fields for altering the MTU | |
12790 | and MAC address while being connected to a specific network | |
12791 | interface. | |
12792 | ||
12793 | * The LUKS logic gained supported for configuring | |
12794 | UUID-specific key files. There's also new support for naming | |
12795 | LUKS device from the kernel command line, using the new | |
12796 | luks.name= argument. | |
12797 | ||
12798 | * Timer units may now be transiently created via the bus API | |
12799 | (this was previously already available for scope and service | |
12800 | units). In addition it is now possible to create multiple | |
12801 | transient units at the same time with a single bus call. The | |
12802 | "systemd-run" tool has been updated to make use of this for | |
12803 | running commands on a specified time, in at(1)-style. | |
12804 | ||
12805 | * tmpfiles gained support for "t" lines, for assigning | |
12806 | extended attributes to files. Among other uses this may be | |
12807 | used to assign SMACK labels to files. | |
12808 | ||
13e92f39 LP |
12809 | Contributions from: Alin Rauta, Alison Chaiken, Andrej |
12810 | Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris | |
12811 | Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez, | |
12812 | Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave | |
12813 | Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin | |
12814 | Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan | |
12815 | Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe | |
12816 | Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering, | |
12817 | Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas | |
12818 | Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi, | |
12819 | Michael Biebl, Michael Chapman, Michael Marineau, Michal | |
7da81d33 LP |
12820 | Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter |
12821 | Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode, | |
12822 | Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross | |
12823 | Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani, | |
12824 | Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen, | |
12825 | Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert | |
12826 | Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek | |
13e92f39 | 12827 | |
ccddd104 | 12828 | — Berlin, 2014-12-10 |
f9e00a9f | 12829 | |
b62a309a ZJS |
12830 | CHANGES WITH 217: |
12831 | ||
78b6b7ce LP |
12832 | * journalctl gained the new options -t/--identifier= to match |
12833 | on the syslog identifier (aka "tag"), as well as --utc to | |
12834 | show log timestamps in the UTC timezone. journalctl now also | |
12835 | accepts -n/--lines=all to disable line capping in a pager. | |
b62a309a | 12836 | |
a65b8245 ZJS |
12837 | * journalctl gained a new switch, --flush, that synchronously |
12838 | flushes logs from /run/log/journal to /var/log/journal if | |
12839 | persistent storage is enabled. systemd-journal-flush.service | |
12840 | now waits until the operation is complete. | |
2a97b03b | 12841 | |
b62a309a ZJS |
12842 | * Services can notify the manager before they start a reload |
12843 | (by sending RELOADING=1) or shutdown (by sending | |
4bdc60cb LP |
12844 | STOPPING=1). This allows the manager to track and show the |
12845 | internal state of daemons and closes a race condition when | |
78b6b7ce | 12846 | the process is still running but has closed its D-Bus |
4bdc60cb | 12847 | connection. |
b62a309a | 12848 | |
78b6b7ce LP |
12849 | * Services with Type=oneshot do not have to have any ExecStart |
12850 | commands anymore. | |
b62a309a ZJS |
12851 | |
12852 | * User units are now loaded also from | |
12853 | $XDG_RUNTIME_DIR/systemd/user/. This is similar to the | |
12854 | /run/systemd/user directory that was already previously | |
12855 | supported, but is under the control of the user. | |
12856 | ||
3f9a0a52 | 12857 | * Job timeouts (i.e. timeouts on the time a job that is |
4ffd29fd LP |
12858 | queued stays in the run queue) can now optionally result in |
12859 | immediate reboot or power-off actions (JobTimeoutAction= and | |
12860 | JobTimeoutRebootArgument=). This is useful on ".target" | |
12861 | units, to limit the maximum time a target remains | |
12862 | undispatched in the run queue, and to trigger an emergency | |
12863 | operation in such a case. This is now used by default to | |
12864 | turn off the system if boot-up (as defined by everything in | |
12865 | basic.target) hangs and does not complete for at least | |
12866 | 15min. Also, if power-off or reboot hang for at least 30min | |
12867 | an immediate power-off/reboot operation is triggered. This | |
12868 | functionality is particularly useful to increase reliability | |
12869 | on embedded devices, but also on laptops which might | |
12870 | accidentally get powered on when carried in a backpack and | |
12871 | whose boot stays stuck in a hard disk encryption passphrase | |
12872 | question. | |
12873 | ||
b62a309a ZJS |
12874 | * systemd-logind can be configured to also handle lid switch |
12875 | events even when the machine is docked or multiple displays | |
12876 | are attached (HandleLidSwitchDocked= option). | |
12877 | ||
12878 | * A helper binary and a service have been added which can be | |
12879 | used to resume from hibernation in the initramfs. A | |
12880 | generator will parse the resume= option on the kernel | |
81c7dd89 | 12881 | command line to trigger resume. |
b62a309a | 12882 | |
78b6b7ce LP |
12883 | * A user console daemon systemd-consoled has been |
12884 | added. Currently, it is a preview, and will so far open a | |
12885 | single terminal on each session of the user marked as | |
09077149 | 12886 | Desktop=systemd-console. |
b62a309a ZJS |
12887 | |
12888 | * Route metrics can be specified for DHCP routes added by | |
12889 | systemd-networkd. | |
12890 | ||
ba8df74b | 12891 | * The SELinux context of socket-activated services can be set |
78b6b7ce | 12892 | from the information provided by the networking stack |
b62a309a ZJS |
12893 | (SELinuxContextFromNet= option). |
12894 | ||
12895 | * Userspace firmware loading support has been removed and | |
12896 | the minimum supported kernel version is thus bumped to 3.7. | |
12897 | ||
12898 | * Timeout for udev workers has been increased from 1 to 3 | |
12899 | minutes, but a warning will be printed after 1 minute to | |
12900 | help diagnose kernel modules that take a long time to load. | |
12901 | ||
78b6b7ce | 12902 | * Udev rules can now remove tags on devices with TAG-="foobar". |
b62a309a | 12903 | |
4bdc60cb | 12904 | * systemd's readahead implementation has been removed. In many |
f6d1de85 | 12905 | circumstances it didn't give expected benefits even for |
b62a309a | 12906 | rotational disk drives and was becoming less relevant in the |
78b6b7ce LP |
12907 | age of SSDs. As none of the developers has been using |
12908 | rotating media anymore, and nobody stepped up to actively | |
12909 | maintain this component of systemd it has now been removed. | |
b62a309a | 12910 | |
c4ac9900 | 12911 | * Swap units can use Options= to specify discard options. |
b62a309a ZJS |
12912 | Discard options specified for swaps in /etc/fstab are now |
12913 | respected. | |
12914 | ||
12915 | * Docker containers are now detected as a separate type of | |
12916 | virtualization. | |
12917 | ||
12918 | * The Password Agent protocol gained support for queries where | |
ba8df74b | 12919 | the user input is shown, useful e.g. for user names. |
78b6b7ce LP |
12920 | systemd-ask-password gained a new --echo option to turn that |
12921 | on. | |
b62a309a | 12922 | |
e6c253e3 MS |
12923 | * The default sysctl.d/ snippets will now set: |
12924 | ||
12925 | net.core.default_qdisc = fq_codel | |
12926 | ||
ba8df74b KS |
12927 | This selects Fair Queuing Controlled Delay as the default |
12928 | queuing discipline for network interfaces. fq_codel helps | |
e6c253e3 MS |
12929 | fight the network bufferbloat problem. It is believed to be |
12930 | a good default with no tuning required for most workloads. | |
12931 | Downstream distributions may override this choice. On 10Gbit | |
12932 | servers that do not do forwarding, "fq" may perform better. | |
12933 | Systems without a good clocksource should use "pfifo_fast". | |
12934 | ||
4bdc60cb LP |
12935 | * If kdbus is enabled during build a new option BusPolicy= is |
12936 | available for service units, that allows locking all service | |
12937 | processes into a stricter bus policy, in order to limit | |
12938 | access to various bus services, or even hide most of them | |
12939 | from the service's view entirely. | |
12940 | ||
12941 | * networkctl will now show the .network and .link file | |
12942 | networkd has applied to a specific interface. | |
12943 | ||
12944 | * sd-login gained a new API call sd_session_get_desktop() to | |
12945 | query which desktop environment has been selected for a | |
12946 | session. | |
12947 | ||
12948 | * UNIX utmp support is now compile-time optional to support | |
12949 | legacy-free systems. | |
12950 | ||
78b6b7ce LP |
12951 | * systemctl gained two new commands "add-wants" and |
12952 | "add-requires" for pulling in units from specific targets | |
12953 | easily. | |
12954 | ||
12955 | * If the word "rescue" is specified on the kernel command line | |
12956 | the system will now boot into rescue mode (aka | |
12957 | rescue.target), which was previously available only by | |
12958 | specifying "1" or "systemd.unit=rescue.target" on the kernel | |
12959 | command line. This new kernel command line option nicely | |
12960 | mirrors the already existing "emergency" kernel command line | |
12961 | option. | |
12962 | ||
12963 | * New kernel command line options mount.usr=, mount.usrflags=, | |
d4474c41 | 12964 | mount.usrfstype= have been added that match root=, rootflags=, |
78b6b7ce LP |
12965 | rootfstype= but allow mounting a specific file system to |
12966 | /usr. | |
12967 | ||
f6d1de85 | 12968 | * The $NOTIFY_SOCKET is now also passed to control processes of |
78b6b7ce LP |
12969 | services, not only the main process. |
12970 | ||
12971 | * This version reenables support for fsck's -l switch. This | |
12972 | means at least version v2.25 of util-linux is required for | |
12973 | operation, otherwise dead-locks on device nodes may | |
12974 | occur. Again: you need to update util-linux to at least | |
12975 | v2.25 when updating systemd to v217. | |
12976 | ||
3769415e TT |
12977 | * The "multi-seat-x" tool has been removed from systemd, as |
12978 | its functionality has been integrated into X servers 1.16, | |
12979 | and the tool is hence redundant. It is recommended to update | |
12980 | display managers invoking this tool to simply invoke X | |
12981 | directly from now on, again. | |
12982 | ||
fae9332b | 12983 | * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus |
d35f51ea ZJS |
12984 | message flag has been added for all of systemd's polkit |
12985 | authenticated method calls has been added. In particular this | |
12986 | now allows optional interactive authorization via polkit for | |
12987 | many of PID1's privileged operations such as unit file | |
12988 | enabling and disabling. | |
fae9332b | 12989 | |
cfa1571b LP |
12990 | * "udevadm hwdb --update" learnt a new switch "--usr" for |
12991 | placing the rebuilt hardware database in /usr instead of | |
12992 | /etc. When used only hardware database entries stored in | |
12993 | /usr will be used, and any user database entries in /etc are | |
12994 | ignored. This functionality is useful for vendors to ship a | |
12995 | pre-built database on systems where local configuration is | |
12996 | unnecessary or unlikely. | |
12997 | ||
7e63dd10 LP |
12998 | * Calendar time specifications in .timer units now also |
12999 | understand the strings "semi-annually", "quarterly" and | |
ba8df74b | 13000 | "minutely" as shortcuts (in addition to the preexisting |
1d3a473b | 13001 | "annually", "hourly", …). |
7e63dd10 | 13002 | |
d4474c41 TG |
13003 | * systemd-tmpfiles will now correctly create files in /dev |
13004 | at boot which are marked for creation only at boot. It is | |
13005 | recommended to always create static device nodes with 'c!' | |
13006 | and 'b!', so that they are created only at boot and not | |
13007 | overwritten at runtime. | |
13008 | ||
3b187c5c LP |
13009 | * When the watchdog logic is used for a service (WatchdogSec=) |
13010 | and the watchdog timeout is hit the service will now be | |
13011 | terminated with SIGABRT (instead of just SIGTERM), in order | |
13012 | to make sure a proper coredump and backtrace is | |
13013 | generated. This ensures that hanging services will result in | |
13014 | similar coredump/backtrace behaviour as services that hit a | |
13015 | segmentation fault. | |
13016 | ||
4b08dd87 LP |
13017 | Contributions from: Andreas Henriksson, Andrei Borzenkov, |
13018 | Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L. | |
13019 | Black, Christian Hesse, Cristian Rodríguez, Daniel Buch, | |
13020 | Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David | |
13021 | Herrmann, David Sommerseth, David Strauss, Emil Renner | |
13022 | Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger, | |
13023 | Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo | |
13024 | Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan | |
13025 | Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus | |
13026 | Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz | |
13027 | Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann, | |
13028 | Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl, | |
13029 | Michael Marineau, Michael Olbrich, Michael Scherer, Michal | |
13030 | Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt, | |
13031 | Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard | |
13032 | Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof, | |
13033 | Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd | |
13034 | Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant | |
13035 | Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen, | |
13036 | Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein | |
13037 | Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew | |
13e92f39 | 13038 | Jędrzejewski-Szmek |
4b08dd87 | 13039 | |
ccddd104 | 13040 | — Berlin, 2014-10-28 |
4b08dd87 | 13041 | |
b72ddf0f | 13042 | CHANGES WITH 216: |
b2ca0d63 LP |
13043 | |
13044 | * timedated no longer reads NTP implementation unit names from | |
b72ddf0f | 13045 | /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP |
b2ca0d63 LP |
13046 | implementations should add a |
13047 | ||
b72ddf0f | 13048 | Conflicts=systemd-timesyncd.service |
b2ca0d63 LP |
13049 | |
13050 | to their unit files to take over and replace systemd's NTP | |
13051 | default functionality. | |
13052 | ||
13053 | * systemd-sysusers gained a new line type "r" for configuring | |
13054 | which UID/GID ranges to allocate system users/groups | |
13055 | from. Lines of type "u" may now add an additional column | |
13056 | that specifies the home directory for the system user to be | |
13057 | created. Also, systemd-sysusers may now optionally read user | |
13058 | information from STDIN instead of a file. This is useful for | |
13059 | invoking it from RPM preinst scriptlets that need to create | |
13060 | users before the first RPM file is installed since these | |
13061 | files might need to be owned by them. A new | |
13062 | %sysusers_create_inline RPM macro has been introduced to do | |
13063 | just that. systemd-sysusers now updates the shadow files as | |
13064 | well as the user/group databases, which should enhance | |
13065 | compatibility with certain tools like grpck. | |
13066 | ||
d35f51ea ZJS |
13067 | * A number of bus APIs of PID 1 now optionally consult polkit to |
13068 | permit access for otherwise unprivileged clients under certain | |
13069 | conditions. Note that this currently doesn't support | |
13070 | interactive authentication yet, but this is expected to be | |
13071 | added eventually, too. | |
b2ca0d63 LP |
13072 | |
13073 | * /etc/machine-info now has new fields for configuring the | |
13074 | deployment environment of the machine, as well as the | |
13075 | location of the machine. hostnamectl has been updated with | |
13076 | new command to update these fields. | |
13077 | ||
13078 | * systemd-timesyncd has been updated to automatically acquire | |
13079 | NTP server information from systemd-networkd, which might | |
13080 | have been discovered via DHCP. | |
13081 | ||
13082 | * systemd-resolved now includes a caching DNS stub resolver | |
13083 | and a complete LLMNR name resolution implementation. A new | |
daa05349 AB |
13084 | NSS module "nss-resolve" has been added which can be used |
13085 | instead of glibc's own "nss-dns" to resolve hostnames via | |
b2ca0d63 LP |
13086 | systemd-resolved. Hostnames, addresses and arbitrary RRs may |
13087 | be resolved via systemd-resolved D-Bus APIs. In contrast to | |
13088 | the glibc internal resolver systemd-resolved is aware of | |
13089 | multi-homed system, and keeps DNS server and caches separate | |
5f02e26c | 13090 | and per-interface. Queries are sent simultaneously on all |
b2ca0d63 LP |
13091 | interfaces that have DNS servers configured, in order to |
13092 | properly handle VPNs and local LANs which might resolve | |
13093 | separate sets of domain names. systemd-resolved may acquire | |
a1a4a25e | 13094 | DNS server information from systemd-networkd automatically, |
b2ca0d63 LP |
13095 | which in turn might have discovered them via DHCP. A tool |
13096 | "systemd-resolve-host" has been added that may be used to | |
13097 | query the DNS logic in resolved. systemd-resolved implements | |
13098 | IDNA and automatically uses IDNA or UTF-8 encoding depending | |
13099 | on whether classic DNS or LLMNR is used as transport. In the | |
13100 | next releases we intend to add a DNSSEC and mDNS/DNS-SD | |
13101 | implementation to systemd-resolved. | |
13102 | ||
13103 | * A new NSS module nss-mymachines has been added, that | |
13104 | automatically resolves the names of all local registered | |
13105 | containers to their respective IP addresses. | |
13106 | ||
13107 | * A new client tool "networkctl" for systemd-networkd has been | |
13108 | added. It currently is entirely passive and will query | |
13109 | networking configuration from udev, rtnetlink and networkd, | |
5f02e26c | 13110 | and present it to the user in a very friendly |
b2ca0d63 LP |
13111 | way. Eventually, we hope to extend it to become a full |
13112 | control utility for networkd. | |
13113 | ||
13114 | * .socket units gained a new DeferAcceptSec= setting that | |
13115 | controls the kernels' TCP_DEFER_ACCEPT sockopt for | |
a8eaaee7 | 13116 | TCP. Similarly, support for controlling TCP keep-alive |
b2ca0d63 LP |
13117 | settings has been added (KeepAliveTimeSec=, |
13118 | KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for | |
13119 | turning off Nagle's algorithm on TCP has been added | |
13120 | (NoDelay=). | |
13121 | ||
a1a4a25e | 13122 | * logind learned a new session type "web", for use in projects |
b2ca0d63 LP |
13123 | like Cockpit which register web clients as PAM sessions. |
13124 | ||
13125 | * timer units with at least one OnCalendar= setting will now | |
46ae28d8 | 13126 | be started only after time-sync.target has been |
b2ca0d63 LP |
13127 | reached. This way they will not elapse before the system |
13128 | clock has been corrected by a local NTP client or | |
13129 | similar. This is particular useful on RTC-less embedded | |
13130 | machines, that come up with an invalid system clock. | |
13131 | ||
13132 | * systemd-nspawn's --network-veth= switch should now result in | |
13133 | stable MAC addresses for both the outer and the inner side | |
13134 | of the link. | |
13135 | ||
13136 | * systemd-nspawn gained a new --volatile= switch for running | |
13137 | container instances with /etc or /var unpopulated. | |
13138 | ||
13139 | * The kdbus client code has been updated to use the new Linux | |
13140 | 3.17 memfd subsystem instead of the old kdbus-specific one. | |
13141 | ||
13142 | * systemd-networkd's DHCP client and server now support | |
01da80b1 LP |
13143 | FORCERENEW. There are also new configuration options to |
13144 | configure the vendor client identifier and broadcast mode | |
13145 | for DHCP. | |
b2ca0d63 LP |
13146 | |
13147 | * systemd will no longer inform the kernel about the current | |
13148 | timezone, as this is necessarily incorrect and racy as the | |
13149 | kernel has no understanding of DST and similar | |
13150 | concepts. This hence means FAT timestamps will be always | |
13151 | considered UTC, similar to what Android is already | |
13152 | doing. Also, when the RTC is configured to the local time | |
13153 | (rather than UTC) systemd will never synchronize back to it, | |
13154 | as this might confuse Windows at a later boot. | |
13155 | ||
13156 | * systemd-analyze gained a new command "verify" for offline | |
13157 | validation of unit files. | |
13158 | ||
13159 | * systemd-networkd gained support for a couple of additional | |
13160 | settings for bonding networking setups. Also, the metric for | |
13161 | statically configured routes may now be configured. For | |
13162 | network interfaces where this is appropriate the peer IP | |
13163 | address may now be configured. | |
13164 | ||
26568403 TG |
13165 | * systemd-networkd's DHCP client will no longer request |
13166 | broadcasting by default, as this tripped up some networks. | |
13167 | For hardware where broadcast is required the feature should | |
13168 | be switched back on using RequestBroadcast=yes. | |
13169 | ||
13170 | * systemd-networkd will now set up IPv4LL addresses (when | |
13171 | enabled) even if DHCP is configured successfully. | |
13172 | ||
13173 | * udev will now default to respect network device names given | |
13174 | by the kernel when the kernel indicates that these are | |
13175 | predictable. This behavior can be tweaked by changing | |
13176 | NamePolicy= in the relevant .link file. | |
13177 | ||
b2ca0d63 LP |
13178 | * A new library systemd-terminal has been added that |
13179 | implements full TTY stream parsing and rendering. This | |
13180 | library is supposed to be used later on for implementing a | |
13181 | full userspace VT subsystem, replacing the current kernel | |
13182 | implementation. | |
13183 | ||
13184 | * A new tool systemd-journal-upload has been added to push | |
13185 | journal data to a remote system running | |
13186 | systemd-journal-remote. | |
13187 | ||
13188 | * journald will no longer forward all local data to another | |
13189 | running syslog daemon. This change has been made because | |
13190 | rsyslog (which appears to be the most commonly used syslog | |
13191 | implementation these days) no longer makes use of this, and | |
13192 | instead pulls the data out of the journal on its own. Since | |
5f02e26c | 13193 | forwarding the messages to a non-existent syslog server is |
b2ca0d63 LP |
13194 | more expensive than we assumed we have now turned this |
13195 | off. If you run a syslog server that is not a recent rsyslog | |
13196 | version, you have to turn this option on again | |
13197 | (ForwardToSyslog= in journald.conf). | |
13198 | ||
13199 | * journald now optionally supports the LZ4 compressor for | |
13200 | larger journal fields. This compressor should perform much | |
13201 | better than XZ which was the previous default. | |
13202 | ||
13203 | * machinectl now shows the IP addresses of local containers, | |
13204 | if it knows them, plus the interface name of the container. | |
13205 | ||
13206 | * A new tool "systemd-escape" has been added that makes it | |
13207 | easy to escape strings to build unit names and similar. | |
13208 | ||
13209 | * sd_notify() messages may now include a new ERRNO= field | |
13210 | which is parsed and collected by systemd and shown among the | |
13211 | "systemctl status" output for a service. | |
13212 | ||
13213 | * A new component "systemd-firstboot" has been added that | |
13214 | queries the most basic systemd information (timezone, | |
a1a4a25e | 13215 | hostname, root password) interactively on first |
b2ca0d63 LP |
13216 | boot. Alternatively it may also be used to provision these |
13217 | things offline on OS images installed into directories. | |
13218 | ||
01da80b1 LP |
13219 | * The default sysctl.d/ snippets will now set |
13220 | ||
13221 | net.ipv4.conf.default.promote_secondaries=1 | |
13222 | ||
13223 | This has the benefit of no flushing secondary IP addresses | |
13224 | when primary addresses are removed. | |
13225 | ||
b2ca0d63 LP |
13226 | Contributions from: Ansgar Burchardt, Bastien Nocera, Colin |
13227 | Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel | |
13228 | Mack, Dan Williams, Dave Reisner, David Herrmann, Denis | |
13229 | Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald | |
13230 | Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann | |
13231 | B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin | |
13232 | Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, | |
13233 | Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael | |
13234 | Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar, | |
13235 | Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert | |
13236 | Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef | |
13237 | Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas | |
13238 | Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets, | |
13239 | Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut | |
13240 | Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek | |
13241 | ||
ccddd104 | 13242 | — Berlin, 2014-08-19 |
b72ddf0f | 13243 | |
3dff3e00 | 13244 | CHANGES WITH 215: |
24a2bf4c LP |
13245 | |
13246 | * A new tool systemd-sysusers has been added. This tool | |
13247 | creates system users and groups in /etc/passwd and | |
13248 | /etc/group, based on static declarative system user/group | |
13249 | definitions in /usr/lib/sysusers.d/. This is useful to | |
13250 | enable factory resets and volatile systems that boot up with | |
13251 | an empty /etc directory, and thus need system users and | |
13252 | groups created during early boot. systemd now also ships | |
13253 | with two default sysusers.d/ files for the most basic | |
13254 | users and groups systemd and the core operating system | |
13255 | require. | |
13256 | ||
13257 | * A new tmpfiles snippet has been added that rebuilds the | |
13258 | essential files in /etc on boot, should they be missing. | |
13259 | ||
13260 | * A directive for ensuring automatic clean-up of | |
13261 | /var/cache/man/ has been removed from the default | |
13262 | configuration. This line should now be shipped by the man | |
13263 | implementation. The necessary change has been made to the | |
13264 | man-db implementation. Note that you need to update your man | |
13265 | implementation to one that ships this line, otherwise no | |
13266 | automatic clean-up of /var/cache/man will take place. | |
13267 | ||
13268 | * A new condition ConditionNeedsUpdate= has been added that | |
13269 | may conditionalize services to only run when /etc or /var | |
13270 | are "older" than the vendor operating system resources in | |
13271 | /usr. This is useful for reconstructing or updating /etc | |
13272 | after an offline update of /usr or a factory reset, on the | |
13273 | next reboot. Services that want to run once after such an | |
13274 | update or reset should use this condition and order | |
13275 | themselves before the new systemd-update-done.service, which | |
13276 | will mark the two directories as fully updated. A number of | |
13277 | service files have been added making use of this, to rebuild | |
13278 | the udev hardware database, the journald message catalog and | |
13279 | dynamic loader cache (ldconfig). The systemd-sysusers tool | |
13280 | described above also makes use of this now. With this in | |
13281 | place it is now possible to start up a minimal operating | |
ce1dde29 | 13282 | system with /etc empty cleanly. For more information on the |
24a2bf4c LP |
13283 | concepts involved see this recent blog story: |
13284 | ||
dc7e580e | 13285 | https://0pointer.de/blog/projects/stateless.html |
24a2bf4c LP |
13286 | |
13287 | * A new system group "input" has been introduced, and all | |
13288 | input device nodes get this group assigned. This is useful | |
13289 | for system-level software to get access to input devices. It | |
3dff3e00 KS |
13290 | complements what is already done for "audio" and "video". |
13291 | ||
24a2bf4c LP |
13292 | * systemd-networkd learnt minimal DHCPv4 server support in |
13293 | addition to the existing DHCPv4 client support. It also | |
13294 | learnt DHCPv6 client and IPv6 Router Solicitation client | |
13295 | support. The DHCPv4 client gained support for static routes | |
13296 | passed in from the server. Note that the [DHCPv4] section | |
13297 | known in older systemd-networkd versions has been renamed to | |
13298 | [DHCP] and is now also used by the DHCPv6 client. Existing | |
c7435cc9 LP |
13299 | .network files using settings of this section should be |
13300 | updated, though compatibility is maintained. Optionally, the | |
13301 | client hostname may now be sent to the DHCP server. | |
24a2bf4c | 13302 | |
c7435cc9 LP |
13303 | * networkd gained support for vxlan virtual networks as well |
13304 | as tun/tap and dummy devices. | |
24a2bf4c LP |
13305 | |
13306 | * networkd gained support for automatic allocation of address | |
13307 | ranges for interfaces from a system-wide pool of | |
13308 | addresses. This is useful for dynamically managing a large | |
13309 | number of interfaces with a single network configuration | |
13310 | file. In particular this is useful to easily assign | |
13311 | appropriate IP addresses to the veth links of a large number | |
13312 | of nspawn instances. | |
13313 | ||
13314 | * RPM macros for processing sysusers, sysctl and binfmt | |
13315 | drop-in snippets at package installation time have been | |
13316 | added. | |
13317 | ||
13318 | * The /etc/os-release file should now be placed in | |
13319 | /usr/lib/os-release. The old location is automatically | |
13320 | created as symlink. /usr/lib is the more appropriate | |
13321 | location of this file, since it shall actually describe the | |
13322 | vendor operating system shipped in /usr, and not the | |
13323 | configuration stored in /etc. | |
13324 | ||
13325 | * .mount units gained a new boolean SloppyOptions= setting | |
13326 | that maps to mount(8)'s -s option which enables permissive | |
13327 | parsing of unknown mount options. | |
13328 | ||
13329 | * tmpfiles learnt a new "L+" directive which creates a symlink | |
13330 | but (unlike "L") deletes a pre-existing file first, should | |
13331 | it already exist and not already be the correct | |
a8eaaee7 | 13332 | symlink. Similarly, "b+", "c+" and "p+" directives have been |
24a2bf4c LP |
13333 | added as well, which create block and character devices, as |
13334 | well as fifos in the filesystem, possibly removing any | |
13335 | pre-existing files of different types. | |
13336 | ||
13337 | * For tmpfiles' "L", "L+", "C" and "C+" directives the final | |
13338 | 'argument' field (which so far specified the source to | |
ce1dde29 | 13339 | symlink/copy the files from) is now optional. If omitted the |
24a2bf4c LP |
13340 | same file os copied from /usr/share/factory/ suffixed by the |
13341 | full destination path. This is useful for populating /etc | |
13342 | with essential files, by copying them from vendor defaults | |
13343 | shipped in /usr/share/factory/etc. | |
13344 | ||
13345 | * A new command "systemctl preset-all" has been added that | |
13346 | applies the service preset settings to all installed unit | |
13347 | files. A new switch --preset-mode= has been added that | |
13348 | controls whether only enable or only disable operations | |
13349 | shall be executed. | |
13350 | ||
13351 | * A new command "systemctl is-system-running" has been added | |
13352 | that allows checking the overall state of the system, for | |
ce1dde29 | 13353 | example whether it is fully up and running. |
24a2bf4c LP |
13354 | |
13355 | * When the system boots up with an empty /etc, the equivalent | |
13356 | to "systemctl preset-all" is executed during early boot, to | |
13357 | make sure all default services are enabled after a factory | |
13358 | reset. | |
13359 | ||
13360 | * systemd now contains a minimal preset file that enables the | |
13361 | most basic services systemd ships by default. | |
13362 | ||
13363 | * Unit files' [Install] section gained a new DefaultInstance= | |
13364 | field for defining the default instance to create if a | |
13365 | template unit is enabled with no instance specified. | |
13366 | ||
13367 | * A new passive target cryptsetup-pre.target has been added | |
13368 | that may be used by services that need to make they run and | |
13369 | finish before the first LUKS cryptographic device is set up. | |
13370 | ||
13371 | * The /dev/loop-control and /dev/btrfs-control device nodes | |
13372 | are now owned by the "disk" group by default, opening up | |
13373 | access to this group. | |
13374 | ||
13375 | * systemd-coredump will now automatically generate a | |
13376 | stack trace of all core dumps taking place on the system, | |
13377 | based on elfutils' libdw library. This stack trace is logged | |
13378 | to the journal. | |
13379 | ||
13380 | * systemd-coredump may now optionally store coredumps directly | |
13381 | on disk (in /var/lib/systemd/coredump, possibly compressed), | |
13382 | instead of storing them unconditionally in the journal. This | |
13383 | mode is the new default. A new configuration file | |
13384 | /etc/systemd/coredump.conf has been added to configure this | |
13385 | and other parameters of systemd-coredump. | |
13386 | ||
13387 | * coredumpctl gained a new "info" verb to show details about a | |
13388 | specific coredump. A new switch "-1" has also been added | |
13389 | that makes sure to only show information about the most | |
13390 | recent entry instead of all entries. Also, as the tool is | |
13391 | generally useful now the "systemd-" prefix of the binary | |
13392 | name has been removed. Distributions that want to maintain | |
13393 | compatibility with the old name should add a symlink from | |
13394 | the old name to the new name. | |
13395 | ||
13396 | * journald's SplitMode= now defaults to "uid". This makes sure | |
ce1dde29 | 13397 | that unprivileged users can access their own coredumps with |
24a2bf4c LP |
13398 | coredumpctl without restrictions. |
13399 | ||
13400 | * New kernel command line options "systemd.wants=" (for | |
13401 | pulling an additional unit during boot), "systemd.mask=" | |
13402 | (for masking a specific unit for the boot), and | |
13403 | "systemd.debug-shell" (for enabling the debug shell on tty9) | |
13404 | have been added. This is implemented in the new generator | |
13405 | "systemd-debug-generator". | |
13406 | ||
13407 | * systemd-nspawn will now by default filter a couple of | |
13408 | syscalls for containers, among them those required for | |
13409 | kernel module loading, direct x86 IO port access, swap | |
13410 | management, and kexec. Most importantly though | |
13411 | open_by_handle_at() is now prohibited for containers, | |
13412 | closing a hole similar to a recently discussed vulnerability | |
13413 | in docker regarding access to files on file hierarchies the | |
b938cb90 JE |
13414 | container should normally not have access to. Note that, for |
13415 | nspawn, we generally make no security claims anyway (and | |
24a2bf4c LP |
13416 | this is explicitly documented in the man page), so this is |
13417 | just a fix for one of the most obvious problems. | |
13418 | ||
13419 | * A new man page file-hierarchy(7) has been added that | |
13420 | contains a minimized, modernized version of the file system | |
13421 | layout systemd expects, similar in style to the FHS | |
c7435cc9 LP |
13422 | specification or hier(5). A new tool systemd-path(1) has |
13423 | been added to query many of these paths for the local | |
13424 | machine and user. | |
24a2bf4c LP |
13425 | |
13426 | * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no | |
13427 | longer done. Since the directory now has a per-user size | |
13428 | limit, and is cleaned on logout this appears unnecessary, | |
13429 | in particular since this now brings the lifecycle of this | |
13430 | directory closer in line with how IPC objects are handled. | |
13431 | ||
13432 | * systemd.pc now exports a number of additional directories, | |
13433 | including $libdir (which is useful to identify the library | |
13434 | path for the primary architecture of the system), and a | |
13435 | couple of drop-in directories. | |
13436 | ||
3058e017 TLSC |
13437 | * udev's predictable network interface names now use the dev_port |
13438 | sysfs attribute, introduced in linux 3.15 instead of dev_id to | |
13439 | distinguish between ports of the same PCI function. dev_id should | |
13440 | only be used for ports using the same HW address, hence the need | |
13441 | for dev_port. | |
13442 | ||
c7435cc9 LP |
13443 | * machined has been updated to export the OS version of a |
13444 | container (read from /etc/os-release and | |
13445 | /usr/lib/os-release) on the bus. This is now shown in | |
13446 | "machinectl status" for a machine. | |
13447 | ||
13448 | * A new service setting RestartForceExitStatus= has been | |
13449 | added. If configured to a set of exit signals or process | |
13450 | return values, the service will be restarted when the main | |
13451 | daemon process exits with any of them, regardless of the | |
13452 | Restart= setting. | |
13453 | ||
13454 | * systemctl's -H switch for connecting to remote systemd | |
13455 | machines has been extended so that it may be used to | |
13456 | directly connect to a specific container on the | |
13457 | host. "systemctl -H root@foobar:waldi" will now connect as | |
13458 | user "root" to host "foobar", and then proceed directly to | |
13459 | the container named "waldi". Note that currently you have to | |
13460 | authenticate as user "root" for this to work, as entering | |
13461 | containers is a privileged operation. | |
13462 | ||
13463 | Contributions from: Andreas Henriksson, Benjamin Steinwender, | |
13464 | Carl Schaefer, Christian Hesse, Colin Ian King, Cristian | |
13465 | Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene | |
13466 | Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo | |
13467 | Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart | |
13468 | Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine | |
13469 | Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich, | |
13470 | Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le | |
13471 | Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan, | |
13472 | Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe | |
13473 | Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar | |
13474 | Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek | |
13475 | ||
ccddd104 | 13476 | — Berlin, 2014-07-03 |
c7435cc9 | 13477 | |
4196a3ea KS |
13478 | CHANGES WITH 214: |
13479 | ||
13480 | * As an experimental feature, udev now tries to lock the | |
13481 | disk device node (flock(LOCK_SH|LOCK_NB)) while it | |
13482 | executes events for the disk or any of its partitions. | |
13483 | Applications like partitioning programs can lock the | |
13484 | disk device node (flock(LOCK_EX)) and claim temporary | |
13485 | device ownership that way; udev will entirely skip all event | |
13486 | handling for this disk and its partitions. If the disk | |
13487 | was opened for writing, the close will trigger a partition | |
13488 | table rescan in udev's "watch" facility, and if needed | |
71449caf | 13489 | synthesize "change" events for the disk and all its partitions. |
8d0e0ddd | 13490 | This is now unconditionally enabled, and if it turns out to |
4196a3ea | 13491 | cause major problems, we might turn it on only for specific |
45df8656 | 13492 | devices, or might need to disable it entirely. Device Mapper |
4196a3ea KS |
13493 | devices are excluded from this logic. |
13494 | ||
04e91da2 LP |
13495 | * We temporarily dropped the "-l" switch for fsck invocations, |
13496 | since they collide with the flock() logic above. util-linux | |
13497 | upstream has been changed already to avoid this conflict, | |
5238e957 | 13498 | and we will re-add "-l" as soon as util-linux with this |
04e91da2 LP |
13499 | change has been released. |
13500 | ||
13501 | * The dependency on libattr has been removed. Since a long | |
8d0e0ddd | 13502 | time, the extended attribute calls have moved to glibc, and |
04e91da2 LP |
13503 | libattr is thus unnecessary. |
13504 | ||
ce830873 | 13505 | * Virtualization detection works without privileges now. This |
04e91da2 LP |
13506 | means the systemd-detect-virt binary no longer requires |
13507 | CAP_SYS_PTRACE file capabilities, and our daemons can run | |
71449caf | 13508 | with fewer privileges. |
04e91da2 LP |
13509 | |
13510 | * systemd-networkd now runs under its own "systemd-network" | |
13511 | user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE, | |
13512 | CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but | |
13513 | loses the ability to write to files owned by root this way. | |
13514 | ||
a8eaaee7 | 13515 | * Similarly, systemd-resolved now runs under its own |
04e91da2 LP |
13516 | "systemd-resolve" user with no capabilities remaining. |
13517 | ||
a8eaaee7 | 13518 | * Similarly, systemd-bus-proxyd now runs under its own |
04e91da2 LP |
13519 | "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining. |
13520 | ||
13521 | * systemd-networkd gained support for setting up "veth" | |
a8eaaee7 | 13522 | virtual Ethernet devices for container connectivity, as well |
04e91da2 LP |
13523 | as GRE and VTI tunnels. |
13524 | ||
13525 | * systemd-networkd will no longer automatically attempt to | |
13526 | manually load kernel modules necessary for certain tunnel | |
8d0e0ddd | 13527 | transports. Instead, it is assumed the kernel loads them |
04e91da2 LP |
13528 | automatically when required. This only works correctly on |
13529 | very new kernels. On older kernels, please consider adding | |
c54bed5d | 13530 | the kernel modules to /etc/modules-load.d/ as a work-around. |
04e91da2 | 13531 | |
cd14eda3 | 13532 | * The resolv.conf file systemd-resolved generates has been |
8d0e0ddd JE |
13533 | moved to /run/systemd/resolve/. If you have a symlink from |
13534 | /etc/resolv.conf, it might be necessary to correct it. | |
cd14eda3 | 13535 | |
ef392da6 | 13536 | * Two new service settings, ProtectHome= and ProtectSystem=, |
8d0e0ddd | 13537 | have been added. When enabled, they will make the user data |
04e91da2 LP |
13538 | (such as /home) inaccessible or read-only and the system |
13539 | (such as /usr) read-only, for specific services. This allows | |
13540 | very light-weight per-service sandboxing to avoid | |
13541 | modifications of user data or system files from | |
13542 | services. These two new switches have been enabled for all | |
13543 | of systemd's long-running services, where appropriate. | |
13544 | ||
13545 | * Socket units gained new SocketUser= and SocketGroup= | |
13546 | settings to set the owner user and group of AF_UNIX sockets | |
13547 | and FIFOs in the file system. | |
13548 | ||
8d0e0ddd | 13549 | * Socket units gained a new RemoveOnStop= setting. If enabled, |
04e91da2 LP |
13550 | all FIFOS and sockets in the file system will be removed |
13551 | when the specific socket unit is stopped. | |
13552 | ||
13553 | * Socket units gained a new Symlinks= setting. It takes a list | |
13554 | of symlinks to create to file system sockets or FIFOs | |
45df8656 | 13555 | created by the specific Unix sockets. This is useful to |
de04bbdc | 13556 | manage symlinks to socket nodes with the same lifecycle as |
04e91da2 LP |
13557 | the socket itself. |
13558 | ||
13559 | * The /dev/log socket and /dev/initctl FIFO have been moved to | |
13560 | /run, and have been replaced by symlinks. This allows | |
13561 | connecting to these facilities even if PrivateDevices=yes is | |
13562 | used for a service (which makes /dev/log itself unavailable, | |
13563 | but /run is left). This also has the benefit of ensuring | |
13564 | that /dev only contains device nodes, directories and | |
13565 | symlinks, and nothing else. | |
13566 | ||
13567 | * sd-daemon gained two new calls sd_pid_notify() and | |
13568 | sd_pid_notifyf(). They are similar to sd_notify() and | |
13569 | sd_notifyf(), but allow overriding of the source PID of | |
13570 | notification messages if permissions permit this. This is | |
13571 | useful to send notify messages on behalf of a different | |
13572 | process (for example, the parent process). The | |
13573 | systemd-notify tool has been updated to make use of this | |
13574 | when sending messages (so that notification messages now | |
13575 | originate from the shell script invoking systemd-notify and | |
13576 | not the systemd-notify process itself. This should minimize | |
13577 | a race where systemd fails to associate notification | |
13578 | messages to services when the originating process already | |
13579 | vanished. | |
13580 | ||
13581 | * A new "on-abnormal" setting for Restart= has been added. If | |
8d0e0ddd | 13582 | set, it will result in automatic restarts on all "abnormal" |
04e91da2 LP |
13583 | reasons for a process to exit, which includes unclean |
13584 | signals, core dumps, timeouts and watchdog timeouts, but | |
13585 | does not include clean and unclean exit codes or clean | |
13586 | signals. Restart=on-abnormal is an alternative for | |
13587 | Restart=on-failure for services that shall be able to | |
13588 | terminate and avoid restarts on certain errors, by | |
13589 | indicating so with an unclean exit code. Restart=on-failure | |
13590 | or Restart=on-abnormal is now the recommended setting for | |
13591 | all long-running services. | |
13592 | ||
13593 | * If the InaccessibleDirectories= service setting points to a | |
13594 | mount point (or if there are any submounts contained within | |
13595 | it), it is now attempted to completely unmount it, to make | |
13596 | the file systems truly unavailable for the respective | |
13597 | service. | |
13598 | ||
13599 | * The ReadOnlyDirectories= service setting and | |
13600 | systemd-nspawn's --read-only parameter are now recursively | |
13601 | applied to all submounts, too. | |
13602 | ||
13603 | * Mount units may now be created transiently via the bus APIs. | |
13604 | ||
13605 | * The support for SysV and LSB init scripts has been removed | |
13606 | from the systemd daemon itself. Instead, it is now | |
13607 | implemented as a generator that creates native systemd units | |
13608 | from these scripts when needed. This enables us to remove a | |
13609 | substantial amount of legacy code from PID 1, following the | |
13610 | fact that many distributions only ship a very small number | |
13611 | of LSB/SysV init scripts nowadays. | |
13612 | ||
cc98b302 | 13613 | * Privileged Xen (dom0) domains are not considered |
04e91da2 LP |
13614 | virtualization anymore by the virtualization detection |
13615 | logic. After all, they generally have unrestricted access to | |
71449caf | 13616 | the hardware and usually are used to manage the unprivileged |
04e91da2 LP |
13617 | (domU) domains. |
13618 | ||
13619 | * systemd-tmpfiles gained a new "C" line type, for copying | |
13620 | files or entire directories. | |
13621 | ||
13622 | * systemd-tmpfiles "m" lines are now fully equivalent to "z" | |
8d0e0ddd JE |
13623 | lines. So far, they have been non-globbing versions of the |
13624 | latter, and have thus been redundant. In future, it is | |
13625 | recommended to only use "z". "m" has hence been removed | |
04e91da2 LP |
13626 | from the documentation, even though it stays supported. |
13627 | ||
13628 | * A tmpfiles snippet to recreate the most basic structure in | |
13629 | /var has been added. This is enough to create the /var/run → | |
13630 | /run symlink and create a couple of structural | |
13631 | directories. This allows systems to boot up with an empty or | |
8d0e0ddd JE |
13632 | volatile /var. Of course, while with this change, the core OS |
13633 | now is capable with dealing with a volatile /var, not all | |
04e91da2 | 13634 | user services are ready for it. However, we hope that sooner |
8d0e0ddd | 13635 | or later, many service daemons will be changed upstream so |
04e91da2 LP |
13636 | that they are able to automatically create their necessary |
13637 | directories in /var at boot, should they be missing. This is | |
13638 | the first step to allow state-less systems that only require | |
13639 | the vendor image for /usr to boot. | |
13640 | ||
13641 | * systemd-nspawn has gained a new --tmpfs= switch to mount an | |
13642 | empty tmpfs instance to a specific directory. This is | |
13643 | particularly useful for making use of the automatic | |
13644 | reconstruction of /var (see above), by passing --tmpfs=/var. | |
13645 | ||
13646 | * Access modes specified in tmpfiles snippets may now be | |
13647 | prefixed with "~", which indicates that they shall be masked | |
daa05349 | 13648 | by whether the existing file or directory is currently |
8d0e0ddd | 13649 | writable, readable or executable at all. Also, if specified, |
04e91da2 LP |
13650 | the sgid/suid/sticky bits will be masked for all |
13651 | non-directories. | |
13652 | ||
13653 | * A new passive target unit "network-pre.target" has been | |
13654 | added which is useful for services that shall run before any | |
13655 | network is configured, for example firewall scripts. | |
13656 | ||
4c0d13bd LP |
13657 | * The "floppy" group that previously owned the /dev/fd* |
13658 | devices is no longer used. The "disk" group is now used | |
13659 | instead. Distributions should probably deprecate usage of | |
13660 | this group. | |
13661 | ||
dc1d6c02 LP |
13662 | Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian |
13663 | King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David | |
13664 | Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers, | |
13665 | Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny | |
13666 | Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel | |
13667 | Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew | |
13668 | Jędrzejewski-Szmek | |
13669 | ||
ccddd104 | 13670 | — Berlin, 2014-06-11 |
dc1d6c02 | 13671 | |
6936cd89 LP |
13672 | CHANGES WITH 213: |
13673 | ||
13674 | * A new "systemd-timesyncd" daemon has been added for | |
69beda1f | 13675 | synchronizing the system clock across the network. It |
6936cd89 | 13676 | implements an SNTP client. In contrast to NTP |
8d0e0ddd | 13677 | implementations such as chrony or the NTP reference server, |
6936cd89 | 13678 | this only implements a client side, and does not bother with |
c9679c65 LP |
13679 | the full NTP complexity, focusing only on querying time from |
13680 | one remote server and synchronizing the local clock to | |
6936cd89 | 13681 | it. Unless you intend to serve NTP to networked clients or |
8d0e0ddd | 13682 | want to connect to local hardware clocks, this simple NTP |
6936cd89 LP |
13683 | client should be more than appropriate for most |
13684 | installations. The daemon runs with minimal privileges, and | |
13685 | has been hooked up with networkd to only operate when | |
13686 | network connectivity is available. The daemon saves the | |
13687 | current clock to disk every time a new NTP sync has been | |
13688 | acquired, and uses this to possibly correct the system clock | |
69beda1f | 13689 | early at bootup, in order to accommodate for systems that |
6936cd89 | 13690 | lack an RTC such as the Raspberry Pi and embedded devices, |
8d0e0ddd | 13691 | and to make sure that time monotonically progresses on these |
c9679c65 | 13692 | systems, even if it is not always correct. To make use of |
8d0e0ddd | 13693 | this daemon, a new system user and group "systemd-timesync" |
c9679c65 | 13694 | needs to be created on installation of systemd. |
6936cd89 | 13695 | |
69beda1f KS |
13696 | * The queue "seqnum" interface of libudev has been disabled, as |
13697 | it was generally incompatible with device namespacing as | |
6936cd89 LP |
13698 | sequence numbers of devices go "missing" if the devices are |
13699 | part of a different namespace. | |
13700 | ||
13701 | * "systemctl list-timers" and "systemctl list-sockets" gained | |
13702 | a --recursive switch for showing units of these types also | |
499b604b ZJS |
13703 | for all local containers, similar in style to the already |
13704 | supported --recursive switch for "systemctl list-units". | |
6936cd89 LP |
13705 | |
13706 | * A new RebootArgument= setting has been added for service | |
13707 | units, which may be used to specify a kernel reboot argument | |
499b604b | 13708 | to use when triggering reboots with StartLimitAction=. |
6936cd89 LP |
13709 | |
13710 | * A new FailureAction= setting has been added for service | |
13711 | units which may be used to specify an operation to trigger | |
499b604b | 13712 | when a service fails. This works similarly to |
8d0e0ddd | 13713 | StartLimitAction=, but unlike it, controls what is done |
6936cd89 LP |
13714 | immediately rather than only after several attempts to |
13715 | restart the service in question. | |
13716 | ||
13717 | * hostnamed got updated to also expose the kernel name, | |
499b604b ZJS |
13718 | release, and version on the bus. This is useful for |
13719 | executing commands like hostnamectl with the -H switch. | |
13720 | systemd-analyze makes use of this to properly display | |
13721 | details when running non-locally. | |
6936cd89 LP |
13722 | |
13723 | * The bootchart tool can now show cgroup information in the | |
13724 | graphs it generates. | |
13725 | ||
13726 | * The CFS CPU quota cgroup attribute is now exposed for | |
13727 | services. The new CPUQuota= switch has been added for this | |
13728 | which takes a percentage value. Setting this will have the | |
13729 | result that a service may never get more CPU time than the | |
13730 | specified percentage, even if the machine is otherwise idle. | |
13731 | ||
13732 | * systemd-networkd learned IPIP and SIT tunnel support. | |
13733 | ||
13734 | * LSB init scripts exposing a dependency on $network will now | |
13735 | get a dependency on network-online.target rather than simply | |
13736 | network.target. This should bring LSB handling closer to | |
13737 | what it was on SysV systems. | |
13738 | ||
13739 | * A new fsck.repair= kernel option has been added to control | |
13740 | how fsck shall deal with unclean file systems at boot. | |
13741 | ||
2ad98889 ZJS |
13742 | * The (.ini) configuration file parser will now silently ignore |
13743 | sections whose names begin with "X-". This may be used to maintain | |
13744 | application-specific extension sections in unit files. | |
6936cd89 LP |
13745 | |
13746 | * machined gained a new API to query the IP addresses of | |
13747 | registered containers. "machinectl status" has been updated | |
13748 | to show these addresses in its output. | |
13749 | ||
13750 | * A new call sd_uid_get_display() has been added to the | |
13751 | sd-login APIs for querying the "primary" session of a | |
13752 | user. The "primary" session of the user is elected from the | |
13753 | user's sessions and generally a graphical session is | |
13754 | preferred over a text one. | |
13755 | ||
13756 | * A minimal systemd-resolved daemon has been added. It | |
13757 | currently simply acts as a companion to systemd-networkd and | |
13758 | manages resolv.conf based on per-interface DNS | |
13759 | configuration, possibly supplied via DHCP. In the long run | |
13760 | we hope to extend this into a local DNSSEC enabled DNS and | |
13761 | mDNS cache. | |
13762 | ||
68dd0956 TG |
13763 | * The systemd-networkd-wait-online tool is now enabled by |
13764 | default. It will delay network-online.target until a network | |
13765 | connection has been configured. The tool primarily integrates | |
13766 | with networkd, but will also make a best effort to make sense | |
13767 | of network configuration performed in some other way. | |
13768 | ||
6936cd89 | 13769 | * Two new service options StartupCPUShares= and |
499b604b | 13770 | StartupBlockIOWeight= have been added that work similarly to |
6936cd89 | 13771 | CPUShares= and BlockIOWeight= however only apply during |
69beda1f | 13772 | system startup. This is useful to prioritize certain services |
6936cd89 LP |
13773 | differently during bootup than during normal runtime. |
13774 | ||
8e7acf67 LP |
13775 | * hostnamed has been changed to prefer the statically |
13776 | configured hostname in /etc/hostname (unless set to | |
13777 | 'localhost' or empty) over any dynamic one supplied by | |
8d0e0ddd | 13778 | dhcp. With this change, the rules for picking the hostname |
8e7acf67 LP |
13779 | match more closely the rules of other configuration settings |
13780 | where the local administrator's configuration in /etc always | |
13781 | overrides any other settings. | |
13782 | ||
5238e957 | 13783 | Contributions from: Ali H. Caliskan, Alison Chaiken, Bas van |
6936cd89 LP |
13784 | den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch, |
13785 | Dan Kilman, Dave Reisner, David Härdeman, David Herrmann, | |
13786 | David Strauss, Dimitris Spingos, Djalal Harouni, Eelco | |
13787 | Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg | |
13788 | Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan | |
13789 | Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark, | |
13790 | Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas | |
13791 | Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas, | |
8e7acf67 LP |
13792 | Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael |
13793 | Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis | |
13794 | Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode, | |
13795 | Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter, | |
13796 | Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler, | |
13797 | Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar | |
13798 | Lindskog, WaLyong Cho, Will Woods, Zbigniew | |
6936cd89 LP |
13799 | Jędrzejewski-Szmek |
13800 | ||
ccddd104 | 13801 | — Beijing, 2014-05-28 |
6936cd89 | 13802 | |
51c61cda LP |
13803 | CHANGES WITH 212: |
13804 | ||
13805 | * When restoring the screen brightness at boot, stay away from | |
13806 | the darkest setting or from the lowest 5% of the available | |
13807 | range, depending on which is the larger value of both. This | |
13808 | should effectively protect the user from rebooting into a | |
13809 | black screen, should the brightness have been set to minimum | |
13810 | by accident. | |
13811 | ||
13812 | * sd-login gained a new sd_machine_get_class() call to | |
13813 | determine the class ("vm" or "container") of a machine | |
13814 | registered with machined. | |
13815 | ||
13816 | * sd-login gained new calls | |
13817 | sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(), | |
13818 | to query the identity of the peer of a local AF_UNIX | |
499b604b | 13819 | connection. They operate similarly to their sd_pid_get_xyz() |
51c61cda LP |
13820 | counterparts. |
13821 | ||
13822 | * PID 1 will now maintain a system-wide system state engine | |
13823 | with the states "starting", "running", "degraded", | |
13824 | "maintenance", "stopping". These states are bound to system | |
13825 | startup, normal runtime, runtime with at least one failed | |
13826 | service, rescue/emergency mode and system shutdown. This | |
13827 | state is shown in the "systemctl status" output when no unit | |
13828 | name is passed. It is useful to determine system state, in | |
13829 | particularly when doing so for many systems or containers at | |
13830 | once. | |
13831 | ||
13832 | * A new command "list-machines" has been added to "systemctl" | |
13833 | that lists all local OS containers and shows their system | |
13834 | state (see above), if systemd runs inside of them. | |
13835 | ||
13836 | * systemctl gained a new "-r" switch to recursively enumerate | |
13837 | units on all local containers, when used with the | |
13838 | "list-unit" command (which is the default one that is | |
13839 | executed when no parameters are specified). | |
13840 | ||
13841 | * The GPT automatic partition discovery logic will now honour | |
13842 | two GPT partition flags: one may be set on a partition to | |
13843 | cause it to be mounted read-only, and the other may be set | |
13844 | on a partition to ignore it during automatic discovery. | |
13845 | ||
13846 | * Two new GPT type UUIDs have been added for automatic root | |
70a44afe | 13847 | partition discovery, for 32-bit and 64-bit ARM. This is not |
51c61cda LP |
13848 | particularly useful for discovering the root directory on |
13849 | these architectures during bare-metal boots (since UEFI is | |
13850 | not common there), but still very useful to allow booting of | |
13851 | ARM disk images in nspawn with the -i option. | |
13852 | ||
13853 | * MAC addresses of interfaces created with nspawn's | |
13854 | --network-interface= switch will now be generated from the | |
13855 | machine name, and thus be stable between multiple invocations | |
13856 | of the container. | |
13857 | ||
13858 | * logind will now automatically remove all IPC objects owned | |
13859 | by a user if she or he fully logs out. This makes sure that | |
13860 | users who are logged out cannot continue to consume IPC | |
13861 | resources. This covers SysV memory, semaphores and message | |
13862 | queues as well as POSIX shared memory and message | |
de04bbdc | 13863 | queues. Traditionally, SysV and POSIX IPC had no lifecycle |
b8bde116 JE |
13864 | limits. With this functionality, that is corrected. This may |
13865 | be turned off by using the RemoveIPC= switch of logind.conf. | |
51c61cda LP |
13866 | |
13867 | * The systemd-machine-id-setup and tmpfiles tools gained a | |
13868 | --root= switch to operate on a specific root directory, | |
13869 | instead of /. | |
13870 | ||
13871 | * journald can now forward logged messages to the TTYs of all | |
13872 | logged in users ("wall"). This is the default for all | |
13873 | emergency messages now. | |
13874 | ||
13875 | * A new tool systemd-journal-remote has been added to stream | |
13876 | journal log messages across the network. | |
13877 | ||
13878 | * /sys/fs/cgroup/ is now mounted read-only after all cgroup | |
13879 | controller trees are mounted into it. Note that the | |
13880 | directories mounted beneath it are not read-only. This is a | |
13881 | security measure and is particularly useful because glibc | |
13882 | actually includes a search logic to pick any tmpfs it can | |
13883 | find to implement shm_open() if /dev/shm is not available | |
13884 | (which it might very well be in namespaced setups). | |
13885 | ||
13886 | * machinectl gained a new "poweroff" command to cleanly power | |
13887 | down a local OS container. | |
13888 | ||
13889 | * The PrivateDevices= unit file setting will now also drop the | |
13890 | CAP_MKNOD capability from the capability bound set, and | |
13891 | imply DevicePolicy=closed. | |
13892 | ||
13893 | * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used | |
13894 | comprehensively on all long-running systemd services where | |
13895 | this is appropriate. | |
13896 | ||
13897 | * systemd-udevd will now run in a disassociated mount | |
b8bde116 | 13898 | namespace. To mount directories from udev rules, make sure to |
51c61cda LP |
13899 | pull in mount units via SYSTEMD_WANTS properties. |
13900 | ||
13901 | * The kdbus support gained support for uploading policy into | |
13902 | the kernel. sd-bus gained support for creating "monitoring" | |
13903 | connections that can eavesdrop into all bus communication | |
13904 | for debugging purposes. | |
13905 | ||
13906 | * Timestamps may now be specified in seconds since the UNIX | |
13907 | epoch Jan 1st, 1970 by specifying "@" followed by the value | |
13908 | in seconds. | |
13909 | ||
13910 | * Native tcpwrap support in systemd has been removed. tcpwrap | |
13911 | is old code, not really maintained anymore and has serious | |
13912 | shortcomings, and better options such as firewalls | |
13913 | exist. For setups that require tcpwrap usage, please | |
13914 | consider invoking your socket-activated service via tcpd, | |
13915 | like on traditional inetd. | |
13916 | ||
13917 | * A new system.conf configuration option | |
13918 | DefaultTimerAccuracySec= has been added that controls the | |
13919 | default AccuracySec= setting of .timer units. | |
13920 | ||
b8bde116 | 13921 | * Timer units gained a new WakeSystem= switch. If enabled, |
51c61cda LP |
13922 | timers configured this way will cause the system to resume |
13923 | from system suspend (if the system supports that, which most | |
13924 | do these days). | |
13925 | ||
b8bde116 | 13926 | * Timer units gained a new Persistent= switch. If enabled, |
51c61cda LP |
13927 | timers configured this way will save to disk when they have |
13928 | been last triggered. This information is then used on next | |
13929 | reboot to possible execute overdue timer events, that | |
d28315e4 JE |
13930 | could not take place because the system was powered off. |
13931 | This enables simple anacron-like behaviour for timer units. | |
51c61cda LP |
13932 | |
13933 | * systemctl's "list-timers" will now also list the time a | |
13934 | timer unit was last triggered in addition to the next time | |
13935 | it will be triggered. | |
13936 | ||
13937 | * systemd-networkd will now assign predictable IPv4LL | |
13938 | addresses to its local interfaces. | |
13939 | ||
13940 | Contributions from: Brandon Philips, Daniel Buch, Daniel Mack, | |
13941 | Dave Reisner, David Herrmann, Gerd Hoffmann, Greg | |
13942 | Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh | |
13943 | Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine | |
13944 | Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna, | |
13945 | Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler, | |
13946 | Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen, | |
13947 | Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew | |
13948 | Jędrzejewski-Szmek | |
13949 | ||
ccddd104 | 13950 | — Berlin, 2014-03-25 |
51c61cda | 13951 | |
699b6b34 LP |
13952 | CHANGES WITH 211: |
13953 | ||
13954 | * A new unit file setting RestrictAddressFamilies= has been | |
13955 | added to restrict which socket address families unit | |
13956 | processes gain access to. This takes address family names | |
13957 | like "AF_INET" or "AF_UNIX", and is useful to minimize the | |
13958 | attack surface of services via exotic protocol stacks. This | |
13959 | is built on seccomp system call filters. | |
13960 | ||
13961 | * Two new unit file settings RuntimeDirectory= and | |
13962 | RuntimeDirectoryMode= have been added that may be used to | |
13963 | manage a per-daemon runtime directories below /run. This is | |
13964 | an alternative for setting up directory permissions with | |
13965 | tmpfiles snippets, and has the advantage that the runtime | |
13966 | directory's lifetime is bound to the daemon runtime and that | |
13967 | the daemon starts up with an empty directory each time. This | |
13968 | is particularly useful when writing services that drop | |
f1721625 | 13969 | privileges using the User= or Group= setting. |
699b6b34 LP |
13970 | |
13971 | * The DeviceAllow= unit setting now supports globbing for | |
13972 | matching against device group names. | |
13973 | ||
13974 | * The systemd configuration file system.conf gained new | |
13975 | settings DefaultCPUAccounting=, DefaultBlockIOAccounting=, | |
13976 | DefaultMemoryAccounting= to globally turn on/off accounting | |
13977 | for specific resources (cgroups) for all units. These | |
22e7062d | 13978 | settings may still be overridden individually in each unit |
699b6b34 LP |
13979 | though. |
13980 | ||
13981 | * systemd-gpt-auto-generator is now able to discover /srv and | |
13982 | root partitions in addition to /home and swap partitions. It | |
13983 | also supports LUKS-encrypted partitions now. With this in | |
b8bde116 | 13984 | place, automatic discovery of partitions to mount following |
699b6b34 | 13985 | the Discoverable Partitions Specification |
a794a4d8 | 13986 | (https://systemd.io/DISCOVERABLE_PARTITIONS/) |
699b6b34 LP |
13987 | is now a lot more complete. This allows booting without |
13988 | /etc/fstab and without root= on the kernel command line on | |
b8bde116 | 13989 | systems prepared appropriately. |
699b6b34 LP |
13990 | |
13991 | * systemd-nspawn gained a new --image= switch which allows | |
13992 | booting up disk images and Linux installations on any block | |
13993 | device that follow the Discoverable Partitions Specification | |
13994 | (see above). This means that installations made with | |
13995 | appropriately updated installers may now be started and | |
13996 | deployed using container managers, completely | |
13997 | unmodified. (We hope that libvirt-lxc will add support for | |
13998 | this feature soon, too.) | |
13999 | ||
14000 | * systemd-nspawn gained a new --network-macvlan= setting to | |
14001 | set up a private macvlan interface for the | |
499b604b | 14002 | container. Similarly, systemd-networkd gained a new |
699b6b34 LP |
14003 | Kind=macvlan setting in .netdev files. |
14004 | ||
14005 | * systemd-networkd now supports configuring local addresses | |
14006 | using IPv4LL. | |
14007 | ||
14008 | * A new tool systemd-network-wait-online has been added to | |
14009 | synchronously wait for network connectivity using | |
14010 | systemd-networkd. | |
14011 | ||
14012 | * The sd-bus.h bus API gained a new sd_bus_track object for | |
de04bbdc | 14013 | tracking the lifecycle of bus peers. Note that sd-bus.h is |
699b6b34 LP |
14014 | still not a public API though (unless you specify |
14015 | --enable-kdbus on the configure command line, which however | |
14016 | voids your warranty and you get no API stability guarantee). | |
14017 | ||
14018 | * The $XDG_RUNTIME_DIR runtime directories for each user are | |
14019 | now individual tmpfs instances, which has the benefit of | |
14020 | introducing separate pools for each user, with individual | |
4ef6e535 | 14021 | size limits, and thus making sure that unprivileged clients |
699b6b34 LP |
14022 | can no longer negatively impact the system or other users by |
14023 | filling up their $XDG_RUNTIME_DIR. A new logind.conf setting | |
14024 | RuntimeDirectorySize= has been introduced that allows | |
14025 | controlling the default size limit for all users. It | |
14026 | defaults to 10% of the available physical memory. This is no | |
14027 | replacement for quotas on tmpfs though (which the kernel | |
14028 | still does not support), as /dev/shm and /tmp are still | |
4ef6e535 | 14029 | shared resources used by both the system and unprivileged |
699b6b34 LP |
14030 | users. |
14031 | ||
14032 | * logind will now automatically turn off automatic suspending | |
14033 | on laptop lid close when more than one display is | |
14034 | connected. This was previously expected to be implemented | |
14035 | individually in desktop environments (such as GNOME), | |
14036 | however has been added to logind now, in order to fix a | |
14037 | boot-time race where a desktop environment might not have | |
14038 | been started yet and thus not been able to take an inhibitor | |
14039 | lock at the time where logind already suspends the system | |
14040 | due to a closed lid. | |
14041 | ||
14042 | * logind will now wait at least 30s after each system | |
14043 | suspend/resume cycle, and 3min after system boot before | |
14044 | suspending the system due to a closed laptop lid. This | |
14045 | should give USB docking stations and similar enough time to | |
4ef6e535 | 14046 | be probed and configured after system resume and boot in |
699b6b34 LP |
14047 | order to then act as suspend blocker. |
14048 | ||
14049 | * systemd-run gained a new --property= setting which allows | |
14050 | initialization of resource control properties (and others) | |
14051 | for the created scope or service unit. Example: "systemd-run | |
14052 | --property=BlockIOWeight=10 updatedb" may be used to run | |
14053 | updatedb at a low block IO scheduling weight. | |
14054 | ||
14055 | * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches | |
14056 | now also work in --scope mode. | |
14057 | ||
14058 | * When systemd is compiled with kdbus support, basic support | |
14059 | for enforced policies is now in place. (Note that enabling | |
14060 | kdbus still voids your warranty and no API compatibility | |
14061 | promises are made.) | |
14062 | ||
14063 | Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin | |
14064 | K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni, | |
14065 | Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay | |
14066 | Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn, | |
14067 | Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt, | |
14068 | Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef | |
14069 | Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas | |
14070 | Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom | |
14071 | Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook, | |
14072 | Zbigniew Jędrzejewski-Szmek | |
14073 | ||
ccddd104 | 14074 | — Berlin, 2014-03-12 |
699b6b34 | 14075 | |
43c71255 LP |
14076 | CHANGES WITH 210: |
14077 | ||
14078 | * systemd will now relabel /dev after loading the SMACK policy | |
14079 | according to SMACK rules. | |
14080 | ||
67dd87c5 | 14081 | * A new unit file option AppArmorProfile= has been added to |
43c71255 LP |
14082 | set the AppArmor profile for the processes of a unit. |
14083 | ||
14084 | * A new condition check ConditionArchitecture= has been added | |
14085 | to conditionalize units based on the system architecture, as | |
14086 | reported by uname()'s "machine" field. | |
14087 | ||
14088 | * systemd-networkd now supports matching on the system | |
38b38500 | 14089 | virtualization, architecture, kernel command line, hostname |
43c71255 LP |
14090 | and machine ID. |
14091 | ||
ed28905e | 14092 | * logind is now a lot more aggressive when suspending the |
43c71255 | 14093 | machine due to a closed laptop lid. Instead of acting only |
b8bde116 | 14094 | on the lid close action, it will continuously watch the lid |
43c71255 LP |
14095 | status and act on it. This is useful for laptops where the |
14096 | power button is on the outside of the chassis so that it can | |
ed28905e | 14097 | be reached without opening the lid (such as the Lenovo |
b8bde116 | 14098 | Yoga). On those machines, logind will now immediately |
ed28905e | 14099 | re-suspend the machine if the power button has been |
43c71255 LP |
14100 | accidentally pressed while the laptop was suspended and in a |
14101 | backpack or similar. | |
14102 | ||
14103 | * logind will now watch SW_DOCK switches and inhibit reaction | |
14104 | to the lid switch if it is pressed. This means that logind | |
d27893ef | 14105 | will not suspend the machine anymore if the lid is closed |
949138cc | 14106 | and the system is docked, if the laptop supports SW_DOCK |
43c71255 LP |
14107 | notifications via the input layer. Note that ACPI docking |
14108 | stations do not generate this currently. Also note that this | |
14109 | logic is usually not fully sufficient and Desktop | |
14110 | Environments should take a lid switch inhibitor lock when an | |
14111 | external display is connected, as systemd will not watch | |
14112 | this on its own. | |
14113 | ||
14114 | * nspawn will now make use of the devices cgroup controller by | |
14115 | default, and only permit creation of and access to the usual | |
14116 | API device nodes like /dev/null or /dev/random, as well as | |
14117 | access to (but not creation of) the pty devices. | |
14118 | ||
14119 | * We will now ship a default .network file for | |
14120 | systemd-networkd that automatically configures DHCP for | |
14121 | network interfaces created by nspawn's --network-veth or | |
14122 | --network-bridge= switches. | |
14123 | ||
14124 | * systemd will now understand the usual M, K, G, T suffixes | |
14125 | according to SI conventions (i.e. to the base 1000) when | |
14126 | referring to throughput and hardware metrics. It will stay | |
14127 | with IEC conventions (i.e. to the base 1024) for software | |
14128 | metrics, according to what is customary according to | |
14129 | Wikipedia. We explicitly document which base applies for | |
14130 | each configuration option. | |
14131 | ||
6b000af4 LP |
14132 | * The DeviceAllow= setting in unit files now supports a syntax to |
14133 | allow-list an entire group of devices node majors at once, based on | |
14134 | the /proc/devices listing. For example, with the string "char-pts", | |
14135 | it is now possible to allow-list all current and future pseudo-TTYs | |
14136 | at once. | |
43c71255 LP |
14137 | |
14138 | * sd-event learned a new "post" event source. Event sources of | |
14139 | this type are triggered by the dispatching of any event | |
14140 | source of a type that is not "post". This is useful for | |
14141 | implementing clean-up and check event sources that are | |
14142 | triggered by other work being done in the program. | |
14143 | ||
14144 | * systemd-networkd is no longer statically enabled, but uses | |
14145 | the usual [Install] sections so that it can be | |
14146 | enabled/disabled using systemctl. It still is enabled by | |
14147 | default however. | |
14148 | ||
b8bde116 | 14149 | * When creating a veth interface pair with systemd-nspawn, the |
43c71255 LP |
14150 | host side will now be prefixed with "vb-" if |
14151 | --network-bridge= is used, and with "ve-" if --network-veth | |
b8bde116 | 14152 | is used. This way, it is easy to distinguish these cases on |
43c71255 LP |
14153 | the host, for example to apply different configuration to |
14154 | them with systemd-networkd. | |
14155 | ||
d27893ef LP |
14156 | * The compatibility libraries for libsystemd-journal.so, |
14157 | libsystem-id128.so, libsystemd-login.so and | |
14158 | libsystemd-daemon.so do not make use of IFUNC | |
b8bde116 | 14159 | anymore. Instead, we now build libsystemd.so multiple times |
d27893ef LP |
14160 | under these alternative names. This means that the footprint |
14161 | is drastically increased, but given that these are | |
b8bde116 | 14162 | transitional compatibility libraries, this should not matter |
d27893ef LP |
14163 | much. This change has been made necessary to support the ARM |
14164 | platform for these compatibility libraries, as the ARM | |
d28315e4 | 14165 | toolchain is not really at the same level as the toolchain |
ed28905e | 14166 | for other architectures like x86 and does not support |
d27893ef LP |
14167 | IFUNC. Please make sure to use --enable-compat-libs only |
14168 | during a transitional period! | |
14169 | ||
f26ad321 ZJS |
14170 | * The .include syntax has been deprecated and is not documented |
14171 | anymore. Drop-in files in .d directories should be used instead. | |
14172 | ||
13b28d82 | 14173 | Contributions from: Andreas Fuchs, Armin K., Colin Walters, |
43c71255 LP |
14174 | Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni, |
14175 | Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper | |
14176 | St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach, | |
14177 | Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike | |
14178 | Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe | |
14179 | Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog, | |
14180 | Zbigniew Jędrzejewski-Szmek | |
14181 | ||
ccddd104 | 14182 | — Berlin, 2014-02-24 |
43c71255 | 14183 | |
e49b5aad LP |
14184 | CHANGES WITH 209: |
14185 | ||
14186 | * A new component "systemd-networkd" has been added that can | |
14187 | be used to configure local network interfaces statically or | |
8b7d0494 JSJ |
14188 | via DHCP. It is capable of bringing up bridges, VLANs, and |
14189 | bonding. Currently, no hook-ups for interactive network | |
4670e9d5 | 14190 | configuration are provided. Use this for your initrd, |
8b7d0494 JSJ |
14191 | container, embedded, or server setup if you need a simple, |
14192 | yet powerful, network configuration solution. This | |
4670e9d5 | 14193 | configuration subsystem is quite nifty, as it allows wildcard |
1e190502 | 14194 | hotplug matching in interfaces. For example, with a single |
4670e9d5 | 14195 | configuration snippet, you can configure that all Ethernet |
1e190502 ZJS |
14196 | interfaces showing up are automatically added to a bridge, |
14197 | or similar. It supports link-sensing and more. | |
e49b5aad LP |
14198 | |
14199 | * A new tool "systemd-socket-proxyd" has been added which can | |
4c2413bf | 14200 | act as a bidirectional proxy for TCP sockets. This is |
e49b5aad LP |
14201 | useful for adding socket activation support to services that |
14202 | do not actually support socket activation, including virtual | |
4c2413bf | 14203 | machines and the like. |
e49b5aad LP |
14204 | |
14205 | * Add a new tool to save/restore rfkill state on | |
14206 | shutdown/boot. | |
14207 | ||
8b7d0494 JSJ |
14208 | * Save/restore state of keyboard backlights in addition to |
14209 | display backlights on shutdown/boot. | |
e49b5aad LP |
14210 | |
14211 | * udev learned a new SECLABEL{} construct to label device | |
14212 | nodes with a specific security label when they appear. For | |
4c2413bf | 14213 | now, only SECLABEL{selinux} is supported, but the syntax is |
e49b5aad LP |
14214 | prepared for additional security frameworks. |
14215 | ||
14216 | * udev gained a new scheme to configure link-level attributes | |
14217 | from files in /etc/systemd/network/*.link. These files can | |
8b7d0494 | 14218 | match against MAC address, device path, driver name and type, |
4c2413bf | 14219 | and will apply attributes like the naming policy, link speed, |
8b7d0494 | 14220 | MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC |
1d3a473b | 14221 | address assignment policy (randomized, …). |
e49b5aad | 14222 | |
dfb08b05 ZJS |
14223 | * The configuration of network interface naming rules for |
14224 | "permanent interface names" has changed: a new NamePolicy= | |
14225 | setting in the [Link] section of .link files determines the | |
a8eaaee7 | 14226 | priority of possible naming schemes (onboard, slot, MAC, |
dfb08b05 ZJS |
14227 | path). The default value of this setting is determined by |
14228 | /usr/lib/net/links/99-default.link. Old | |
14229 | 80-net-name-slot.rules udev configuration file has been | |
14230 | removed, so local configuration overriding this file should | |
ce830873 | 14231 | be adapted to override 99-default.link instead. |
dfb08b05 | 14232 | |
e49b5aad | 14233 | * When the User= switch is used in a unit file, also |
4c2413bf | 14234 | initialize $SHELL= based on the user database entry. |
e49b5aad LP |
14235 | |
14236 | * systemd no longer depends on libdbus. All communication is | |
14237 | now done with sd-bus, systemd's low-level bus library | |
14238 | implementation. | |
14239 | ||
14240 | * kdbus support has been added to PID 1 itself. When kdbus is | |
4c2413bf | 14241 | enabled, this causes PID 1 to set up the system bus and |
e49b5aad LP |
14242 | enable support for a new ".busname" unit type that |
14243 | encapsulates bus name activation on kdbus. It works a little | |
14244 | bit like ".socket" units, except for bus names. A new | |
14245 | generator has been added that converts classic dbus1 service | |
14246 | activation files automatically into native systemd .busname | |
14247 | and .service units. | |
14248 | ||
14249 | * sd-bus: add a light-weight vtable implementation that allows | |
14250 | defining objects on the bus with a simple static const | |
14251 | vtable array of its methods, signals and properties. | |
14252 | ||
8b7d0494 | 14253 | * systemd will not generate or install static dbus |
e49b5aad | 14254 | introspection data anymore to /usr/share/dbus-1/interfaces, |
1e190502 | 14255 | as the precise format of these files is unclear, and |
e49b5aad LP |
14256 | nothing makes use of it. |
14257 | ||
14258 | * A proxy daemon is now provided to proxy clients connecting | |
14259 | via classic D-Bus AF_UNIX sockets to kdbus, to provide full | |
14260 | compatibility with classic D-Bus. | |
14261 | ||
14262 | * A bus driver implementation has been added that supports the | |
14263 | classic D-Bus bus driver calls on kdbus, also for | |
14264 | compatibility purposes. | |
14265 | ||
14266 | * A new API "sd-event.h" has been added that implements a | |
14267 | minimal event loop API built around epoll. It provides a | |
14268 | couple of features that direct epoll usage is lacking: | |
b9761003 | 14269 | prioritization of events, scales to large numbers of timer |
e49b5aad LP |
14270 | events, per-event timer slack (accuracy), system-wide |
14271 | coalescing of timer events, exit handlers, watchdog | |
14272 | supervision support using systemd's sd_notify() API, child | |
14273 | process handling. | |
14274 | ||
14275 | * A new API "sd-rntl.h" has been added that provides an API | |
14276 | around the route netlink interface of the kernel, similar in | |
14277 | style to "sd-bus.h". | |
14278 | ||
7e95eda5 PF |
14279 | * A new API "sd-dhcp-client.h" has been added that provides a |
14280 | small DHCPv4 client-side implementation. This is used by | |
e49b5aad LP |
14281 | "systemd-networkd". |
14282 | ||
4c2413bf | 14283 | * There is a new kernel command line option |
8b7d0494 JSJ |
14284 | "systemd.restore_state=0|1". When set to "0", none of the |
14285 | systemd tools will restore saved runtime state to hardware | |
14286 | devices. More specifically, the rfkill and backlight states | |
14287 | are not restored. | |
e49b5aad LP |
14288 | |
14289 | * The FsckPassNo= compatibility option in mount/service units | |
14290 | has been removed. The fstab generator will now add the | |
14291 | necessary dependencies automatically, and does not require | |
14292 | PID1's support for that anymore. | |
14293 | ||
8b7d0494 | 14294 | * journalctl gained a new switch, --list-boots, that lists |
e49b5aad LP |
14295 | recent boots with their times and boot IDs. |
14296 | ||
14297 | * The various tools like systemctl, loginctl, timedatectl, | |
1d3a473b | 14298 | busctl, systemd-run, … have gained a new switch "-M" to |
e49b5aad LP |
14299 | connect to a specific, local OS container (as direct |
14300 | connection, without requiring SSH). This works on any | |
14301 | container that is registered with machined, such as those | |
14302 | created by libvirt-lxc or nspawn. | |
14303 | ||
14304 | * systemd-run and systemd-analyze also gained support for "-H" | |
4c2413bf | 14305 | to connect to remote hosts via SSH. This is particularly |
8b7d0494 JSJ |
14306 | useful for systemd-run because it enables queuing of jobs |
14307 | onto remote systems. | |
e49b5aad LP |
14308 | |
14309 | * machinectl gained a new command "login" to open a getty | |
14310 | login in any local container. This works with any container | |
14311 | that is registered with machined (such as those created by | |
8e420494 | 14312 | libvirt-lxc or nspawn), and which runs systemd inside. |
e49b5aad LP |
14313 | |
14314 | * machinectl gained a new "reboot" command that may be used to | |
14315 | trigger a reboot on a specific container that is registered | |
14316 | with machined. This works on any container that runs an init | |
14317 | system of some kind. | |
14318 | ||
14319 | * systemctl gained a new "list-timers" command to print a nice | |
14320 | listing of installed timer units with the times they elapse | |
14321 | next. | |
14322 | ||
14323 | * Alternative reboot() parameters may now be specified on the | |
14324 | "systemctl reboot" command line and are passed to the | |
14325 | reboot() system call. | |
14326 | ||
14327 | * systemctl gained a new --job-mode= switch to configure the | |
14328 | mode to queue a job with. This is a more generic version of | |
8b7d0494 | 14329 | --fail, --irreversible, and --ignore-dependencies, which are |
e49b5aad LP |
14330 | still available but not advertised anymore. |
14331 | ||
e49b5aad LP |
14332 | * /etc/systemd/system.conf gained new settings to configure |
14333 | various default timeouts of units, as well as the default | |
b9761003 | 14334 | start limit interval and burst. These may still be overridden |
e49b5aad LP |
14335 | within each Unit. |
14336 | ||
270f1624 LP |
14337 | * PID1 will now export on the bus profile data of the security |
14338 | policy upload process (such as the SELinux policy upload to | |
8e420494 | 14339 | the kernel). |
e49b5aad | 14340 | |
4670e9d5 | 14341 | * journald: when forwarding logs to the console, include |
1e190502 ZJS |
14342 | timestamps (following the setting in |
14343 | /sys/module/printk/parameters/time). | |
e49b5aad LP |
14344 | |
14345 | * OnCalendar= in timer units now understands the special | |
14346 | strings "yearly" and "annually". (Both are equivalent) | |
14347 | ||
14348 | * The accuracy of timer units is now configurable with the new | |
14349 | AccuracySec= setting. It defaults to 1min. | |
14350 | ||
14351 | * A new dependency type JoinsNamespaceOf= has been added that | |
14352 | allows running two services within the same /tmp and network | |
14353 | namespace, if PrivateNetwork= or PrivateTmp= are used. | |
14354 | ||
14355 | * A new command "cat" has been added to systemctl. It outputs | |
14356 | the original unit file of a unit, and concatenates the | |
1e190502 ZJS |
14357 | contents of additional "drop-in" unit file snippets, so that |
14358 | the full configuration is shown. | |
e49b5aad LP |
14359 | |
14360 | * systemctl now supports globbing on the various "list-xyz" | |
14361 | commands, like "list-units" or "list-sockets", as well as on | |
1e190502 ZJS |
14362 | those commands which take multiple unit names. |
14363 | ||
14364 | * journalctl's --unit= switch gained support for globbing. | |
e49b5aad LP |
14365 | |
14366 | * All systemd daemons now make use of the watchdog logic so | |
14367 | that systemd automatically notices when they hang. | |
14368 | ||
4c2413bf | 14369 | * If the $container_ttys environment variable is set, |
e49b5aad LP |
14370 | getty-generator will automatically spawn a getty for each |
14371 | listed tty. This is useful for container managers to request | |
14372 | login gettys to be spawned on as many ttys as needed. | |
14373 | ||
14374 | * %h, %s, %U specifier support is not available anymore when | |
14375 | used in unit files for PID 1. This is because NSS calls are | |
14376 | not safe from PID 1. They stay available for --user | |
14377 | instances of systemd, and as special case for the root user. | |
14378 | ||
e49b5aad LP |
14379 | * loginctl gained a new "--no-legend" switch to turn off output |
14380 | of the legend text. | |
14381 | ||
14382 | * The "sd-login.h" API gained three new calls: | |
14383 | sd_session_is_remote(), sd_session_get_remote_user(), | |
14384 | sd_session_get_remote_host() to query information about | |
14385 | remote sessions. | |
14386 | ||
8e420494 LP |
14387 | * The udev hardware database now also carries vendor/product |
14388 | information of SDIO devices. | |
e49b5aad LP |
14389 | |
14390 | * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to | |
14391 | determine whether watchdog notifications are requested by | |
14392 | the system manager. | |
14393 | ||
1e190502 | 14394 | * Socket-activated per-connection services now include a |
e49b5aad LP |
14395 | short description of the connection parameters in the |
14396 | description. | |
14397 | ||
4c2413bf | 14398 | * tmpfiles gained a new "--boot" option. When this is not used, |
e49b5aad | 14399 | only lines where the command character is not suffixed with |
4670e9d5 | 14400 | "!" are executed. When this option is specified, those |
1e190502 ZJS |
14401 | options are executed too. This partitions tmpfiles |
14402 | directives into those that can be safely executed at any | |
14403 | time, and those which should be run only at boot (for | |
14404 | example, a line that creates /run/nologin). | |
e49b5aad | 14405 | |
c0c5af00 | 14406 | * A new API "sd-resolve.h" has been added which provides a simple |
38b38500 | 14407 | asynchronous wrapper around glibc NSS hostname resolution |
e49b5aad | 14408 | calls, such as getaddrinfo(). In contrast to glibc's |
4c2413bf JE |
14409 | getaddrinfo_a(), it does not use signals. In contrast to most |
14410 | other asynchronous name resolution libraries, this one does | |
14411 | not reimplement DNS, but reuses NSS, so that alternate | |
38b38500 | 14412 | hostname resolution systems continue to work, such as mDNS, |
8b7d0494 | 14413 | LDAP, etc. This API is based on libasyncns, but it has been |
e49b5aad LP |
14414 | cleaned up for inclusion in systemd. |
14415 | ||
6300b3ec LP |
14416 | * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h", |
14417 | "sd-daemon.h" are no longer found in individual libraries | |
14418 | libsystemd-journal.so, libsystemd-login.so, | |
14419 | libsystemd-id128.so, libsystemd-daemon.so. Instead, we have | |
8b7d0494 JSJ |
14420 | merged them into a single library, libsystemd.so, which |
14421 | provides all symbols. The reason for this is cyclic | |
e49b5aad | 14422 | dependencies, as these libraries tend to use each other's |
d28315e4 | 14423 | symbols. So far, we have managed to workaround that by linking |
6300b3ec LP |
14424 | a copy of a good part of our code into each of these |
14425 | libraries again and again, which, however, makes certain | |
14426 | things hard to do, like sharing static variables. Also, it | |
14427 | substantially increases footprint. With this change, there | |
14428 | is only one library for the basic APIs systemd | |
14429 | provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h", | |
14430 | "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this | |
14431 | library as well, however are subject to the --enable-kdbus | |
14432 | switch (see below). Note that "sd-dhcp-client.h" is not part | |
14433 | of this library (this is because it only consumes, never | |
14434 | provides, services of/to other APIs). To make the transition | |
8b7d0494 | 14435 | easy from the separate libraries to the unified one, we |
4c2413bf | 14436 | provide the --enable-compat-libs compile-time switch which |
e49b5aad LP |
14437 | will generate stub libraries that are compatible with the |
14438 | old ones but redirect all calls to the new one. | |
14439 | ||
8b7d0494 | 14440 | * All of the kdbus logic and the new APIs "sd-bus.h", |
e49b5aad | 14441 | "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h", |
8b7d0494 JSJ |
14442 | and "sd-utf8.h" are compile-time optional via the |
14443 | "--enable-kdbus" switch, and they are not compiled in by | |
14444 | default. To make use of kdbus, you have to explicitly enable | |
4c2413bf | 14445 | the switch. Note however, that neither the kernel nor the |
e49b5aad LP |
14446 | userspace API for all of this is considered stable yet. We |
14447 | want to maintain the freedom to still change the APIs for | |
4c2413bf | 14448 | now. By specifying this build-time switch, you acknowledge |
e49b5aad | 14449 | that you are aware of the instability of the current |
ad42cf73 KS |
14450 | APIs. |
14451 | ||
14452 | * Also, note that while kdbus is pretty much complete, | |
e49b5aad | 14453 | it lacks one thing: proper policy support. This means you |
8b7d0494 | 14454 | can build a fully working system with all features; however, |
4c2413bf JE |
14455 | it will be highly insecure. Policy support will be added in |
14456 | one of the next releases, at the same time that we will | |
14457 | declare the APIs stable. | |
e49b5aad | 14458 | |
81c7dd89 | 14459 | * When the kernel command line argument "kdbus" is specified, |
ad42cf73 | 14460 | systemd will automatically load the kdbus.ko kernel module. At |
8b7d0494 | 14461 | this stage of development, it is only useful for testing kdbus |
ad42cf73 | 14462 | and should not be used in production. Note: if "--enable-kdbus" |
8b7d0494 | 14463 | is specified, and the kdbus.ko kernel module is available, and |
ad42cf73 KS |
14464 | "kdbus" is added to the kernel command line, the entire system |
14465 | runs with kdbus instead of dbus-daemon, with the above mentioned | |
14466 | problem of missing the system policy enforcement. Also a future | |
14467 | version of kdbus.ko or a newer systemd will not be compatible with | |
14468 | each other, and will unlikely be able to boot the machine if only | |
14469 | one of them is updated. | |
14470 | ||
e49b5aad | 14471 | * systemctl gained a new "import-environment" command which |
4c2413bf | 14472 | uploads the caller's environment (or parts thereof) into the |
e49b5aad LP |
14473 | service manager so that it is inherited by services started |
14474 | by the manager. This is useful to upload variables like | |
14475 | $DISPLAY into the user service manager. | |
14476 | ||
14477 | * A new PrivateDevices= switch has been added to service units | |
14478 | which allows running a service with a namespaced /dev | |
14479 | directory that does not contain any device nodes for | |
4c2413bf | 14480 | physical devices. More specifically, it only includes devices |
8b7d0494 | 14481 | such as /dev/null, /dev/urandom, and /dev/zero which are API |
e49b5aad LP |
14482 | entry points. |
14483 | ||
14484 | * logind has been extended to support behaviour like VT | |
14485 | switching on seats that do not support a VT. This makes | |
14486 | multi-session available on seats that are not the first seat | |
14487 | (seat0), and on systems where kernel support for VTs has | |
8b7d0494 | 14488 | been disabled at compile-time. |
e49b5aad LP |
14489 | |
14490 | * If a process holds a delay lock for system sleep or shutdown | |
1e190502 | 14491 | and fails to release it in time, we will now log its |
e49b5aad LP |
14492 | identity. This makes it easier to identify processes that |
14493 | cause slow suspends or power-offs. | |
14494 | ||
1e190502 ZJS |
14495 | * When parsing /etc/crypttab, support for a new key-slot= |
14496 | option as supported by Debian is added. It allows indicating | |
14497 | which LUKS slot to use on disk, speeding up key loading. | |
e49b5aad | 14498 | |
000b1ba5 | 14499 | * The sd_journal_sendv() API call has been checked and |
1e190502 ZJS |
14500 | officially declared to be async-signal-safe so that it may |
14501 | be invoked from signal handlers for logging purposes. | |
e49b5aad LP |
14502 | |
14503 | * Boot-time status output is now enabled automatically after a | |
14504 | short timeout if boot does not progress, in order to give | |
8e420494 | 14505 | the user an indication what she or he is waiting for. |
1e190502 ZJS |
14506 | |
14507 | * The boot-time output has been improved to show how much time | |
14508 | remains until jobs expire. | |
e49b5aad LP |
14509 | |
14510 | * The KillMode= switch in service units gained a new possible | |
8b7d0494 | 14511 | value "mixed". If set, and the unit is shut down, then the |
e49b5aad | 14512 | initial SIGTERM signal is sent only to the main daemon |
8e420494 | 14513 | process, while the following SIGKILL signal is sent to |
e49b5aad LP |
14514 | all remaining processes of the service. |
14515 | ||
4c2413bf JE |
14516 | * When a scope unit is registered, a new property "Controller" |
14517 | may be set. If set to a valid bus name, systemd will send a | |
e49b5aad LP |
14518 | RequestStop() signal to this name when it would like to shut |
14519 | down the scope. This may be used to hook manager logic into | |
14520 | the shutdown logic of scope units. Also, scope units may now | |
8b7d0494 | 14521 | be put in a special "abandoned" state, in which case the |
e49b5aad LP |
14522 | manager process which created them takes no further |
14523 | responsibilities for it. | |
14524 | ||
1e190502 | 14525 | * When reading unit files, systemd will now verify |
e49b5aad LP |
14526 | the access mode of these files, and warn about certain |
14527 | suspicious combinations. This has been added to make it | |
14528 | easier to track down packaging bugs where unit files are | |
14529 | marked executable or world-writable. | |
14530 | ||
14531 | * systemd-nspawn gained a new "--setenv=" switch to set | |
8b7d0494 | 14532 | container-wide environment variables. The similar option in |
1e190502 ZJS |
14533 | systemd-activate was renamed from "--environment=" to |
14534 | "--setenv=" for consistency. | |
e49b5aad LP |
14535 | |
14536 | * systemd-nspawn has been updated to create a new kdbus domain | |
14537 | for each container that is invoked, thus allowing each | |
b9761003 | 14538 | container to have its own set of system and user buses, |
8b7d0494 | 14539 | independent of the host. |
e49b5aad LP |
14540 | |
14541 | * systemd-nspawn gained a new --drop-capability= switch to run | |
14542 | the container with less capabilities than the default. Both | |
b9761003 | 14543 | --drop-capability= and --capability= now take the special |
e49b5aad LP |
14544 | string "all" for dropping or keeping all capabilities. |
14545 | ||
14546 | * systemd-nspawn gained new switches for executing containers | |
14547 | with specific SELinux labels set. | |
14548 | ||
14549 | * systemd-nspawn gained a new --quiet switch to not generate | |
14550 | any additional output but the container's own console | |
14551 | output. | |
14552 | ||
14553 | * systemd-nspawn gained a new --share-system switch to run a | |
14554 | container without PID namespacing enabled. | |
14555 | ||
14556 | * systemd-nspawn gained a new --register= switch to control | |
1e190502 | 14557 | whether the container is registered with systemd-machined or |
8e420494 | 14558 | not. This is useful for containers that do not run full |
e49b5aad LP |
14559 | OS images, but only specific apps. |
14560 | ||
14561 | * systemd-nspawn gained a new --keep-unit which may be used | |
8b7d0494 | 14562 | when invoked as the only program from a service unit, and |
e49b5aad | 14563 | results in registration of the unit service itself in |
1e190502 | 14564 | systemd-machined, instead of a newly opened scope unit. |
e49b5aad LP |
14565 | |
14566 | * systemd-nspawn gained a new --network-interface= switch for | |
14567 | moving arbitrary interfaces to the container. The new | |
4c2413bf | 14568 | --network-veth switch creates a virtual Ethernet connection |
8b7d0494 JSJ |
14569 | between host and container. The new --network-bridge= |
14570 | switch then allows assigning the host side of this virtual | |
14571 | Ethernet connection to a bridge device. | |
e49b5aad | 14572 | |
6afc95b7 LP |
14573 | * systemd-nspawn gained a new --personality= switch for |
14574 | setting the kernel personality for the container. This is | |
70a44afe | 14575 | useful when running a 32-bit container on a 64-bit host. A |
b8bde116 JE |
14576 | similar option Personality= is now also available for service |
14577 | units to use. | |
6afc95b7 | 14578 | |
e49b5aad LP |
14579 | * logind will now also track a "Desktop" identifier for each |
14580 | session which encodes the desktop environment of it. This is | |
14581 | useful for desktop environments that want to identify | |
14582 | multiple running sessions of itself easily. | |
14583 | ||
14584 | * A new SELinuxContext= setting for service units has been | |
14585 | added that allows setting a specific SELinux execution | |
14586 | context for a service. | |
14587 | ||
14588 | * Most systemd client tools will now honour $SYSTEMD_LESS for | |
14589 | settings of the "less" pager. By default, these tools will | |
8b7d0494 JSJ |
14590 | override $LESS to allow certain operations to work, such as |
14591 | jump-to-the-end. With $SYSTEMD_LESS, it is possible to | |
e49b5aad LP |
14592 | influence this logic. |
14593 | ||
14594 | * systemd's "seccomp" hook-up has been changed to make use of | |
14595 | the libseccomp library instead of using its own | |
14596 | implementation. This has benefits for portability among | |
14597 | other things. | |
14598 | ||
4c2413bf | 14599 | * For usage together with SystemCallFilter=, a new |
8b7d0494 | 14600 | SystemCallErrorNumber= setting has been introduced that |
b8bde116 JE |
14601 | allows configuration of a system error number to be returned |
14602 | on filtered system calls, instead of immediately killing the | |
e49b5aad LP |
14603 | process. Also, SystemCallArchitectures= has been added to |
14604 | limit access to system calls of a particular architecture | |
14605 | (in order to turn off support for unused secondary | |
4c2413bf | 14606 | architectures). There is also a global |
8b7d0494 | 14607 | SystemCallArchitectures= setting in system.conf now to turn |
e49b5aad LP |
14608 | off support for non-native system calls system-wide. |
14609 | ||
210054d7 KS |
14610 | * systemd requires a kernel with a working name_to_handle_at(), |
14611 | please see the kernel config requirements in the README file. | |
14612 | ||
e49b5aad LP |
14613 | Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov, |
14614 | Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera, | |
14615 | Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters, | |
14616 | Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J | |
14617 | Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa, | |
14618 | David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov, | |
14619 | Elia Pinto, Florian Weimer, George McCollister, Goffredo | |
14620 | Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor | |
14621 | Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld, | |
14622 | Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose | |
14623 | Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg, | |
14624 | Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz | |
14625 | Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas, | |
14626 | Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de | |
14627 | Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael | |
14628 | Marineau, Michael Scherer, Michał Górny, Michal Sekletar, | |
14629 | Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt, | |
14630 | Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien | |
14631 | Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters, | |
14632 | Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else, | |
14633 | Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen, | |
14634 | Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav | |
14635 | Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang | |
14636 | Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek | |
14637 | ||
ccddd104 | 14638 | — Berlin, 2014-02-20 |
e49b5aad | 14639 | |
cd4010b3 LP |
14640 | CHANGES WITH 208: |
14641 | ||
14642 | * logind has gained support for facilitating privileged input | |
14643 | and drm device access for unprivileged clients. This work is | |
14644 | useful to allow Wayland display servers (and similar | |
14645 | programs, such as kmscon) to run under the user's ID and | |
14646 | access input and drm devices which are normally | |
14647 | protected. When this is used (and the kernel is new enough) | |
14648 | logind will "mute" IO on the file descriptors passed to | |
14649 | Wayland as long as it is in the background and "unmute" it | |
14650 | if it returns into the foreground. This allows secure | |
14651 | session switching without allowing background sessions to | |
14652 | eavesdrop on input and display data. This also introduces | |
14653 | session switching support if VT support is turned off in the | |
14654 | kernel, and on seats that are not seat0. | |
14655 | ||
14656 | * A new kernel command line option luks.options= is understood | |
06b643e7 | 14657 | now which allows specifying LUKS options for usage for LUKS |
cd4010b3 LP |
14658 | encrypted partitions specified with luks.uuid=. |
14659 | ||
14660 | * tmpfiles.d(5) snippets may now use specifier expansion in | |
14661 | path names. More specifically %m, %b, %H, %v, are now | |
14662 | replaced by the local machine id, boot id, hostname, and | |
14663 | kernel version number. | |
14664 | ||
14665 | * A new tmpfiles.d(5) command "m" has been introduced which | |
14666 | may be used to change the owner/group/access mode of a file | |
d28315e4 | 14667 | or directory if it exists, but do nothing if it does not. |
cd4010b3 LP |
14668 | |
14669 | * This release removes high-level support for the | |
14670 | MemorySoftLimit= cgroup setting. The underlying kernel | |
14671 | cgroup attribute memory.soft_limit= is currently badly | |
14672 | designed and likely to be removed from the kernel API in its | |
d28315e4 | 14673 | current form, hence we should not expose it for now. |
cd4010b3 LP |
14674 | |
14675 | * The memory.use_hierarchy cgroup attribute is now enabled for | |
14676 | all cgroups systemd creates in the memory cgroup | |
14677 | hierarchy. This option is likely to be come the built-in | |
cc98b302 TH |
14678 | default in the kernel anyway, and the non-hierarchical mode |
14679 | never made much sense in the intrinsically hierarchical | |
cd4010b3 LP |
14680 | cgroup system. |
14681 | ||
14682 | * A new field _SYSTEMD_SLICE= is logged along with all journal | |
14683 | messages containing the slice a message was generated | |
14684 | from. This is useful to allow easy per-customer filtering of | |
14685 | logs among other things. | |
14686 | ||
14687 | * systemd-journald will no longer adjust the group of journal | |
14688 | files it creates to the "systemd-journal" group. Instead we | |
14689 | rely on the journal directory to be owned by the | |
14690 | "systemd-journal" group, and its setgid bit set, so that the | |
14691 | kernel file system layer will automatically enforce that | |
14692 | journal files inherit this group assignment. The reason for | |
14693 | this change is that we cannot allow NSS look-ups from | |
14694 | journald which would be necessary to resolve | |
14695 | "systemd-journal" to a numeric GID, because this might | |
14696 | create deadlocks if NSS involves synchronous queries to | |
14697 | other daemons (such as nscd, or sssd) which in turn are | |
14698 | logging clients of journald and might block on it, which | |
14699 | would then dead lock. A tmpfiles.d(5) snippet included in | |
14700 | systemd will make sure the setgid bit and group are | |
14701 | properly set on the journal directory if it exists on every | |
14702 | boot. However, we recommend adjusting it manually after | |
14703 | upgrades too (or from RPM scriptlets), so that the change is | |
14704 | not delayed until next reboot. | |
14705 | ||
14706 | * Backlight and random seed files in /var/lib/ have moved into | |
14707 | the /var/lib/systemd/ directory, in order to centralize all | |
14708 | systemd generated files in one directory. | |
14709 | ||
14710 | * Boot time performance measurements (as displayed by | |
14711 | "systemd-analyze" for example) will now read ACPI 5.0 FPDT | |
14712 | performance information if that's available to determine how | |
14713 | much time BIOS and boot loader initialization required. With | |
14714 | a sufficiently new BIOS you hence no longer need to boot | |
14715 | with Gummiboot to get access to such information. | |
14716 | ||
14717 | Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters, | |
14718 | Cristian Rodríguez, Dave Reisner, David Herrmann, David | |
14719 | Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao | |
14720 | feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart | |
14721 | Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt, | |
14722 | Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty, | |
14723 | Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek | |
14724 | ||
ccddd104 | 14725 | — Berlin, 2013-10-02 |
cd4010b3 | 14726 | |
4f0be680 LP |
14727 | CHANGES WITH 207: |
14728 | ||
14729 | * The Restart= option for services now understands a new | |
f3a165b0 | 14730 | on-watchdog setting, which will restart the service |
4f0be680 LP |
14731 | automatically if the service stops sending out watchdog keep |
14732 | alive messages (as configured with WatchdogSec=). | |
14733 | ||
14734 | * The getty generator (which is responsible for bringing up a | |
14735 | getty on configured serial consoles) will no longer only | |
14736 | start a getty on the primary kernel console but on all | |
14737 | others, too. This makes the order in which console= is | |
14738 | specified on the kernel command line less important. | |
14739 | ||
14740 | * libsystemd-logind gained a new sd_session_get_vt() call to | |
14741 | retrieve the VT number of a session. | |
14742 | ||
14743 | * If the option "tries=0" is set for an entry of /etc/crypttab | |
14744 | its passphrase is queried indefinitely instead of any | |
14745 | maximum number of tries. | |
14746 | ||
14747 | * If a service with a configure PID file terminates its PID | |
14748 | file will now be removed automatically if it still exists | |
14749 | afterwards. This should put an end to stale PID files. | |
14750 | ||
14751 | * systemd-run will now also take relative binary path names | |
14752 | for execution and no longer insists on absolute paths. | |
14753 | ||
14754 | * InaccessibleDirectories= and ReadOnlyDirectories= now take | |
14755 | paths that are optionally prefixed with "-" to indicate that | |
d28315e4 | 14756 | it should not be considered a failure if they do not exist. |
4f0be680 | 14757 | |
f3a165b0 KS |
14758 | * journalctl -o (and similar commands) now understands a new |
14759 | output mode "short-precise", it is similar to "short" but | |
4f0be680 LP |
14760 | shows timestamps with usec accuracy. |
14761 | ||
14762 | * The option "discard" (as known from Debian) is now | |
14763 | synonymous to "allow-discards" in /etc/crypttab. In fact, | |
387abf80 | 14764 | "discard" is preferred now (since it is easier to remember |
4f0be680 LP |
14765 | and type). |
14766 | ||
f3a165b0 | 14767 | * Some licensing clean-ups were made, so that more code is now |
4f0be680 LP |
14768 | LGPL-2.1 licensed than before. |
14769 | ||
14770 | * A minimal tool to save/restore the display backlight | |
14771 | brightness across reboots has been added. It will store the | |
f3a165b0 | 14772 | backlight setting as late as possible at shutdown, and |
4f0be680 LP |
14773 | restore it as early as possible during reboot. |
14774 | ||
14775 | * A logic to automatically discover and enable home and swap | |
14776 | partitions on GPT disks has been added. With this in place | |
14777 | /etc/fstab becomes optional for many setups as systemd can | |
14778 | discover certain partitions located on the root disk | |
14779 | automatically. Home partitions are recognized under their | |
14780 | GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap | |
14781 | partitions are recognized under their GPT type ID | |
14782 | 0657fd6da4ab43c484e50933c84b4f4f. | |
14783 | ||
14784 | * systemd will no longer pass any environment from the kernel | |
14785 | or initrd to system services. If you want to set an | |
14786 | environment for all services, do so via the kernel command | |
14787 | line systemd.setenv= assignment. | |
14788 | ||
387abf80 LP |
14789 | * The systemd-sysctl tool no longer natively reads the file |
14790 | /etc/sysctl.conf. If desired, the file should be symlinked | |
14791 | from /etc/sysctl.d/99-sysctl.conf. Apart from providing | |
14792 | legacy support by a symlink rather than built-in code, it | |
14793 | also makes the otherwise hidden order of application of the | |
14794 | different files visible. (Note that this partly reverts to a | |
14795 | pre-198 application order of sysctl knobs!) | |
04bf3c1a | 14796 | |
4f0be680 LP |
14797 | * The "systemctl set-log-level" and "systemctl dump" commands |
14798 | have been moved to systemd-analyze. | |
14799 | ||
14800 | * systemd-run learned the new --remain-after-exit switch, | |
14801 | which causes the scope unit not to be cleaned up | |
14802 | automatically after the process terminated. | |
14803 | ||
14804 | * tmpfiles learned a new --exclude-prefix= switch to exclude | |
14805 | certain paths from operation. | |
14806 | ||
14807 | * journald will now automatically flush all messages to disk | |
f47ad593 ZJS |
14808 | as soon as a message at the log level CRIT, ALERT or EMERG |
14809 | is received. | |
4f0be680 LP |
14810 | |
14811 | Contributions from: Andrew Cook, Brandon Philips, Christian | |
14812 | Hesse, Christoph Junghans, Colin Walters, Daniel Schaal, | |
14813 | Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George | |
14814 | McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer, | |
14815 | Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt, | |
14816 | Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering, | |
14817 | Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel | |
14818 | Holtmann, Martin Pitt, Michael Biebl, Michael Marineau, | |
14819 | Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał | |
14820 | Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn | |
14821 | Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe | |
14822 | Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao, | |
14823 | William Giokas, Zbigniew Jędrzejewski-Szmek | |
14824 | ||
ccddd104 | 14825 | — Berlin, 2013-09-13 |
4f0be680 | 14826 | |
408f281b LP |
14827 | CHANGES WITH 206: |
14828 | ||
14829 | * The documentation has been updated to cover the various new | |
14830 | concepts introduced with 205. | |
14831 | ||
14832 | * Unit files now understand the new %v specifier which | |
14833 | resolves to the kernel version string as returned by "uname | |
14834 | -r". | |
14835 | ||
14836 | * systemctl now supports filtering the unit list output by | |
14837 | load state, active state and sub state, using the new | |
33b521be | 14838 | --state= parameter. |
408f281b LP |
14839 | |
14840 | * "systemctl status" will now show the results of the | |
14841 | condition checks (like ConditionPathExists= and similar) of | |
14842 | the last start attempts of the unit. They are also logged to | |
14843 | the journal. | |
14844 | ||
14845 | * "journalctl -b" may now be used to look for boot output of a | |
14846 | specific boot. Try "journalctl -b -1" for the previous boot, | |
14847 | but the syntax is substantially more powerful. | |
14848 | ||
14849 | * "journalctl --show-cursor" has been added which prints the | |
14850 | cursor string the last shown log line. This may then be used | |
14851 | with the new "journalctl --after-cursor=" switch to continue | |
14852 | browsing logs from that point on. | |
14853 | ||
14854 | * "journalctl --force" may now be used to force regeneration | |
14855 | of an FSS key. | |
14856 | ||
251cc819 LP |
14857 | * Creation of "dead" device nodes has been moved from udev |
14858 | into kmod and tmpfiles. Previously, udev would read the kmod | |
14859 | databases to pre-generate dead device nodes based on meta | |
14860 | information contained in kernel modules, so that these would | |
14861 | be auto-loaded on access rather then at boot. As this | |
d28315e4 | 14862 | does not really have much to do with the exposing actual |
251cc819 LP |
14863 | kernel devices to userspace this has always been slightly |
14864 | alien in the udev codebase. Following the new scheme kmod | |
14865 | will now generate a runtime snippet for tmpfiles from the | |
14866 | module meta information and it now is tmpfiles' job to the | |
14867 | create the nodes. This also allows overriding access and | |
14868 | other parameters for the nodes using the usual tmpfiles | |
14869 | facilities. As side effect this allows us to remove the | |
14870 | CAP_SYS_MKNOD capability bit from udevd entirely. | |
14871 | ||
14872 | * logind's device ACLs may now be applied to these "dead" | |
14873 | devices nodes too, thus finally allowing managed access to | |
ce830873 | 14874 | devices such as /dev/snd/sequencer without loading the |
251cc819 | 14875 | backing module right-away. |
408f281b LP |
14876 | |
14877 | * A new RPM macro has been added that may be used to apply | |
14878 | tmpfiles configuration during package installation. | |
14879 | ||
14880 | * systemd-detect-virt and ConditionVirtualization= now can | |
14881 | detect User-Mode-Linux machines (UML). | |
14882 | ||
251cc819 LP |
14883 | * journald will now implicitly log the effective capabilities |
14884 | set of processes in the message metadata. | |
408f281b LP |
14885 | |
14886 | * systemd-cryptsetup has gained support for TrueCrypt volumes. | |
14887 | ||
14888 | * The initrd interface has been simplified (more specifically, | |
14889 | support for passing performance data via environment | |
14890 | variables and fsck results via files in /run has been | |
14891 | removed). These features were non-essential, and are | |
14892 | nowadays available in a much nicer way by having systemd in | |
14893 | the initrd serialize its state and have the hosts systemd | |
14894 | deserialize it again. | |
14895 | ||
28f5c779 KS |
14896 | * The udev "keymap" data files and tools to apply keyboard |
14897 | specific mappings of scan to key codes, and force-release | |
14898 | scan code lists have been entirely replaced by a udev | |
14899 | "keyboard" builtin and a hwdb data file. | |
408f281b | 14900 | |
251cc819 LP |
14901 | * systemd will now honour the kernel's "quiet" command line |
14902 | argument also during late shutdown, resulting in a | |
14903 | completely silent shutdown when used. | |
14904 | ||
14905 | * There's now an option to control the SO_REUSEPORT socket | |
14906 | option in .socket units. | |
14907 | ||
14908 | * Instance units will now automatically get a per-template | |
14909 | subslice of system.slice unless something else is explicitly | |
14910 | configured. For example, instances of sshd@.service will now | |
14911 | implicitly be placed in system-sshd.slice rather than | |
14912 | system.slice as before. | |
14913 | ||
14914 | * Test coverage support may now be enabled at build time. | |
14915 | ||
14916 | Contributions from: Dave Reisner, Frederic Crozat, Harald | |
14917 | Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan | |
14918 | Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart | |
14919 | Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael | |
14920 | Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden, | |
14921 | Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William | |
14922 | Giokas, Zbigniew Jędrzejewski-Szmek | |
14923 | ||
ccddd104 | 14924 | — Berlin, 2013-07-23 |
4f0be680 | 14925 | |
00aa832b LP |
14926 | CHANGES WITH 205: |
14927 | ||
14928 | * Two new unit types have been introduced: | |
14929 | ||
14930 | Scope units are very similar to service units, however, are | |
ccddd104 | 14931 | created out of pre-existing processes — instead of PID 1 |
00aa832b LP |
14932 | forking off the processes. By using scope units it is |
14933 | possible for system services and applications to group their | |
14934 | own child processes (worker processes) in a powerful way | |
14935 | which then maybe used to organize them, or kill them | |
14936 | together, or apply resource limits on them. | |
14937 | ||
14938 | Slice units may be used to partition system resources in an | |
cc98b302 | 14939 | hierarchical fashion and then assign other units to them. By |
00aa832b LP |
14940 | default there are now three slices: system.slice (for all |
14941 | system services), user.slice (for all user sessions), | |
14942 | machine.slice (for VMs and containers). | |
14943 | ||
14944 | Slices and scopes have been introduced primarily in | |
14945 | context of the work to move cgroup handling to a | |
14946 | single-writer scheme, where only PID 1 | |
14947 | creates/removes/manages cgroups. | |
14948 | ||
14949 | * There's a new concept of "transient" units. In contrast to | |
14950 | normal units these units are created via an API at runtime, | |
14951 | not from configuration from disk. More specifically this | |
14952 | means it is now possible to run arbitrary programs as | |
14953 | independent services, with all execution parameters passed | |
14954 | in via bus APIs rather than read from disk. Transient units | |
14955 | make systemd substantially more dynamic then it ever was, | |
14956 | and useful as a general batch manager. | |
14957 | ||
14958 | * logind has been updated to make use of scope and slice units | |
14959 | for managing user sessions. As a user logs in he will get | |
14960 | his own private slice unit, to which all sessions are added | |
14961 | as scope units. We also added support for automatically | |
14962 | adding an instance of user@.service for the user into the | |
14963 | slice. Effectively logind will no longer create cgroup | |
14964 | hierarchies on its own now, it will defer entirely to PID 1 | |
14965 | for this by means of scope, service and slice units. Since | |
14966 | user sessions this way become entities managed by PID 1 | |
14967 | the output of "systemctl" is now a lot more comprehensive. | |
14968 | ||
14969 | * A new mini-daemon "systemd-machined" has been added which | |
14970 | may be used by virtualization managers to register local | |
14971 | VMs/containers. nspawn has been updated accordingly, and | |
14972 | libvirt will be updated shortly. machined will collect a bit | |
14973 | of meta information about the VMs/containers, and assign | |
14974 | them their own scope unit (see above). The collected | |
14975 | meta-data is then made available via the "machinectl" tool, | |
14976 | and exposed in "ps" and similar tools. machined/machinectl | |
14977 | is compile-time optional. | |
14978 | ||
14979 | * As discussed earlier, the low-level cgroup configuration | |
14980 | options ControlGroup=, ControlGroupModify=, | |
14981 | ControlGroupPersistent=, ControlGroupAttribute= have been | |
14982 | removed. Please use high-level attribute settings instead as | |
14983 | well as slice units. | |
14984 | ||
14985 | * A new bus call SetUnitProperties() has been added to alter | |
14986 | various runtime parameters of a unit. This is primarily | |
14987 | useful to alter cgroup parameters dynamically in a nice way, | |
14988 | but will be extended later on to make more properties | |
14989 | modifiable at runtime. systemctl gained a new set-properties | |
14990 | command that wraps this call. | |
14991 | ||
14992 | * A new tool "systemd-run" has been added which can be used to | |
14993 | run arbitrary command lines as transient services or scopes, | |
14994 | while configuring a number of settings via the command | |
14995 | line. This tool is currently very basic, however already | |
14996 | very useful. We plan to extend this tool to even allow | |
14997 | queuing of execution jobs with time triggers from the | |
14998 | command line, similar in fashion to "at". | |
14999 | ||
15000 | * nspawn will now inform the user explicitly that kernels with | |
15001 | audit enabled break containers, and suggest the user to turn | |
15002 | off audit. | |
15003 | ||
15004 | * Support for detecting the IMA and AppArmor security | |
15005 | frameworks with ConditionSecurity= has been added. | |
15006 | ||
15007 | * journalctl gained a new "-k" switch for showing only kernel | |
1fda0ab5 ZJS |
15008 | messages, mimicking dmesg output; in addition to "--user" |
15009 | and "--system" switches for showing only user's own logs | |
15010 | and system logs. | |
00aa832b LP |
15011 | |
15012 | * systemd-delta can now show information about drop-in | |
15013 | snippets extending unit files. | |
15014 | ||
15015 | * libsystemd-bus has been substantially updated but is still | |
15016 | not available as public API. | |
15017 | ||
15018 | * systemd will now look for the "debug" argument on the kernel | |
499b604b | 15019 | command line and enable debug logging, similar to what |
00aa832b LP |
15020 | "systemd.log_level=debug" already did before. |
15021 | ||
15022 | * "systemctl set-default", "systemctl get-default" has been | |
15023 | added to configure the default.target symlink, which | |
15024 | controls what to boot into by default. | |
15025 | ||
1fda0ab5 ZJS |
15026 | * "systemctl set-log-level" has been added as a convenient |
15027 | way to raise and lower systemd logging threshold. | |
15028 | ||
00aa832b LP |
15029 | * "systemd-analyze plot" will now show the time the various |
15030 | generators needed for execution, as well as information | |
15031 | about the unit file loading. | |
15032 | ||
00aa832b LP |
15033 | * libsystemd-journal gained a new sd_journal_open_files() call |
15034 | for opening specific journal files. journactl also gained a | |
15035 | new switch to expose this new functionality. Previously we | |
15036 | only supported opening all files from a directory, or all | |
15037 | files from the system, as opening individual files only is | |
15038 | racy due to journal file rotation. | |
15039 | ||
15040 | * systemd gained the new DefaultEnvironment= setting in | |
15041 | /etc/systemd/system.conf to set environment variables for | |
15042 | all services. | |
15043 | ||
15044 | * If a privileged process logs a journal message with the | |
15045 | OBJECT_PID= field set, then journald will automatically | |
15046 | augment this with additional OBJECT_UID=, OBJECT_GID=, | |
1d3a473b | 15047 | OBJECT_COMM=, OBJECT_EXE=, … fields. This is useful if |
00aa832b LP |
15048 | system services want to log events about specific client |
15049 | processes. journactl/systemctl has been updated to make use | |
15050 | of this information if all log messages regarding a specific | |
15051 | unit is requested. | |
15052 | ||
15053 | Contributions from: Auke Kok, Chengwei Yang, Colin Walters, | |
15054 | Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave | |
15055 | Reisner, David Coppa, David King, David Strauss, Eelco | |
15056 | Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander | |
15057 | Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan | |
15058 | Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart | |
15059 | Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer, | |
15060 | Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer, | |
15061 | Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan, | |
15062 | Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern, | |
15063 | Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar, | |
15064 | Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek, | |
15065 | Łukasz Stelmach, 장동준 | |
15066 | ||
606c24e3 LP |
15067 | CHANGES WITH 204: |
15068 | ||
15069 | * The Python bindings gained some minimal support for the APIs | |
15070 | exposed by libsystemd-logind. | |
15071 | ||
15072 | * ConditionSecurity= gained support for detecting SMACK. Since | |
15073 | this condition already supports SELinux and AppArmor we only | |
15074 | miss IMA for this. Patches welcome! | |
15075 | ||
15076 | Contributions from: Karol Lewandowski, Lennart Poettering, | |
15077 | Zbigniew Jędrzejewski-Szmek | |
15078 | ||
2f3fcf85 LP |
15079 | CHANGES WITH 203: |
15080 | ||
15081 | * systemd-nspawn will now create /etc/resolv.conf if | |
15082 | necessary, before bind-mounting the host's file onto it. | |
15083 | ||
15084 | * systemd-nspawn will now store meta information about a | |
15085 | container on the container's cgroup as extended attribute | |
15086 | fields, including the root directory. | |
15087 | ||
15088 | * The cgroup hierarchy has been reworked in many ways. All | |
15089 | objects any of the components systemd creates in the cgroup | |
b82eed9a | 15090 | tree are now suffixed. More specifically, user sessions are |
2f3fcf85 LP |
15091 | now placed in cgroups suffixed with ".session", users in |
15092 | cgroups suffixed with ".user", and nspawn containers in | |
15093 | cgroups suffixed with ".nspawn". Furthermore, all cgroup | |
15094 | names are now escaped in a simple scheme to avoid collision | |
15095 | of userspace object names with kernel filenames. This work | |
15096 | is preparation for making these objects relocatable in the | |
15097 | cgroup tree, in order to allow easy resource partitioning of | |
15098 | these objects without causing naming conflicts. | |
15099 | ||
15100 | * systemctl list-dependencies gained the new switches | |
15101 | --plain, --reverse, --after and --before. | |
15102 | ||
15103 | * systemd-inhibit now shows the process name of processes that | |
15104 | have taken an inhibitor lock. | |
15105 | ||
15106 | * nss-myhostname will now also resolve "localhost" | |
15107 | implicitly. This makes /etc/hosts an optional file and | |
15108 | nicely handles that on IPv6 ::1 maps to both "localhost" and | |
15109 | the local hostname. | |
15110 | ||
15111 | * libsystemd-logind.so gained a new call | |
15112 | sd_get_machine_names() to enumerate running containers and | |
15113 | VMs (currently only supported by very new libvirt and | |
15114 | nspawn). sd_login_monitor can now be used to watch | |
15115 | VMs/containers coming and going. | |
15116 | ||
15117 | * .include is not allowed recursively anymore, and only in | |
15118 | unit files. Usually it is better to use drop-in snippets in | |
15119 | .d/*.conf anyway, as introduced with systemd 198. | |
15120 | ||
15121 | * systemd-analyze gained a new "critical-chain" command that | |
15122 | determines the slowest chain of units run during system | |
15123 | boot-up. It is very useful for tracking down where | |
15124 | optimizing boot time is the most beneficial. | |
15125 | ||
15126 | * systemd will no longer allow manipulating service paths in | |
15127 | the name=systemd:/system cgroup tree using ControlGroup= in | |
15128 | units. (But is still fine with it in all other dirs.) | |
15129 | ||
15130 | * There's a new systemd-nspawn@.service service file that may | |
15131 | be used to easily run nspawn containers as system | |
15132 | services. With the container's root directory in | |
15133 | /var/lib/container/foobar it is now sufficient to run | |
15134 | "systemctl start systemd-nspawn@foobar.service" to boot it. | |
15135 | ||
15136 | * systemd-cgls gained a new parameter "--machine" to list only | |
15137 | the processes within a certain container. | |
15138 | ||
15139 | * ConditionSecurity= now can check for "apparmor". We still | |
15140 | are lacking checks for SMACK and IMA for this condition | |
15141 | check though. Patches welcome! | |
15142 | ||
15143 | * A new configuration file /etc/systemd/sleep.conf has been | |
15144 | added that may be used to configure which kernel operation | |
15145 | systemd is supposed to execute when "suspend", "hibernate" | |
15146 | or "hybrid-sleep" is requested. This makes the new kernel | |
15147 | "freeze" state accessible to the user. | |
15148 | ||
15149 | * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape | |
15150 | the passed argument if applicable. | |
15151 | ||
15152 | Contributions from: Auke Kok, Colin Guthrie, Colin Walters, | |
15153 | Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner, | |
15154 | Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh | |
15155 | Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn, | |
15156 | MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel | |
15157 | Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom | |
15158 | Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew | |
15159 | Jędrzejewski-Szmek | |
15160 | ||
ef3b5246 LP |
15161 | CHANGES WITH 202: |
15162 | ||
15163 | * The output of 'systemctl list-jobs' got some polishing. The | |
15164 | '--type=' argument may now be passed more than once. A new | |
15165 | command 'systemctl list-sockets' has been added which shows | |
15166 | a list of kernel sockets systemd is listening on with the | |
15167 | socket units they belong to, plus the units these socket | |
15168 | units activate. | |
15169 | ||
15170 | * The experimental libsystemd-bus library got substantial | |
15171 | updates to work in conjunction with the (also experimental) | |
15172 | kdbus kernel project. It works well enough to exchange | |
15173 | messages with some sophistication. Note that kdbus is not | |
15174 | ready yet, and the library is mostly an elaborate test case | |
15175 | for now, and not installable. | |
15176 | ||
15177 | * systemd gained a new unit 'systemd-static-nodes.service' | |
15178 | that generates static device nodes earlier during boot, and | |
15179 | can run in conjunction with udev. | |
15180 | ||
15181 | * libsystemd-login gained a new call sd_pid_get_user_unit() | |
15182 | to retrieve the user systemd unit a process is running | |
15183 | in. This is useful for systems where systemd is used as | |
15184 | session manager. | |
15185 | ||
15186 | * systemd-nspawn now places all containers in the new /machine | |
15187 | top-level cgroup directory in the name=systemd | |
15188 | hierarchy. libvirt will soon do the same, so that we get a | |
15189 | uniform separation of /system, /user and /machine for system | |
15190 | services, user processes and containers/virtual | |
15191 | machines. This new cgroup hierarchy is also useful to stick | |
15192 | stable names to specific container instances, which can be | |
7c04ad2d | 15193 | recognized later this way (this name may be controlled |
ef3b5246 LP |
15194 | via systemd-nspawn's new -M switch). libsystemd-login also |
15195 | gained a new call sd_pid_get_machine_name() to retrieve the | |
15196 | name of the container/VM a specific process belongs to. | |
15197 | ||
15198 | * bootchart can now store its data in the journal. | |
15199 | ||
15200 | * libsystemd-journal gained a new call | |
15201 | sd_journal_add_conjunction() for AND expressions to the | |
15202 | matching logic. This can be used to express more complex | |
15203 | logical expressions. | |
15204 | ||
15205 | * journactl can now take multiple --unit= and --user-unit= | |
15206 | switches. | |
15207 | ||
15208 | * The cryptsetup logic now understands the "luks.key=" kernel | |
15209 | command line switch for specifying a file to read the | |
7c04ad2d | 15210 | decryption key from. Also, if a configured key file is not |
ef3b5246 LP |
15211 | found the tool will now automatically fall back to prompting |
15212 | the user. | |
15213 | ||
cbeabcfb ZJS |
15214 | * Python systemd.journal module was updated to wrap recently |
15215 | added functions from libsystemd-journal. The interface was | |
15216 | changed to bring the low level interface in s.j._Reader | |
15217 | closer to the C API, and the high level interface in | |
15218 | s.j.Reader was updated to wrap and convert all data about | |
15219 | an entry. | |
15220 | ||
ef3b5246 LP |
15221 | Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer, |
15222 | Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart | |
15223 | Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer, | |
15224 | Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt, | |
15225 | Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks, | |
15226 | Tom Gundersen, Zbigniew Jędrzejewski-Szmek | |
15227 | ||
d3a86981 LP |
15228 | CHANGES WITH 201: |
15229 | ||
15230 | * journalctl --update-catalog now understands a new --root= | |
15231 | option to operate on catalogs found in a different root | |
15232 | directory. | |
15233 | ||
15234 | * During shutdown after systemd has terminated all running | |
15235 | services a final killing loop kills all remaining left-over | |
15236 | processes. We will now print the name of these processes | |
15237 | when we send SIGKILL to them, since this usually indicates a | |
15238 | problem. | |
15239 | ||
15240 | * If /etc/crypttab refers to password files stored on | |
15241 | configured mount points automatic dependencies will now be | |
15242 | generated to ensure the specific mount is established first | |
15243 | before the key file is attempted to be read. | |
15244 | ||
15245 | * 'systemctl status' will now show information about the | |
15246 | network sockets a socket unit is listening on. | |
15247 | ||
15248 | * 'systemctl status' will also shown information about any | |
15249 | drop-in configuration file for units. (Drop-In configuration | |
15250 | files in this context are files such as | |
5ada98cd | 15251 | /etc/systemd/system/foobar.service.d/*.conf) |
d3a86981 LP |
15252 | |
15253 | * systemd-cgtop now optionally shows summed up CPU times of | |
15254 | cgroups. Press '%' while running cgtop to switch between | |
15255 | percentage and absolute mode. This is useful to determine | |
15256 | which cgroups use up the most CPU time over the entire | |
15257 | runtime of the system. systemd-cgtop has also been updated | |
15258 | to be 'pipeable' for processing with further shell tools. | |
15259 | ||
15260 | * 'hostnamectl set-hostname' will now allow setting of FQDN | |
15261 | hostnames. | |
15262 | ||
15263 | * The formatting and parsing of time span values has been | |
15264 | changed. The parser now understands fractional expressions | |
15265 | such as "5.5h". The formatter will now output fractional | |
15266 | expressions for all time spans under 1min, i.e. "5.123456s" | |
15267 | rather than "5s 123ms 456us". For time spans under 1s | |
15268 | millisecond values are shown, for those under 1ms | |
15269 | microsecond values are shown. This should greatly improve | |
15270 | all time-related output of systemd. | |
15271 | ||
15272 | * libsystemd-login and libsystemd-journal gained new | |
15273 | functions for querying the poll() events mask and poll() | |
15274 | timeout value for integration into arbitrary event | |
15275 | loops. | |
15276 | ||
15277 | * localectl gained the ability to list available X11 keymaps | |
15278 | (models, layouts, variants, options). | |
15279 | ||
15280 | * 'systemd-analyze dot' gained the ability to filter for | |
15281 | specific units via shell-style globs, to create smaller, | |
d28315e4 | 15282 | more useful graphs. I.e. it is now possible to create simple |
d3a86981 LP |
15283 | graphs of all the dependencies between only target units, or |
15284 | of all units that Avahi has dependencies with. | |
15285 | ||
15286 | Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck, | |
15287 | Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly | |
15288 | Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau, | |
15289 | Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal | |
15290 | Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie, | |
15291 | Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav | |
15292 | Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach | |
15293 | ||
9ca3c17f LP |
15294 | CHANGES WITH 200: |
15295 | ||
15296 | * The boot-time readahead implementation for rotating media | |
15297 | will now read the read-ahead data in multiple passes which | |
15298 | consist of all read requests made in equidistant time | |
15299 | intervals. This means instead of strictly reading read-ahead | |
15300 | data in its physical order on disk we now try to find a | |
15301 | middle ground between physical and access time order. | |
15302 | ||
15303 | * /etc/os-release files gained a new BUILD_ID= field for usage | |
15304 | on operating systems that provide continuous builds of OS | |
15305 | images. | |
15306 | ||
15307 | Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers, | |
15308 | Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín | |
15309 | William Douglas, Zbigniew Jędrzejewski-Szmek | |
15310 | ||
35911459 LP |
15311 | CHANGES WITH 199: |
15312 | ||
15313 | * systemd-python gained an API exposing libsystemd-daemon. | |
15314 | ||
15315 | * The SMACK setup logic gained support for uploading CIPSO | |
15316 | security policy. | |
15317 | ||
15318 | * Behaviour of PrivateTmp=, ReadWriteDirectories=, | |
15319 | ReadOnlyDirectories= and InaccessibleDirectories= has | |
15320 | changed. The private /tmp and /var/tmp directories are now | |
15321 | shared by all processes of a service (which means | |
15322 | ExecStartPre= may now leave data in /tmp that ExecStart= of | |
15323 | the same service can still access). When a service is | |
15324 | stopped its temporary directories are immediately deleted | |
a87197f5 | 15325 | (normal clean-up with tmpfiles is still done in addition to |
35911459 LP |
15326 | this though). |
15327 | ||
15328 | * By default, systemd will now set a couple of sysctl | |
15329 | variables in the kernel: the safe sysrq options are turned | |
15330 | on, IP route verification is turned on, and source routing | |
15331 | disabled. The recently added hardlink and softlink | |
15332 | protection of the kernel is turned on. These settings should | |
15333 | be reasonably safe, and good defaults for all new systems. | |
15334 | ||
15335 | * The predictable network naming logic may now be turned off | |
a87197f5 | 15336 | with a new kernel command line switch: net.ifnames=0. |
35911459 LP |
15337 | |
15338 | * A new libsystemd-bus module has been added that implements a | |
15339 | pretty complete D-Bus client library. For details see: | |
15340 | ||
56cadcb6 | 15341 | https://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html |
35911459 | 15342 | |
c20d8298 | 15343 | * journald will now explicitly flush the journal files to disk |
a87197f5 ZJS |
15344 | at the latest 5min after each write. The file will then also |
15345 | be marked offline until the next write. This should increase | |
15346 | reliability in case of a crash. The synchronization delay | |
15347 | can be configured via SyncIntervalSec= in journald.conf. | |
35911459 LP |
15348 | |
15349 | * There's a new remote-fs-setup.target unit that can be used | |
15350 | to pull in specific services when at least one remote file | |
15351 | system is to be mounted. | |
15352 | ||
15353 | * There are new targets timers.target and paths.target as | |
15354 | canonical targets to pull user timer and path units in | |
15355 | from. This complements sockets.target with a similar | |
15356 | purpose for socket units. | |
15357 | ||
6a7d3d68 LP |
15358 | * libudev gained a new call udev_device_set_attribute_value() |
15359 | to set sysfs attributes of a device. | |
15360 | ||
a87197f5 ZJS |
15361 | * The udev daemon now sets the default number of worker |
15362 | processes executed in parallel based on the number of available | |
c20d8298 | 15363 | CPUs instead of the amount of available RAM. This is supposed |
ab06eef8 | 15364 | to provide a more reliable default and limit a too aggressive |
ce830873 | 15365 | parallelism for setups with 1000s of devices connected. |
c20d8298 | 15366 | |
35911459 LP |
15367 | Contributions from: Auke Kok, Colin Walters, Cristian |
15368 | Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes | |
15369 | Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan | |
15370 | Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering, | |
15371 | Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl, | |
15372 | Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen, | |
15373 | Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel | |
15374 | Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, | |
15375 | Zbigniew Jędrzejewski-Szmek | |
15376 | ||
85d68397 LP |
15377 | CHANGES WITH 198: |
15378 | ||
15379 | * Configuration of unit files may now be extended via drop-in | |
15380 | files without having to edit/override the unit files | |
15381 | themselves. More specifically, if the administrator wants to | |
15382 | change one value for a service file foobar.service he can | |
15383 | now do so by dropping in a configuration snippet into | |
ad88e758 | 15384 | /etc/systemd/system/foobar.service.d/*.conf. The unit logic |
85d68397 LP |
15385 | will load all these snippets and apply them on top of the |
15386 | main unit configuration file, possibly extending or | |
15387 | overriding its settings. Using these drop-in snippets is | |
40e21da8 KS |
15388 | generally nicer than the two earlier options for changing |
15389 | unit files locally: copying the files from | |
85d68397 LP |
15390 | /usr/lib/systemd/system/ to /etc/systemd/system/ and editing |
15391 | them there; or creating a new file in /etc/systemd/system/ | |
15392 | that incorporates the original one via ".include". Drop-in | |
15393 | snippets into these .d/ directories can be placed in any | |
fd868975 | 15394 | directory systemd looks for units in, and the usual |
85d68397 LP |
15395 | overriding semantics between /usr/lib, /etc and /run apply |
15396 | for them too. | |
15397 | ||
15398 | * Most unit file settings which take lists of items can now be | |
6aa8d43a | 15399 | reset by assigning the empty string to them. For example, |
85d68397 LP |
15400 | normally, settings such as Environment=FOO=BAR append a new |
15401 | environment variable assignment to the environment block, | |
15402 | each time they are used. By assigning Environment= the empty | |
15403 | string the environment block can be reset to empty. This is | |
15404 | particularly useful with the .d/*.conf drop-in snippets | |
156f7d09 KS |
15405 | mentioned above, since this adds the ability to reset list |
15406 | settings from vendor unit files via these drop-ins. | |
85d68397 LP |
15407 | |
15408 | * systemctl gained a new "list-dependencies" command for | |
15409 | listing the dependencies of a unit recursively. | |
15410 | ||
40e21da8 | 15411 | * Inhibitors are now honored and listed by "systemctl |
85d68397 LP |
15412 | suspend", "systemctl poweroff" (and similar) too, not only |
15413 | GNOME. These commands will also list active sessions by | |
15414 | other users. | |
15415 | ||
15416 | * Resource limits (as exposed by the various control group | |
15417 | controllers) can now be controlled dynamically at runtime | |
15418 | for all units. More specifically, you can now use a command | |
15419 | like "systemctl set-cgroup-attr foobar.service cpu.shares | |
15420 | 2000" to alter the CPU shares a specific service gets. These | |
6aa8d43a | 15421 | settings are stored persistently on disk, and thus allow the |
85d68397 LP |
15422 | administrator to easily adjust the resource usage of |
15423 | services with a few simple commands. This dynamic resource | |
6aa8d43a | 15424 | management logic is also available to other programs via the |
85d68397 LP |
15425 | bus. Almost any kernel cgroup attribute and controller is |
15426 | supported. | |
15427 | ||
15428 | * systemd-vconsole-setup will now copy all font settings to | |
6aa8d43a LP |
15429 | all allocated VTs, where it previously applied them only to |
15430 | the foreground VT. | |
85d68397 LP |
15431 | |
15432 | * libsystemd-login gained the new sd_session_get_tty() API | |
15433 | call. | |
15434 | ||
6aa8d43a LP |
15435 | * This release drops support for a few legacy or |
15436 | distribution-specific LSB facility names when parsing init | |
15437 | scripts: $x-display-manager, $mail-transfer-agent, | |
85d68397 LP |
15438 | $mail-transport-agent, $mail-transfer-agent, $smtp, |
15439 | $null. Also, the mail-transfer-agent.target unit backing | |
15440 | this has been removed. Distributions which want to retain | |
6aa8d43a LP |
15441 | compatibility with this should carry the burden for |
15442 | supporting this themselves and patch support for these back | |
15443 | in, if they really need to. Also, the facilities $syslog and | |
15444 | $local_fs are now ignored, since systemd does not support | |
15445 | early-boot LSB init scripts anymore, and these facilities | |
15446 | are implied anyway for normal services. syslog.target has | |
15447 | also been removed. | |
85d68397 | 15448 | |
40e21da8 | 15449 | * There are new bus calls on PID1's Manager object for |
6aa8d43a | 15450 | cancelling jobs, and removing snapshot units. Previously, |
85d68397 LP |
15451 | both calls were only available on the Job and Snapshot |
15452 | objects themselves. | |
15453 | ||
15454 | * systemd-journal-gatewayd gained SSL support. | |
15455 | ||
15456 | * The various "environment" files, such as /etc/locale.conf | |
15457 | now support continuation lines with a backslash ("\") as | |
499b604b | 15458 | last character in the line, similarly in style (but different) |
85d68397 LP |
15459 | to how this is supported in shells. |
15460 | ||
15461 | * For normal user processes the _SYSTEMD_USER_UNIT= field is | |
15462 | now implicitly appended to every log entry logged. systemctl | |
15463 | has been updated to filter by this field when operating on a | |
15464 | user systemd instance. | |
15465 | ||
15466 | * nspawn will now implicitly add the CAP_AUDIT_WRITE and | |
15467 | CAP_AUDIT_CONTROL capabilities to the capabilities set for | |
15468 | the container. This makes it easier to boot unmodified | |
15469 | Fedora systems in a container, which however still requires | |
15470 | audit=0 to be passed on the kernel command line. Auditing in | |
15471 | kernel and userspace is unfortunately still too broken in | |
15472 | context of containers, hence we recommend compiling it out | |
15473 | of the kernel or using audit=0. Hopefully this will be fixed | |
15474 | one day for good in the kernel. | |
15475 | ||
15476 | * nspawn gained the new --bind= and --bind-ro= parameters to | |
15477 | bind mount specific directories from the host into the | |
15478 | container. | |
15479 | ||
40e21da8 | 15480 | * nspawn will now mount its own devpts file system instance |
6aa8d43a | 15481 | into the container, in order not to leak pty devices from |
85d68397 LP |
15482 | the host into the container. |
15483 | ||
15484 | * systemd will now read the firmware boot time performance | |
6aa8d43a LP |
15485 | information from the EFI variables, if the used boot loader |
15486 | supports this, and takes it into account for boot performance | |
15487 | analysis via "systemd-analyze". This is currently supported | |
15488 | only in conjunction with Gummiboot, but could be supported | |
15489 | by other boot loaders too. For details see: | |
85d68397 | 15490 | |
a794a4d8 | 15491 | https://systemd.io/BOOT_LOADER_INTERFACE |
85d68397 LP |
15492 | |
15493 | * A new generator has been added that automatically mounts the | |
15494 | EFI System Partition (ESP) to /boot, if that directory | |
6aa8d43a LP |
15495 | exists, is empty, and no other file system has been |
15496 | configured to be mounted there. | |
85d68397 LP |
15497 | |
15498 | * logind will now send out PrepareForSleep(false) out | |
15499 | unconditionally, after coming back from suspend. This may be | |
15500 | used by applications as asynchronous notification for | |
15501 | system resume events. | |
15502 | ||
15503 | * "systemctl unlock-sessions" has been added, that allows | |
15504 | unlocking the screens of all user sessions at once, similar | |
499b604b | 15505 | to how "systemctl lock-sessions" already locked all users |
40e21da8 | 15506 | sessions. This is backed by a new D-Bus call UnlockSessions(). |
85d68397 LP |
15507 | |
15508 | * "loginctl seat-status" will now show the master device of a | |
15509 | seat. (i.e. the device of a seat that needs to be around for | |
15510 | the seat to be considered available, usually the graphics | |
15511 | card). | |
15512 | ||
15513 | * tmpfiles gained a new "X" line type, that allows | |
15514 | configuration of files and directories (with wildcards) that | |
15515 | shall be excluded from automatic cleanup ("aging"). | |
15516 | ||
bf933560 KS |
15517 | * udev default rules set the device node permissions now only |
15518 | at "add" events, and do not change them any longer with a | |
15519 | later "change" event. | |
85d68397 LP |
15520 | |
15521 | * The log messages for lid events and power/sleep keypresses | |
15522 | now carry a message ID. | |
15523 | ||
15524 | * We now have a substantially larger unit test suite, but this | |
15525 | continues to be work in progress. | |
15526 | ||
15527 | * udevadm hwdb gained a new --root= parameter to change the | |
15528 | root directory to operate relative to. | |
15529 | ||
40e21da8 KS |
15530 | * logind will now issue a background sync() request to the kernel |
15531 | early at shutdown, so that dirty buffers are flushed to disk early | |
85d68397 LP |
15532 | instead of at the last moment, in order to optimize shutdown |
15533 | times a little. | |
15534 | ||
15535 | * A new bootctl tool has been added that is an interface for | |
15536 | certain boot loader operations. This is currently a preview | |
15537 | and is likely to be extended into a small mechanism daemon | |
15538 | like timedated, localed, hostnamed, and can be used by | |
15539 | graphical UIs to enumerate available boot options, and | |
15540 | request boot into firmware operations. | |
15541 | ||
15542 | * systemd-bootchart has been relicensed to LGPLv2.1+ to match | |
15543 | the rest of the package. It also has been updated to work | |
15544 | correctly in initrds. | |
15545 | ||
d35f51ea ZJS |
15546 | * polkit previously has been runtime optional, and is now also |
15547 | compile time optional via a configure switch. | |
85d68397 LP |
15548 | |
15549 | * systemd-analyze has been reimplemented in C. Also "systemctl | |
15550 | dot" has moved into systemd-analyze. | |
15551 | ||
15552 | * "systemctl status" with no further parameters will now print | |
15553 | the status of all active or failed units. | |
15554 | ||
15555 | * Operations such as "systemctl start" can now be executed | |
15556 | with a new mode "--irreversible" which may be used to queue | |
15557 | operations that cannot accidentally be reversed by a later | |
6aa8d43a | 15558 | job queuing. This is by default used to make shutdown |
85d68397 LP |
15559 | requests more robust. |
15560 | ||
15561 | * The Python API of systemd now gained a new module for | |
15562 | reading journal files. | |
15563 | ||
15564 | * A new tool kernel-install has been added that can install | |
15565 | kernel images according to the Boot Loader Specification: | |
15566 | ||
a794a4d8 | 15567 | https://systemd.io/BOOT_LOADER_SPECIFICATION |
85d68397 LP |
15568 | |
15569 | * Boot time console output has been improved to provide | |
6aa8d43a | 15570 | animated boot time output for hanging jobs. |
85d68397 LP |
15571 | |
15572 | * A new tool systemd-activate has been added which can be used | |
15573 | to test socket activation with, directly from the command | |
15574 | line. This should make it much easier to test and debug | |
15575 | socket activation in daemons. | |
15576 | ||
15577 | * journalctl gained a new "--reverse" (or -r) option to show | |
15578 | journal output in reverse order (i.e. newest line first). | |
15579 | ||
43447fb7 LP |
15580 | * journalctl gained a new "--pager-end" (or -e) option to jump |
15581 | to immediately jump to the end of the journal in the | |
15582 | pager. This is only supported in conjunction with "less". | |
15583 | ||
85d68397 | 15584 | * journalctl gained a new "--user-unit=" option, that works |
499b604b | 15585 | similarly to "--unit=" but filters for user units rather than |
85d68397 LP |
15586 | system units. |
15587 | ||
15588 | * A number of unit files to ease adoption of systemd in | |
15589 | initrds has been added. This moves some minimal logic from | |
15590 | the various initrd implementations into systemd proper. | |
15591 | ||
15592 | * The journal files are now owned by a new group | |
15593 | "systemd-journal", which exists specifically to allow access | |
15594 | to the journal, and nothing else. Previously, we used the | |
6aa8d43a | 15595 | "adm" group for that, which however possibly covers more |
85d68397 LP |
15596 | than just journal/log file access. This new group is now |
15597 | already used by systemd-journal-gatewayd to ensure this | |
15598 | daemon gets access to the journal files and as little else | |
15599 | as possible. Note that "make install" will also set FS ACLs | |
15600 | up for /var/log/journal to give "adm" and "wheel" read | |
15601 | access to it, in addition to "systemd-journal" which owns | |
15602 | the journal files. We recommend that packaging scripts also | |
6aa8d43a | 15603 | add read access to "adm" + "wheel" to /var/log/journal, and |
85d68397 LP |
15604 | all existing/future journal files. To normal users and |
15605 | administrators little changes, however packagers need to | |
15606 | ensure to create the "systemd-journal" system group at | |
15607 | package installation time. | |
15608 | ||
15609 | * The systemd-journal-gatewayd now runs as unprivileged user | |
15610 | systemd-journal-gateway:systemd-journal-gateway. Packaging | |
15611 | scripts need to create these system user/group at | |
15612 | installation time. | |
15613 | ||
15614 | * timedated now exposes a new boolean property CanNTP that | |
15615 | indicates whether a local NTP service is available or not. | |
15616 | ||
15617 | * systemd-detect-virt will now also detect xen PVs | |
15618 | ||
40e21da8 KS |
15619 | * The pstore file system is now mounted by default, if it is |
15620 | available. | |
85d68397 | 15621 | |
1aed4590 LP |
15622 | * In addition to the SELinux and IMA policies we will now also |
15623 | load SMACK policies at early boot. | |
15624 | ||
85d68397 LP |
15625 | Contributions from: Adel Gadllah, Aleksander Morgado, Auke |
15626 | Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch, | |
15627 | Daniel Wallace, Dave Reisner, David Herrmann, David Strauss, | |
15628 | Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer, | |
15629 | Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering, | |
15630 | Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin | |
15631 | Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael | |
15632 | Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil, | |
15633 | Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor | |
15634 | Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob | |
15635 | Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven | |
15636 | Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom | |
15637 | Gundersen, Umut Tezduyar, William Giokas, Zbigniew | |
15638 | Jędrzejewski-Szmek, Zeeshan Ali (Khattak) | |
15639 | ||
8ad26859 LP |
15640 | CHANGES WITH 197: |
15641 | ||
15642 | * Timer units now support calendar time events in addition to | |
15643 | monotonic time events. That means you can now trigger a unit | |
15644 | based on a calendar time specification such as "Thu,Fri | |
15645 | 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first | |
15646 | or fifth day of any month of the year 2013, given that it is | |
c3fb1e43 | 15647 | a Thursday or a Friday. This brings timer event support |
8ad26859 LP |
15648 | considerably closer to cron's capabilities. For details on |
15649 | the supported calendar time specification language see | |
15650 | systemd.time(7). | |
15651 | ||
15652 | * udev now supports a number of different naming policies for | |
15653 | network interfaces for predictable names, and a combination | |
15654 | of these policies is now the default. Please see this wiki | |
15655 | document for details: | |
15656 | ||
a794a4d8 | 15657 | https://www.freedesktop.org/software/systemd/man/systemd.net-naming-scheme.html |
8ad26859 LP |
15658 | |
15659 | * Auke Kok's bootchart implementation has been added to the | |
d28315e4 JE |
15660 | systemd tree. It is an optional component that can graph the |
15661 | boot in quite some detail. It is one of the best bootchart | |
8ad26859 LP |
15662 | implementations around and minimal in its code and |
15663 | dependencies. | |
15664 | ||
15665 | * nss-myhostname has been integrated into the systemd source | |
15666 | tree. nss-myhostname guarantees that the local hostname | |
15667 | always stays resolvable via NSS. It has been a weak | |
15668 | requirement of systemd-hostnamed since a long time, and | |
15669 | since its code is actually trivial we decided to just | |
15670 | include it in systemd's source tree. It can be turned off | |
15671 | with a configure switch. | |
15672 | ||
15673 | * The read-ahead logic is now capable of properly detecting | |
15674 | whether a btrfs file system is on SSD or rotating media, in | |
15675 | order to optimize the read-ahead scheme. Previously, it was | |
15676 | only capable of detecting this on traditional file systems | |
15677 | such as ext4. | |
15678 | ||
15679 | * In udev, additional device properties are now read from the | |
15680 | IAB in addition to the OUI database. Also, Bluetooth company | |
15681 | identities are attached to the devices as well. | |
15682 | ||
15683 | * In service files %U may be used as specifier that is | |
15684 | replaced by the configured user name of the service. | |
15685 | ||
15686 | * nspawn may now be invoked without a controlling TTY. This | |
15687 | makes it suitable for invocation as its own service. This | |
15688 | may be used to set up a simple containerized server system | |
15689 | using only core OS tools. | |
15690 | ||
15691 | * systemd and nspawn can now accept socket file descriptors | |
15692 | when they are started for socket activation. This enables | |
15693 | implementation of socket activated nspawn | |
15694 | containers. i.e. think about autospawning an entire OS image | |
15695 | when the first SSH or HTTP connection is received. We expect | |
15696 | that similar functionality will also be added to libvirt-lxc | |
15697 | eventually. | |
15698 | ||
15699 | * journalctl will now suppress ANSI color codes when | |
15700 | presenting log data. | |
15701 | ||
15702 | * systemctl will no longer show control group information for | |
ce830873 | 15703 | a unit if the control group is empty anyway. |
8ad26859 LP |
15704 | |
15705 | * logind can now automatically suspend/hibernate/shutdown the | |
15706 | system on idle. | |
15707 | ||
15708 | * /etc/machine-info and hostnamed now also expose the chassis | |
15709 | type of the system. This can be used to determine whether | |
15710 | the local system is a laptop, desktop, handset or | |
15711 | tablet. This information may either be configured by the | |
15712 | user/vendor or is automatically determined from ACPI and DMI | |
15713 | information if possible. | |
15714 | ||
d35f51ea ZJS |
15715 | * A number of polkit actions are now bound together with "imply" |
15716 | rules. This should simplify creating UIs because many actions | |
15717 | will now authenticate similar ones as well. | |
8ad26859 LP |
15718 | |
15719 | * Unit files learnt a new condition ConditionACPower= which | |
15720 | may be used to conditionalize a unit depending on whether an | |
15721 | AC power source is connected or not, of whether the system | |
15722 | is running on battery power. | |
15723 | ||
15724 | * systemctl gained a new "is-failed" verb that may be used in | |
15725 | shell scripts and suchlike to check whether a specific unit | |
15726 | is in the "failed" state. | |
15727 | ||
15728 | * The EnvironmentFile= setting in unit files now supports file | |
15729 | globbing, and can hence be used to easily read a number of | |
15730 | environment files at once. | |
15731 | ||
15732 | * systemd will no longer detect and recognize specific | |
15733 | distributions. All distribution-specific #ifdeffery has been | |
15734 | removed, systemd is now fully generic and | |
15735 | distribution-agnostic. Effectively, not too much is lost as | |
15736 | a lot of the code is still accessible via explicit configure | |
15737 | switches. However, support for some distribution specific | |
15738 | legacy configuration file formats has been dropped. We | |
15739 | recommend distributions to simply adopt the configuration | |
15740 | files everybody else uses now and convert the old | |
15741 | configuration from packaging scripts. Most distributions | |
15742 | already did that. If that's not possible or desirable, | |
15743 | distributions are welcome to forward port the specific | |
15744 | pieces of code locally from the git history. | |
15745 | ||
15746 | * When logging a message about a unit systemd will now always | |
15747 | log the unit name in the message meta data. | |
15748 | ||
15749 | * localectl will now also discover system locale data that is | |
15750 | not stored in locale archives, but directly unpacked. | |
15751 | ||
15752 | * logind will no longer unconditionally use framebuffer | |
15753 | devices as seat masters, i.e. as devices that are required | |
15754 | to be existing before a seat is considered preset. Instead, | |
15755 | it will now look for all devices that are tagged as | |
b938cb90 JE |
15756 | "seat-master" in udev. By default, framebuffer devices will |
15757 | be marked as such, but depending on local systems, other | |
8ad26859 LP |
15758 | devices might be marked as well. This may be used to |
15759 | integrate graphics cards using closed source drivers (such | |
15760 | as NVidia ones) more nicely into logind. Note however, that | |
15761 | we recommend using the open source NVidia drivers instead, | |
15762 | and no udev rules for the closed-source drivers will be | |
15763 | shipped from us upstream. | |
15764 | ||
15765 | Contributions from: Adam Williamson, Alessandro Crismani, Auke | |
15766 | Kok, Colin Walters, Daniel Wallace, Dave Reisner, David | |
15767 | Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra, | |
15768 | Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik | |
15769 | Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart | |
15770 | Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann, | |
15771 | Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry, | |
15772 | Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg | |
15773 | Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar | |
15774 | Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn | |
15775 | Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch, | |
15776 | Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew | |
15777 | Jędrzejewski-Szmek | |
15778 | ||
0428ddb7 LP |
15779 | CHANGES WITH 196: |
15780 | ||
15781 | * udev gained support for loading additional device properties | |
15782 | from an indexed database that is keyed by vendor/product IDs | |
15783 | and similar device identifiers. For the beginning this | |
15784 | "hwdb" is populated with data from the well-known PCI and | |
15785 | USB database, but also includes PNP, ACPI and OID data. In | |
15786 | the longer run this indexed database shall grow into | |
15787 | becoming the one central database for non-essential | |
15788 | userspace device metadata. Previously, data from the PCI/USB | |
96ec33c0 | 15789 | database was only attached to select devices, since the |
0428ddb7 | 15790 | lookup was a relatively expensive operation due to O(n) time |
96ec33c0 LP |
15791 | complexity (with n being the number of entries in the |
15792 | database). Since this is now O(1), we decided to add in this | |
15793 | data for all devices where this is available, by | |
0428ddb7 LP |
15794 | default. Note that the indexed database needs to be rebuilt |
15795 | when new data files are installed. To achieve this you need | |
15796 | to update your packaging scripts to invoke "udevadm hwdb | |
15797 | --update" after installation of hwdb data files. For | |
15798 | RPM-based distributions we introduced the new | |
15799 | %udev_hwdb_update macro for this purpose. | |
15800 | ||
15801 | * The Journal gained support for the "Message Catalog", an | |
15802 | indexed database to link up additional information with | |
15803 | journal entries. For further details please check: | |
15804 | ||
56cadcb6 | 15805 | https://www.freedesktop.org/wiki/Software/systemd/catalog |
0428ddb7 LP |
15806 | |
15807 | The indexed message catalog database also needs to be | |
15808 | rebuilt after installation of message catalog files. Use | |
15809 | "journalctl --update-catalog" for this. For RPM-based | |
15810 | distributions we introduced the %journal_catalog_update | |
15811 | macro for this purpose. | |
15812 | ||
15813 | * The Python Journal bindings gained support for the standard | |
15814 | Python logging framework. | |
15815 | ||
15816 | * The Journal API gained new functions for checking whether | |
15817 | the underlying file system of a journal file is capable of | |
15818 | properly reporting file change notifications, or whether | |
15819 | applications that want to reflect journal changes "live" | |
ab06eef8 | 15820 | need to recheck journal files continuously in appropriate |
0428ddb7 LP |
15821 | time intervals. |
15822 | ||
15823 | * It is now possible to set the "age" field for tmpfiles | |
15824 | entries to 0, indicating that files matching this entry | |
15825 | shall always be removed when the directories are cleaned up. | |
15826 | ||
15827 | * coredumpctl gained a new "gdb" verb which invokes gdb | |
15828 | right-away on the selected coredump. | |
15829 | ||
15830 | * There's now support for "hybrid sleep" on kernels that | |
15831 | support this, in addition to "suspend" and "hibernate". Use | |
15832 | "systemctl hybrid-sleep" to make use of this. | |
15833 | ||
15834 | * logind's HandleSuspendKey= setting (and related settings) | |
15835 | now gained support for a new "lock" setting to simply | |
15836 | request the screen lock on all local sessions, instead of | |
15837 | actually executing a suspend or hibernation. | |
15838 | ||
15839 | * systemd will now mount the EFI variables file system by | |
15840 | default. | |
15841 | ||
15842 | * Socket units now gained support for configuration of the | |
15843 | SMACK security label. | |
15844 | ||
15845 | * timedatectl will now output the time of the last and next | |
15846 | daylight saving change. | |
15847 | ||
15848 | * We dropped support for various legacy and distro-specific | |
15849 | concepts, such as insserv, early-boot SysV services | |
15850 | (i.e. those for non-standard runlevels such as 'b' or 'S') | |
15851 | or ArchLinux /etc/rc.conf support. We recommend the | |
15852 | distributions who still need support this to either continue | |
15853 | to maintain the necessary patches downstream, or find a | |
15854 | different solution. (Talk to us if you have questions!) | |
15855 | ||
d35f51ea ZJS |
15856 | * Various systemd components will now bypass polkit checks for |
15857 | root and otherwise handle properly if polkit is not found to | |
15858 | be around. This should fix most issues for polkit-less | |
15859 | systems. Quite frankly this should have been this way since | |
15860 | day one. It is absolutely our intention to make systemd work | |
15861 | fine on polkit-less systems, and we consider it a bug if | |
15862 | something does not work as it should if polkit is not around. | |
0428ddb7 LP |
15863 | |
15864 | * For embedded systems it is now possible to build udev and | |
15865 | systemd without blkid and/or kmod support. | |
15866 | ||
15867 | * "systemctl switch-root" is now capable of switching root | |
15868 | more than once. I.e. in addition to transitions from the | |
15869 | initrd to the host OS it is now possible to transition to | |
15870 | further OS images from the host. This is useful to implement | |
15871 | offline updating tools. | |
15872 | ||
15873 | * Various other additions have been made to the RPM macros | |
15874 | shipped with systemd. Use %udev_rules_update() after | |
15875 | installing new udev rules files. %_udevhwdbdir, | |
15876 | %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir, | |
15877 | %_sysctldir are now available which resolve to the right | |
15878 | directories for packages to place various data files in. | |
15879 | ||
15880 | * journalctl gained the new --full switch (in addition to | |
15881 | --all, to disable ellipsation for long messages. | |
15882 | ||
15883 | Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel, | |
15884 | Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner, | |
15885 | Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers, | |
15886 | Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas, | |
15887 | Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl, | |
15888 | Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen, | |
15889 | Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas | |
15890 | Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony | |
15891 | Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek | |
15892 | ||
139ee8cc LP |
15893 | CHANGES WITH 195: |
15894 | ||
6827101a | 15895 | * journalctl gained new --since= and --until= switches to |
139ee8cc LP |
15896 | filter by time. It also now supports nice filtering for |
15897 | units via --unit=/-u. | |
15898 | ||
6827101a | 15899 | * Type=oneshot services may use ExecReload= and do the |
139ee8cc LP |
15900 | right thing. |
15901 | ||
15902 | * The journal daemon now supports time-based rotation and | |
15903 | vacuuming, in addition to the usual disk-space based | |
15904 | rotation. | |
15905 | ||
15906 | * The journal will now index the available field values for | |
15907 | each field name. This enables clients to show pretty drop | |
15908 | downs of available match values when filtering. The bash | |
15909 | completion of journalctl has been updated | |
15910 | accordingly. journalctl gained a new switch -F to list all | |
15911 | values a certain field takes in the journal database. | |
15912 | ||
15913 | * More service events are now written as structured messages | |
15914 | to the journal, and made recognizable via message IDs. | |
15915 | ||
15916 | * The timedated, localed and hostnamed mini-services which | |
15917 | previously only provided support for changing time, locale | |
15918 | and hostname settings from graphical DEs such as GNOME now | |
15919 | also have a minimal (but very useful) text-based client | |
15920 | utility each. This is probably the nicest way to changing | |
15921 | these settings from the command line now, especially since | |
15922 | it lists available options and is fully integrated with bash | |
15923 | completion. | |
15924 | ||
15925 | * There's now a new tool "systemd-coredumpctl" to list and | |
15926 | extract coredumps from the journal. | |
15927 | ||
15928 | * We now install a README each in /var/log/ and | |
15929 | /etc/rc.d/init.d explaining where the system logs and init | |
15930 | scripts went. This hopefully should help folks who go to | |
15931 | that dirs and look into the otherwise now empty void and | |
15932 | scratch their heads. | |
15933 | ||
15934 | * When user-services are invoked (by systemd --user) the | |
15935 | $MANAGERPID env var is set to the PID of systemd. | |
15936 | ||
15937 | * SIGRTMIN+24 when sent to a --user instance will now result | |
15938 | in immediate termination of systemd. | |
15939 | ||
15940 | * gatewayd received numerous feature additions such as a | |
15941 | "follow" mode, for live syncing and filtering. | |
15942 | ||
15943 | * browse.html now allows filtering and showing detailed | |
15944 | information on specific entries. Keyboard navigation and | |
15945 | mouse screen support has been added. | |
15946 | ||
15947 | * gatewayd/journalctl now supports HTML5/JSON | |
15948 | Server-Sent-Events as output. | |
15949 | ||
1cb88f2c | 15950 | * The SysV init script compatibility logic will now |
139ee8cc LP |
15951 | heuristically determine whether a script supports the |
15952 | "reload" verb, and only then make this available as | |
15953 | "systemctl reload". | |
15954 | ||
15f47220 | 15955 | * "systemctl status --follow" has been removed, use "journalctl |
139ee8cc LP |
15956 | -u" instead. |
15957 | ||
15958 | * journald.conf's RuntimeMinSize=, PersistentMinSize= settings | |
15959 | have been removed since they are hardly useful to be | |
15960 | configured. | |
15961 | ||
15962 | * And I'd like to take the opportunity to specifically mention | |
15963 | Zbigniew for his great contributions. Zbigniew, you rock! | |
15964 | ||
15965 | Contributions from: Andrew Eikum, Christian Hesse, Colin | |
15966 | Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc | |
4d92e078 LP |
15967 | Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas |
15968 | Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich, | |
15969 | Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas | |
15970 | Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew | |
15971 | Jędrzejewski-Szmek, Сковорода Никита Андреевич | |
139ee8cc | 15972 | |
f9b55720 LP |
15973 | CHANGES WITH 194: |
15974 | ||
15975 | * If /etc/vconsole.conf is non-existent or empty we will no | |
15976 | longer load any console font or key map at boot by | |
15977 | default. Instead the kernel defaults will be left | |
15978 | intact. This is definitely the right thing to do, as no | |
15979 | configuration should mean no configuration, and hard-coding | |
15980 | font names that are different on all archs is probably a bad | |
15981 | idea. Also, the kernel default key map and font should be | |
15982 | good enough for most cases anyway, and mostly identical to | |
15983 | the userspace fonts/key maps we previously overloaded them | |
15984 | with. If distributions want to continue to default to a | |
15985 | non-kernel font or key map they should ship a default | |
15986 | /etc/vconsole.conf with the appropriate contents. | |
15987 | ||
15988 | Contributions from: Colin Walters, Daniel J Walsh, Dave | |
15989 | Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef | |
15990 | Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek | |
15991 | ||
597c52cf LP |
15992 | CHANGES WITH 193: |
15993 | ||
15994 | * journalctl gained a new --cursor= switch to show entries | |
15995 | starting from the specified location in the journal. | |
15996 | ||
15997 | * We now enforce a size limit on journal entry fields exported | |
15998 | with "-o json" in journalctl. Fields larger than 4K will be | |
15999 | assigned null. This can be turned off with --all. | |
16000 | ||
16001 | * An (optional) journal gateway daemon is now available as | |
16002 | "systemd-journal-gatewayd.service". This service provides | |
16003 | access to the journal via HTTP and JSON. This functionality | |
16004 | will be used to implement live log synchronization in both | |
16005 | pull and push modes, but has various other users too, such | |
16006 | as easy log access for debugging of embedded devices. Right | |
16007 | now it is already useful to retrieve the journal via HTTP: | |
16008 | ||
16009 | # systemctl start systemd-journal-gatewayd.service | |
16010 | # wget http://localhost:19531/entries | |
16011 | ||
16012 | This will download the journal contents in a | |
16013 | /var/log/messages compatible format. The same as JSON: | |
16014 | ||
16015 | # curl -H"Accept: application/json" http://localhost:19531/entries | |
16016 | ||
16017 | This service is also accessible via a web browser where a | |
16018 | single static HTML5 app is served that uses the JSON logic | |
16019 | to enable the user to do some basic browsing of the | |
16020 | journal. This will be extended later on. Here's an example | |
16021 | screenshot of this app in its current state: | |
16022 | ||
dc7e580e | 16023 | https://0pointer.de/public/journal-gatewayd |
597c52cf LP |
16024 | |
16025 | Contributions from: Kay Sievers, Lennart Poettering, Robert | |
16026 | Milasan, Tom Gundersen | |
16027 | ||
075d4ecb LP |
16028 | CHANGES WITH 192: |
16029 | ||
16030 | * The bash completion logic is now available for journalctl | |
16031 | too. | |
16032 | ||
d28315e4 | 16033 | * We do not mount the "cpuset" controller anymore together with |
075d4ecb LP |
16034 | "cpu" and "cpuacct", as "cpuset" groups generally cannot be |
16035 | started if no parameters are assigned to it. "cpuset" hence | |
61233823 | 16036 | broke code that assumed it could create "cpu" groups and |
075d4ecb LP |
16037 | just start them. |
16038 | ||
16039 | * journalctl -f will now subscribe to terminal size changes, | |
16040 | and line break accordingly. | |
16041 | ||
597c52cf LP |
16042 | Contributions from: Dave Reisner, Kay Sievers, Lennart |
16043 | Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín | |
075d4ecb | 16044 | |
b6a86739 LP |
16045 | CHANGES WITH 191: |
16046 | ||
16047 | * nspawn will now create a symlink /etc/localtime in the | |
16048 | container environment, copying the host's timezone | |
16049 | setting. Previously this has been done via a bind mount, but | |
16050 | since symlinks cannot be bind mounted this has now been | |
16051 | changed to create/update the appropriate symlink. | |
16052 | ||
16053 | * journalctl -n's line number argument is now optional, and | |
16054 | will default to 10 if omitted. | |
16055 | ||
16056 | * journald will now log the maximum size the journal files may | |
16057 | take up on disk. This is particularly useful if the default | |
16058 | built-in logic of determining this parameter from the file | |
16059 | system size is used. Use "systemctl status | |
6563b535 | 16060 | systemd-journald.service" to see this information. |
b6a86739 LP |
16061 | |
16062 | * The multi-seat X wrapper tool has been stripped down. As X | |
16063 | is now capable of enumerating graphics devices via udev in a | |
16064 | seat-aware way the wrapper is not strictly necessary | |
16065 | anymore. A stripped down temporary stop-gap is still shipped | |
16066 | until the upstream display managers have been updated to | |
16067 | fully support the new X logic. Expect this wrapper to be | |
6563b535 | 16068 | removed entirely in one of the next releases. |
b6a86739 LP |
16069 | |
16070 | * HandleSleepKey= in logind.conf has been split up into | |
16071 | HandleSuspendKey= and HandleHibernateKey=. The old setting | |
6563b535 | 16072 | is not available anymore. X11 and the kernel are |
45afd519 | 16073 | distinguishing between these keys and we should too. This |
b6a86739 LP |
16074 | also means the inhibition lock for these keys has been split |
16075 | into two. | |
16076 | ||
597c52cf LP |
16077 | Contributions from: Dave Airlie, Eelco Dolstra, Lennart |
16078 | Poettering, Lukas Nykryn, Václav Pavlín | |
b6a86739 | 16079 | |
0c11f949 LP |
16080 | CHANGES WITH 190: |
16081 | ||
d28315e4 | 16082 | * Whenever a unit changes state we will now log this to the |
0c11f949 LP |
16083 | journal and show along the unit's own log output in |
16084 | "systemctl status". | |
16085 | ||
16086 | * ConditionPathIsMountPoint= can now properly detect bind | |
16087 | mount points too. (Previously, a bind mount of one file | |
8d0256b7 | 16088 | system to another place in the same file system could not be |
0c11f949 LP |
16089 | detected as mount, since they shared struct stat's st_dev |
16090 | field.) | |
16091 | ||
16092 | * We will now mount the cgroup controllers cpu, cpuacct, | |
16093 | cpuset and the controllers net_cls, net_prio together by | |
16094 | default. | |
16095 | ||
16096 | * nspawn containers will now have a virtualized boot | |
16097 | ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted | |
16098 | over with a randomized ID at container initialization). This | |
16099 | has the effect of making "journalctl -b" do the right thing | |
16100 | in a container. | |
16101 | ||
16102 | * The JSON output journal serialization has been updated not | |
16103 | to generate "endless" list objects anymore, but rather one | |
16104 | JSON object per line. This is more in line how most JSON | |
16105 | parsers expect JSON objects. The new output mode | |
16106 | "json-pretty" has been added to provide similar output, but | |
16107 | neatly aligned for readability by humans. | |
16108 | ||
16109 | * We dropped all explicit sync() invocations in the shutdown | |
16110 | code. The kernel does this implicitly anyway in the kernel | |
16111 | reboot() syscall. halt(8)'s -n option is now a compatibility | |
16112 | no-op. | |
16113 | ||
16114 | * We now support virtualized reboot() in containers, as | |
16115 | supported by newer kernels. We will fall back to exit() if | |
16116 | CAP_SYS_REBOOT is not available to the container. Also, | |
16117 | nspawn makes use of this now and will actually reboot the | |
16118 | container if the containerized OS asks for that. | |
16119 | ||
16120 | * journalctl will only show local log output by default | |
16121 | now. Use --merge (-m) to show remote log output, too. | |
16122 | ||
16123 | * libsystemd-journal gained the new sd_journal_get_usage() | |
16124 | call to determine the current disk usage of all journal | |
16125 | files. This is exposed in the new "journalctl --disk-usage" | |
16126 | command. | |
16127 | ||
16128 | * journald gained a new configuration setting SplitMode= in | |
16129 | journald.conf which may be used to control how user journals | |
16130 | are split off. See journald.conf(5) for details. | |
16131 | ||
16132 | * A new condition type ConditionFileNotEmpty= has been added. | |
16133 | ||
16134 | * tmpfiles' "w" lines now support file globbing, to write | |
16135 | multiple files at once. | |
16136 | ||
16137 | * We added Python bindings for the journal submission | |
16138 | APIs. More Python APIs for a number of selected APIs will | |
16139 | likely follow. Note that we intend to add native bindings | |
16140 | only for the Python language, as we consider it common | |
16141 | enough to deserve bindings shipped within systemd. There are | |
16142 | various projects outside of systemd that provide bindings | |
16143 | for languages such as PHP or Lua. | |
16144 | ||
a98d5d64 LP |
16145 | * Many conditions will now resolve specifiers such as %i. In |
16146 | addition, PathChanged= and related directives of .path units | |
16147 | now support specifiers as well. | |
0c11f949 LP |
16148 | |
16149 | * There's now a new RPM macro definition for the system preset | |
16150 | dir: %_presetdir. | |
16151 | ||
d28315e4 | 16152 | * journald will now warn if it ca not forward a message to the |
dca348bc | 16153 | syslog daemon because its socket is full. |
0c11f949 LP |
16154 | |
16155 | * timedated will no longer write or process /etc/timezone, | |
16156 | except on Debian. As we do not support late mounted /usr | |
16157 | anymore /etc/localtime always being a symlink is now safe, | |
16158 | and hence the information in /etc/timezone is not necessary | |
16159 | anymore. | |
16160 | ||
aaccc32c | 16161 | * logind will now always reserve one VT for a text getty (VT6 |
0c11f949 LP |
16162 | by default). Previously if more than 6 X sessions where |
16163 | started they took up all the VTs with auto-spawned gettys, | |
16164 | so that no text gettys were available anymore. | |
16165 | ||
16166 | * udev will now automatically inform the btrfs kernel logic | |
16167 | about btrfs RAID components showing up. This should make | |
16168 | simple hotplug based btrfs RAID assembly work. | |
16169 | ||
16170 | * PID 1 will now increase its RLIMIT_NOFILE to 64K by default | |
16171 | (but not for its children which will stay at the kernel | |
16172 | default). This should allow setups with a lot more listening | |
16173 | sockets. | |
16174 | ||
16175 | * systemd will now always pass the configured timezone to the | |
16176 | kernel at boot. timedated will do the same when the timezone | |
16177 | is changed. | |
16178 | ||
16179 | * logind's inhibition logic has been updated. By default, | |
16180 | logind will now handle the lid switch, the power and sleep | |
16181 | keys all the time, even in graphical sessions. If DEs want | |
16182 | to handle these events on their own they should take the new | |
16183 | handle-power-key, handle-sleep-key and handle-lid-switch | |
f131770b | 16184 | inhibitors during their runtime. A simple way to achieve |
0c11f949 LP |
16185 | that is to invoke the DE wrapped in an invocation of: |
16186 | ||
1d3a473b | 16187 | systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch … |
0c11f949 LP |
16188 | |
16189 | * Access to unit operations is now checked via SELinux taking | |
16190 | the unit file label and client process label into account. | |
16191 | ||
aad803af LP |
16192 | * systemd will now notify the administrator in the journal |
16193 | when he over-mounts a non-empty directory. | |
16194 | ||
16195 | * There are new specifiers that are resolved in unit files, | |
38b38500 | 16196 | for the hostname (%H), the machine ID (%m) and the boot ID |
aad803af LP |
16197 | (%b). |
16198 | ||
b6a86739 | 16199 | Contributions from: Allin Cottrell, Auke Kok, Brandon Philips, |
0c11f949 LP |
16200 | Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner, |
16201 | Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart | |
16202 | Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas, | |
16203 | Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz, | |
16204 | Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen, | |
16205 | Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek | |
16206 | ||
38a60d71 LP |
16207 | CHANGES WITH 189: |
16208 | ||
16209 | * Support for reading structured kernel messages from | |
16210 | /dev/kmsg has now been added and is enabled by default. | |
16211 | ||
16212 | * Support for reading kernel messages from /proc/kmsg has now | |
16213 | been removed. If you want kernel messages in the journal | |
16214 | make sure to run a recent kernel (>= 3.5) that supports | |
16215 | reading structured messages from /dev/kmsg (see | |
16216 | above). /proc/kmsg is now exclusive property of classic | |
16217 | syslog daemons again. | |
16218 | ||
16219 | * The libudev API gained the new | |
16220 | udev_device_new_from_device_id() call. | |
16221 | ||
16222 | * The logic for file system namespace (ReadOnlyDirectory=, | |
16223 | ReadWriteDirectoy=, PrivateTmp=) has been reworked not to | |
16224 | require pivot_root() anymore. This means fewer temporary | |
16225 | directories are created below /tmp for this feature. | |
16226 | ||
16227 | * nspawn containers will now see and receive all submounts | |
16228 | made on the host OS below the root file system of the | |
16229 | container. | |
16230 | ||
16231 | * Forward Secure Sealing is now supported for Journal files, | |
16232 | which provide cryptographical sealing of journal files so | |
16233 | that attackers cannot alter log history anymore without this | |
16234 | being detectable. Lennart will soon post a blog story about | |
16235 | this explaining it in more detail. | |
16236 | ||
16237 | * There are two new service settings RestartPreventExitStatus= | |
16238 | and SuccessExitStatus= which allow configuration of exit | |
16239 | status (exit code or signal) which will be excepted from the | |
16240 | restart logic, resp. consider successful. | |
16241 | ||
16242 | * journalctl gained the new --verify switch that can be used | |
16243 | to check the integrity of the structure of journal files and | |
16244 | (if Forward Secure Sealing is enabled) the contents of | |
16245 | journal files. | |
16246 | ||
16247 | * nspawn containers will now be run with /dev/stdin, /dev/fd/ | |
16248 | and similar symlinks pre-created. This makes running shells | |
16249 | as container init process a lot more fun. | |
16250 | ||
16251 | * The fstab support can now handle PARTUUID= and PARTLABEL= | |
16252 | entries. | |
16253 | ||
16254 | * A new ConditionHost= condition has been added to match | |
16255 | against the hostname (with globs) and machine ID. This is | |
16256 | useful for clusters where a single OS image is used to | |
16257 | provision a large number of hosts which shall run slightly | |
16258 | different sets of services. | |
16259 | ||
16260 | * Services which hit the restart limit will now be placed in a | |
16261 | failure state. | |
16262 | ||
b6a86739 | 16263 | Contributions from: Bertram Poettering, Dave Reisner, Huang |
38a60d71 LP |
16264 | Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin |
16265 | Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek | |
16266 | ||
c269cec3 LP |
16267 | CHANGES WITH 188: |
16268 | ||
16269 | * When running in --user mode systemd will now become a | |
16270 | subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps | |
16271 | tree a lot more organized. | |
16272 | ||
16273 | * A new PartOf= unit dependency type has been introduced that | |
16274 | may be used to group services in a natural way. | |
16275 | ||
16276 | * "systemctl enable" may now be used to enable instances of | |
16277 | services. | |
16278 | ||
16279 | * journalctl now prints error log levels in red, and | |
16280 | warning/notice log levels in bright white. It also supports | |
16281 | filtering by log level now. | |
16282 | ||
16283 | * cgtop gained a new -n switch (similar to top), to configure | |
16284 | the maximum number of iterations to run for. It also gained | |
16285 | -b, to run in batch mode (accepting no input). | |
16286 | ||
ab06eef8 | 16287 | * The suffix ".service" may now be omitted on most systemctl |
c269cec3 LP |
16288 | command lines involving service unit names. |
16289 | ||
16290 | * There's a new bus call in logind to lock all sessions, as | |
16291 | well as a loginctl verb for it "lock-sessions". | |
16292 | ||
16293 | * libsystemd-logind.so gained a new call sd_journal_perror() | |
16294 | that works similar to libc perror() but logs to the journal | |
16295 | and encodes structured information about the error number. | |
16296 | ||
16297 | * /etc/crypttab entries now understand the new keyfile-size= | |
16298 | option. | |
16299 | ||
16300 | * shutdown(8) now can send a (configurable) wall message when | |
16301 | a shutdown is cancelled. | |
16302 | ||
16303 | * The mount propagation mode for the root file system will now | |
16304 | default to "shared", which is useful to make containers work | |
16305 | nicely out-of-the-box so that they receive new mounts from | |
16306 | the host. This can be undone locally by running "mount | |
16307 | --make-rprivate /" if needed. | |
16308 | ||
16309 | * The prefdm.service file has been removed. Distributions | |
16310 | should maintain this unit downstream if they intend to keep | |
16311 | it around. However, we recommend writing normal unit files | |
16312 | for display managers instead. | |
16313 | ||
16314 | * Since systemd is a crucial part of the OS we will now | |
16315 | default to a number of compiler switches that improve | |
16316 | security (hardening) such as read-only relocations, stack | |
16317 | protection, and suchlike. | |
16318 | ||
16319 | * The TimeoutSec= setting for services is now split into | |
16320 | TimeoutStartSec= and TimeoutStopSec= to allow configuration | |
16321 | of individual time outs for the start and the stop phase of | |
16322 | the service. | |
16323 | ||
16324 | Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke | |
16325 | Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer, | |
16326 | Jim Meyering, Kay Sievers, Lennart Poettering, Mantas | |
16327 | Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter | |
16328 | Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom | |
16329 | Gundersen, Zbigniew Jędrzejewski-Szmek | |
16330 | ||
c4f1b862 LP |
16331 | CHANGES WITH 187: |
16332 | ||
16333 | * The journal and id128 C APIs are now fully documented as man | |
16334 | pages. | |
16335 | ||
16336 | * Extra safety checks have been added when transitioning from | |
16337 | the initial RAM disk to the main system to avoid accidental | |
16338 | data loss. | |
16339 | ||
c269cec3 | 16340 | * /etc/crypttab entries now understand the new keyfile-offset= |
c4f1b862 LP |
16341 | option. |
16342 | ||
16343 | * systemctl -t can now be used to filter by unit load state. | |
16344 | ||
16345 | * The journal C API gained the new sd_journal_wait() call to | |
16346 | make writing synchronous journal clients easier. | |
16347 | ||
16348 | * journalctl gained the new -D switch to show journals from a | |
16349 | specific directory. | |
16350 | ||
16351 | * journalctl now displays a special marker between log | |
16352 | messages of two different boots. | |
16353 | ||
16354 | * The journal is now explicitly flushed to /var via a service | |
16355 | systemd-journal-flush.service, rather than implicitly simply | |
16356 | by seeing /var/log/journal to be writable. | |
16357 | ||
16358 | * journalctl (and the journal C APIs) can now match for much | |
16359 | more complex expressions, with alternatives and | |
16360 | disjunctions. | |
16361 | ||
16362 | * When transitioning from the initial RAM disk to the main | |
16363 | system we will now kill all processes in a killing spree to | |
16364 | ensure no processes stay around by accident. | |
16365 | ||
16366 | * Three new specifiers may be used in unit files: %u, %h, %s | |
16367 | resolve to the user name, user home directory resp. user | |
16368 | shell. This is useful for running systemd user instances. | |
16369 | ||
16370 | * We now automatically rotate journal files if their data | |
16371 | object hash table gets a fill level > 75%. We also size the | |
16372 | hash table based on the configured maximum file size. This | |
16373 | together should lower hash collisions drastically and thus | |
16374 | speed things up a bit. | |
16375 | ||
16376 | * journalctl gained the new "--header" switch to introspect | |
16377 | header data of journal files. | |
16378 | ||
6b000af4 LP |
16379 | * A new setting SystemCallFilters= has been added to services which may |
16380 | be used to apply deny lists or allow lists to system calls. This is | |
16381 | based on SECCOMP Mode 2 of Linux 3.5. | |
c4f1b862 LP |
16382 | |
16383 | * nspawn gained a new --link-journal= switch (and quicker: -j) | |
16384 | to link the container journal with the host. This makes it | |
16385 | very easy to centralize log viewing on the host for all | |
16386 | guests while still keeping the journal files separated. | |
16387 | ||
16388 | * Many bugfixes and optimizations | |
16389 | ||
16390 | Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay | |
16391 | Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex | |
16392 | Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew | |
16393 | Jędrzejewski-Szmek | |
16394 | ||
b5b4c94a LP |
16395 | CHANGES WITH 186: |
16396 | ||
16397 | * Several tools now understand kernel command line arguments, | |
16398 | which are only read when run in an initial RAM disk. They | |
16399 | usually follow closely their normal counterparts, but are | |
16400 | prefixed with rd. | |
16401 | ||
16402 | * There's a new tool to analyze the readahead files that are | |
16403 | automatically generated at boot. Use: | |
16404 | ||
16405 | /usr/lib/systemd/systemd-readahead analyze /.readahead | |
16406 | ||
16407 | * We now provide an early debug shell on tty9 if this enabled. Use: | |
16408 | ||
d1f9edaf | 16409 | systemctl enable debug-shell.service |
b5b4c94a LP |
16410 | |
16411 | * All plymouth related units have been moved into the Plymouth | |
16412 | package. Please make sure to upgrade your Plymouth version | |
16413 | as well. | |
16414 | ||
16415 | * systemd-tmpfiles now supports getting passed the basename of | |
16416 | a configuration file only, in which case it will look for it | |
16417 | in all appropriate directories automatically. | |
16418 | ||
16419 | * udevadm info now takes a /dev or /sys path as argument, and | |
16420 | does the right thing. Example: | |
16421 | ||
16422 | udevadm info /dev/sda | |
16423 | udevadm info /sys/class/block/sda | |
16424 | ||
16425 | * systemctl now prints a warning if a unit is stopped but a | |
16426 | unit that might trigger it continues to run. Example: a | |
16427 | service is stopped but the socket that activates it is left | |
16428 | running. | |
16429 | ||
16430 | * "systemctl status" will now mention if the log output was | |
16431 | shortened due to rotation since a service has been started. | |
16432 | ||
16433 | * The journal API now exposes functions to determine the | |
16434 | "cutoff" times due to rotation. | |
16435 | ||
16436 | * journald now understands SIGUSR1 and SIGUSR2 for triggering | |
16437 | immediately flushing of runtime logs to /var if possible, | |
16438 | resp. for triggering immediate rotation of the journal | |
16439 | files. | |
16440 | ||
16441 | * It is now considered an error if a service is attempted to | |
16442 | be stopped that is not loaded. | |
16443 | ||
16444 | * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames. | |
16445 | ||
16446 | * systemd-analyze now supports Python 3 | |
16447 | ||
16448 | * tmpfiles now supports cleaning up directories via aging | |
16449 | where the first level dirs are always kept around but | |
16450 | directories beneath it automatically aged. This is enabled | |
16451 | by prefixing the age field with '~'. | |
16452 | ||
16453 | * Seat objects now expose CanGraphical, CanTTY properties | |
16454 | which is required to deal with very fast bootups where the | |
16455 | display manager might be running before the graphics drivers | |
16456 | completed initialization. | |
16457 | ||
16458 | * Seat objects now expose a State property. | |
16459 | ||
16460 | * We now include RPM macros for service enabling/disabling | |
16461 | based on the preset logic. We recommend RPM based | |
16462 | distributions to make use of these macros if possible. This | |
16463 | makes it simpler to reuse RPM spec files across | |
16464 | distributions. | |
16465 | ||
16466 | * We now make sure that the collected systemd unit name is | |
16467 | always valid when services log to the journal via | |
16468 | STDOUT/STDERR. | |
16469 | ||
16470 | * There's a new man page kernel-command-line(7) detailing all | |
16471 | command line options we understand. | |
16472 | ||
16473 | * The fstab generator may now be disabled at boot by passing | |
16474 | fstab=0 on the kernel command line. | |
16475 | ||
91ac7425 | 16476 | * A new kernel command line option modules-load= is now understood |
b5b4c94a LP |
16477 | to load a specific kernel module statically, early at boot. |
16478 | ||
16479 | * Unit names specified on the systemctl command line are now | |
16480 | automatically escaped as needed. Also, if file system or | |
16481 | device paths are specified they are automatically turned | |
16482 | into the appropriate mount or device unit names. Example: | |
16483 | ||
16484 | systemctl status /home | |
16485 | systemctl status /dev/sda | |
16486 | ||
16487 | * The SysVConsole= configuration option has been removed from | |
16488 | system.conf parsing. | |
16489 | ||
16490 | * The SysV search path is no longer exported on the D-Bus | |
16491 | Manager object. | |
16492 | ||
ce830873 | 16493 | * The Names= option has been removed from unit file parsing. |
b5b4c94a LP |
16494 | |
16495 | * There's a new man page bootup(7) detailing the boot process. | |
16496 | ||
16497 | * Every unit and every generator we ship with systemd now | |
16498 | comes with full documentation. The self-explanatory boot is | |
16499 | complete. | |
16500 | ||
16501 | * A couple of services gained "systemd-" prefixes in their | |
16502 | name if they wrap systemd code, rather than only external | |
16503 | code. Among them fsck@.service which is now | |
16504 | systemd-fsck@.service. | |
16505 | ||
16506 | * The HaveWatchdog property has been removed from the D-Bus | |
16507 | Manager object. | |
16508 | ||
16509 | * systemd.confirm_spawn= on the kernel command line should now | |
16510 | work sensibly. | |
16511 | ||
16512 | * There's a new man page crypttab(5) which details all options | |
16513 | we actually understand. | |
16514 | ||
16515 | * systemd-nspawn gained a new --capability= switch to pass | |
16516 | additional capabilities to the container. | |
16517 | ||
16518 | * timedated will now read known NTP implementation unit names | |
5b00c016 | 16519 | from /usr/lib/systemd/ntp-units.d/*.list, |
b5b4c94a LP |
16520 | systemd-timedated-ntp.target has been removed. |
16521 | ||
16522 | * journalctl gained a new switch "-b" that lists log data of | |
16523 | the current boot only. | |
16524 | ||
16525 | * The notify socket is in the abstract namespace again, in | |
16526 | order to support daemons which chroot() at start-up. | |
16527 | ||
16528 | * There is a new Storage= configuration option for journald | |
16529 | which allows configuration of where log data should go. This | |
16530 | also provides a way to disable journal logging entirely, so | |
16531 | that data collected is only forwarded to the console, the | |
16532 | kernel log buffer or another syslog implementation. | |
16533 | ||
c4f1b862 | 16534 | * Many bugfixes and optimizations |
b5b4c94a | 16535 | |
2d938ac7 LP |
16536 | Contributions from: Auke Kok, Colin Guthrie, Dave Reisner, |
16537 | David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering, | |
16538 | Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel, | |
16539 | Shawn Landden, Tom Gundersen | |
b5b4c94a | 16540 | |
2d197285 | 16541 | CHANGES WITH 185: |
b6a86739 | 16542 | |
2d197285 KS |
16543 | * "systemctl help <unit>" now shows the man page if one is |
16544 | available. | |
16545 | ||
16546 | * Several new man pages have been added. | |
16547 | ||
b5b4c94a LP |
16548 | * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=, |
16549 | MaxLevelConsole= can now be specified in | |
16550 | journald.conf. These options allow reducing the amount of | |
16551 | data stored on disk or forwarded by the log level. | |
2d197285 | 16552 | |
b5b4c94a LP |
16553 | * TimerSlackNSec= can now be specified in system.conf for |
16554 | PID1. This allows system-wide power savings. | |
2d197285 KS |
16555 | |
16556 | Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen, | |
16557 | Lennart Poettering, Malte Starostik, Marc-Antoine Perennou, | |
16558 | Matthias Clasen | |
16559 | ||
4c8cd173 | 16560 | CHANGES WITH 184: |
b6a86739 | 16561 | |
4c8cd173 LP |
16562 | * logind is now capable of (optionally) handling power and |
16563 | sleep keys as well as the lid switch. | |
16564 | ||
16565 | * journalctl now understands the syntax "journalctl | |
16566 | /usr/bin/avahi-daemon" to get all log output of a specific | |
16567 | daemon. | |
16568 | ||
16569 | * CapabilityBoundingSet= in system.conf now also influences | |
16570 | the capability bound set of usermode helpers of the kernel. | |
16571 | ||
16572 | Contributions from: Daniel Drake, Daniel J. Walsh, Gert | |
16573 | Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers, | |
16574 | Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul | |
16575 | Menzel, Shawn Landden, Tero Roponen, Tom Gundersen | |
16576 | ||
ea5943d3 | 16577 | CHANGES WITH 183: |
b6a86739 | 16578 | |
187076d4 LP |
16579 | * Note that we skipped 139 releases here in order to set the |
16580 | new version to something that is greater than both udev's | |
16581 | and systemd's most recent version number. | |
16582 | ||
194bbe33 KS |
16583 | * udev: all udev sources are merged into the systemd source tree now. |
16584 | All future udev development will happen in the systemd tree. It | |
16585 | is still fully supported to use the udev daemon and tools without | |
16586 | systemd running, like in initramfs or other init systems. Building | |
16587 | udev though, will require the *build* of the systemd tree, but | |
ea5943d3 | 16588 | udev can be properly *run* without systemd. |
07cd4fc1 | 16589 | |
91cf7e5c | 16590 | * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles |
f13b388f KS |
16591 | should be used to create dead device nodes as workarounds for broken |
16592 | subsystems. | |
64661ee7 | 16593 | |
1d3a473b | 16594 | * udev: RUN+="socket:…" and udev_monitor_new_from_socket() is |
2d13da88 KS |
16595 | no longer supported. udev_monitor_new_from_netlink() needs to be |
16596 | used to subscribe to events. | |
16597 | ||
194bbe33 KS |
16598 | * udev: when udevd is started by systemd, processes which are left |
16599 | behind by forking them off of udev rules, are unconditionally cleaned | |
16600 | up and killed now after the event handling has finished. Services or | |
16601 | daemons must be started as systemd services. Services can be | |
ea5943d3 | 16602 | pulled-in by udev to get started, but they can no longer be directly |
194bbe33 KS |
16603 | forked by udev rules. |
16604 | ||
f13b388f KS |
16605 | * udev: the daemon binary is called systemd-udevd now and installed |
16606 | in /usr/lib/systemd/. Standalone builds or non-systemd systems need | |
16607 | to adapt to that, create symlink, or rename the binary after building | |
16608 | it. | |
16609 | ||
ea5943d3 | 16610 | * libudev no longer provides these symbols: |
c1959569 KS |
16611 | udev_monitor_from_socket() |
16612 | udev_queue_get_failed_list_entry() | |
16613 | udev_get_{dev,sys,run}_path() | |
ea5943d3 | 16614 | The versions number was bumped and symbol versioning introduced. |
c1959569 | 16615 | |
ea5943d3 | 16616 | * systemd-loginctl and systemd-journalctl have been renamed |
9ae9afce | 16617 | to loginctl and journalctl to match systemctl. |
18b754d3 KS |
16618 | |
16619 | * The config files: /etc/systemd/systemd-logind.conf and | |
16620 | /etc/systemd/systemd-journald.conf have been renamed to | |
16621 | logind.conf and journald.conf. Package updates should rename | |
16622 | the files to the new names on upgrade. | |
16623 | ||
ea5943d3 LP |
16624 | * For almost all files the license is now LGPL2.1+, changed |
16625 | from the previous GPL2.0+. Exceptions are some minor stuff | |
16626 | of udev (which will be changed to LGPL2.1 eventually, too), | |
16627 | and the MIT licensed sd-daemon.[ch] library that is suitable | |
16628 | to be used as drop-in files. | |
16629 | ||
16630 | * systemd and logind now handle system sleep states, in | |
49f43d5f | 16631 | particular suspending and hibernating. |
ea5943d3 LP |
16632 | |
16633 | * logind now implements a sleep/shutdown/idle inhibiting logic | |
16634 | suitable for a variety of uses. Soonishly Lennart will blog | |
16635 | about this in more detail. | |
16636 | ||
16637 | * var-run.mount and var-lock.mount are no longer provided | |
ce830873 | 16638 | (which previously bind mounted these directories to their new |
ea5943d3 LP |
16639 | places). Distributions which have not converted these |
16640 | directories to symlinks should consider stealing these files | |
16641 | from git history and add them downstream. | |
16642 | ||
16643 | * We introduced the Documentation= field for units and added | |
16644 | this to all our shipped units. This is useful to make it | |
3943231c | 16645 | easier to explore the boot and the purpose of the various |
ea5943d3 LP |
16646 | units. |
16647 | ||
16648 | * All smaller setup units (such as | |
16649 | systemd-vconsole-setup.service) now detect properly if they | |
16650 | are run in a container and are skipped when | |
16651 | appropriate. This guarantees an entirely noise-free boot in | |
16652 | Linux container environments such as systemd-nspawn. | |
16653 | ||
16654 | * A framework for implementing offline system updates is now | |
16655 | integrated, for details see: | |
a794a4d8 | 16656 | https://www.freedesktop.org/software/systemd/man/systemd.offline-updates.html |
ea5943d3 LP |
16657 | |
16658 | * A new service type Type=idle is available now which helps us | |
16659 | avoiding ugly interleaving of getty output and boot status | |
16660 | messages. | |
16661 | ||
439d6dfd LP |
16662 | * There's now a system-wide CapabilityBoundingSet= option to |
16663 | globally reduce the set of capabilities for the | |
ea5943d3 LP |
16664 | system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO, |
16665 | CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or | |
16666 | even CAP_NET_ADMIN system-wide for secure systems. | |
16667 | ||
16668 | * There are now system-wide DefaultLimitXXX= options to | |
16669 | globally change the defaults of the various resource limits | |
16670 | for all units started by PID 1. | |
16671 | ||
16672 | * Harald Hoyer's systemd test suite has been integrated into | |
16673 | systemd which allows easy testing of systemd builds in qemu | |
16674 | and nspawn. (This is really awesome! Ask us for details!) | |
16675 | ||
3943231c LP |
16676 | * The fstab parser is now implemented as generator, not inside |
16677 | of PID 1 anymore. | |
ea5943d3 LP |
16678 | |
16679 | * systemctl will now warn you if .mount units generated from | |
16680 | /etc/fstab are out of date due to changes in fstab that | |
d28315e4 | 16681 | have not been read by systemd yet. |
ea5943d3 LP |
16682 | |
16683 | * systemd is now suitable for usage in initrds. Dracut has | |
16684 | already been updated to make use of this. With this in place | |
16685 | initrds get a slight bit faster but primarily are much | |
16686 | easier to introspect and debug since "systemctl status" in | |
16687 | the host system can be used to introspect initrd services, | |
16688 | and the journal from the initrd is kept around too. | |
16689 | ||
16690 | * systemd-delta has been added, a tool to explore differences | |
16691 | between user/admin configuration and vendor defaults. | |
16692 | ||
16693 | * PrivateTmp= now affects both /tmp and /var/tmp. | |
16694 | ||
16695 | * Boot time status messages are now much prettier and feature | |
16696 | proper english language. Booting up systemd has never been | |
16697 | so sexy. | |
16698 | ||
16699 | * Read-ahead pack files now include the inode number of all | |
16700 | files to pre-cache. When the inode changes the pre-caching | |
16701 | is not attempted. This should be nicer to deal with updated | |
16702 | packages which might result in changes of read-ahead | |
16703 | patterns. | |
16704 | ||
16705 | * We now temporaritly lower the kernel's read_ahead_kb variable | |
16706 | when collecting read-ahead data to ensure the kernel's | |
16707 | built-in read-ahead does not add noise to our measurements | |
16708 | of necessary blocks to pre-cache. | |
16709 | ||
16710 | * There's now RequiresMountsFor= to add automatic dependencies | |
16711 | for all mounts necessary for a specific file system path. | |
16712 | ||
16713 | * MountAuto= and SwapAuto= have been removed from | |
16714 | system.conf. Mounting file systems at boot has to take place | |
16715 | in systemd now. | |
16716 | ||
16717 | * nspawn now learned a new switch --uuid= to set the machine | |
16718 | ID on the command line. | |
16719 | ||
f8c0a2cb | 16720 | * nspawn now learned the -b switch to automatically search |
ea5943d3 LP |
16721 | for an init system. |
16722 | ||
16723 | * vt102 is now the default TERM for serial TTYs, upgraded from | |
16724 | vt100. | |
16725 | ||
16726 | * systemd-logind now works on VT-less systems. | |
16727 | ||
16728 | * The build tree has been reorganized. The individual | |
3943231c | 16729 | components now have directories of their own. |
ea5943d3 LP |
16730 | |
16731 | * A new condition type ConditionPathIsReadWrite= is now available. | |
16732 | ||
16733 | * nspawn learned the new -C switch to create cgroups for the | |
16734 | container in other hierarchies. | |
16735 | ||
16736 | * We now have support for hardware watchdogs, configurable in | |
16737 | system.conf. | |
16738 | ||
16739 | * The scheduled shutdown logic now has a public API. | |
16740 | ||
16741 | * We now mount /tmp as tmpfs by default, but this can be | |
16742 | masked and /etc/fstab can override it. | |
16743 | ||
d28315e4 | 16744 | * Since udisks does not make use of /media anymore we are not |
ea5943d3 LP |
16745 | mounting a tmpfs on it anymore. |
16746 | ||
16747 | * journalctl gained a new --local switch to only interleave | |
16748 | locally generated journal files. | |
16749 | ||
16750 | * We can now load the IMA policy at boot automatically. | |
16751 | ||
16752 | * The GTK tools have been split off into a systemd-ui. | |
16753 | ||
79849bf9 LP |
16754 | Contributions from: Andreas Schwab, Auke Kok, Ayan George, |
16755 | Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan | |
16756 | Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal, | |
16757 | Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers, | |
16758 | Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure, | |
16759 | Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim | |
16760 | A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal | |
16761 | Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn | |
16762 | Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom | |
16763 | Gundersen | |
16764 | ||
16f1239e | 16765 | CHANGES WITH 44: |
b6a86739 | 16766 | |
16f1239e LP |
16767 | * This is mostly a bugfix release |
16768 | ||
16769 | * Support optional initialization of the machine ID from the | |
16770 | KVM or container configured UUID. | |
16771 | ||
16772 | * Support immediate reboots with "systemctl reboot -ff" | |
16773 | ||
16774 | * Show /etc/os-release data in systemd-analyze output | |
16775 | ||
ab06eef8 | 16776 | * Many bugfixes for the journal, including endianness fixes and |
16f1239e LP |
16777 | ensuring that disk space enforcement works |
16778 | ||
ce830873 | 16779 | * sd-login.h is C++ compatible again |
16f1239e LP |
16780 | |
16781 | * Extend the /etc/os-release format on request of the Debian | |
16782 | folks | |
16783 | ||
16784 | * We now refuse non-UTF8 strings used in various configuration | |
d28315e4 | 16785 | and unit files. This is done to ensure we do not pass invalid |
16f1239e LP |
16786 | data over D-Bus or expose it elsewhere. |
16787 | ||
16788 | * Register Mimo USB Screens as suitable for automatic seat | |
16789 | configuration | |
16790 | ||
16791 | * Read SELinux client context from journal clients in a race | |
16792 | free fashion | |
16793 | ||
16794 | * Reorder configuration file lookup order. /etc now always | |
16795 | overrides /run in order to allow the administrator to always | |
b938cb90 | 16796 | and unconditionally override vendor-supplied or |
16f1239e LP |
16797 | automatically generated data. |
16798 | ||
16799 | * The various user visible bits of the journal now have man | |
16800 | pages. We still lack man pages for the journal API calls | |
16801 | however. | |
16802 | ||
16803 | * We now ship all man pages in HTML format again in the | |
16804 | tarball. | |
16805 | ||
16806 | Contributions from: Dave Reisner, Dirk Eibach, Frederic | |
16807 | Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti | |
16808 | Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry | |
16809 | Reding | |
16810 | ||
437b7dee | 16811 | CHANGES WITH 43: |
b6a86739 | 16812 | |
437b7dee LP |
16813 | * This is mostly a bugfix release |
16814 | ||
16815 | * systems lacking /etc/os-release are no longer supported. | |
16816 | ||
16817 | * Various functionality updates to libsystemd-login.so | |
16818 | ||
45afd519 | 16819 | * Track class of PAM logins to distinguish greeters from |
437b7dee LP |
16820 | normal user logins. |
16821 | ||
16822 | Contributions from: Kay Sievers, Lennart Poettering, Michael | |
16823 | Biebl | |
16824 | ||
204fa33c | 16825 | CHANGES WITH 42: |
b6a86739 | 16826 | |
204fa33c LP |
16827 | * This is an important bugfix release for v41. |
16828 | ||
16829 | * Building man pages is now optional which should be useful | |
16830 | for those building systemd from git but unwilling to install | |
16831 | xsltproc. | |
16832 | ||
16833 | * Watchdog support for supervising services is now usable. In | |
16834 | a future release support for hardware watchdogs | |
16835 | (i.e. /dev/watchdog) will be added building on this. | |
16836 | ||
16837 | * Service start rate limiting is now configurable and can be | |
16838 | turned off per service. When a start rate limit is hit a | |
16839 | reboot can automatically be triggered. | |
16840 | ||
16841 | * New CanReboot(), CanPowerOff() bus calls in systemd-logind. | |
16842 | ||
16843 | Contributions from: Benjamin Franzke, Bill Nottingham, | |
16844 | Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal | |
16845 | Schmidt, Michał Górny, Piotr Drąg | |
16846 | ||
e0d25329 | 16847 | CHANGES WITH 41: |
b6a86739 | 16848 | |
e0d25329 KS |
16849 | * The systemd binary is installed /usr/lib/systemd/systemd now; |
16850 | An existing /sbin/init symlink needs to be adapted with the | |
16851 | package update. | |
16852 | ||
b13df964 LP |
16853 | * The code that loads kernel modules has been ported to invoke |
16854 | libkmod directly, instead of modprobe. This means we do not | |
16855 | support systems with module-init-tools anymore. | |
16856 | ||
16857 | * Watchdog support is now already useful, but still not | |
16858 | complete. | |
16859 | ||
16860 | * A new kernel command line option systemd.setenv= is | |
16861 | understood to set system wide environment variables | |
16862 | dynamically at boot. | |
16863 | ||
e9c1ea9d | 16864 | * We now limit the set of capabilities of systemd-journald. |
ccd07a08 | 16865 | |
353e12c2 LP |
16866 | * We now set SIGPIPE to ignore by default, since it only is |
16867 | useful in shell pipelines, and has little use in general | |
16868 | code. This can be disabled with IgnoreSIPIPE=no in unit | |
16869 | files. | |
16870 | ||
b13df964 LP |
16871 | Contributions from: Benjamin Franzke, Kay Sievers, Lennart |
16872 | Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen, | |
16873 | William Douglas | |
16874 | ||
d26e4270 | 16875 | CHANGES WITH 40: |
b6a86739 | 16876 | |
d26e4270 LP |
16877 | * This is mostly a bugfix release |
16878 | ||
16879 | * We now expose the reason why a service failed in the | |
16880 | "Result" D-Bus property. | |
16881 | ||
16882 | * Rudimentary service watchdog support (will be completed over | |
16883 | the next few releases.) | |
16884 | ||
16885 | * When systemd forks off in order execute some service we will | |
16886 | now immediately changes its argv[0] to reflect which process | |
16887 | it will execute. This is useful to minimize the time window | |
16888 | with a generic argv[0], which makes bootcharts more useful | |
16889 | ||
b13df964 LP |
16890 | Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay |
16891 | Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt, | |
16892 | Mike Kazantsev, Ray Strode | |
16893 | ||
220a21d3 | 16894 | CHANGES WITH 39: |
b6a86739 | 16895 | |
220a21d3 LP |
16896 | * This is mostly a test release, but incorporates many |
16897 | bugfixes. | |
16898 | ||
16899 | * New systemd-cgtop tool to show control groups by their | |
16900 | resource usage. | |
16901 | ||
16902 | * Linking against libacl for ACLs is optional again. If | |
16903 | disabled, support tracking device access for active logins | |
16904 | goes becomes unavailable, and so does access to the user | |
16905 | journals by the respective users. | |
16906 | ||
16907 | * If a group "adm" exists, journal files are automatically | |
16908 | owned by them, thus allow members of this group full access | |
16909 | to the system journal as well as all user journals. | |
16910 | ||
16911 | * The journal now stores the SELinux context of the logging | |
16912 | client for all entries. | |
16913 | ||
16914 | * Add C++ inclusion guards to all public headers | |
16915 | ||
16916 | * New output mode "cat" in the journal to print only text | |
16917 | messages, without any meta data like date or time. | |
16918 | ||
16919 | * Include tiny X server wrapper as a temporary stop-gap to | |
16920 | teach XOrg udev display enumeration. This is used by display | |
16921 | managers such as gdm, and will go away as soon as XOrg | |
16922 | learned native udev hotplugging for display devices. | |
16923 | ||
16924 | * Add new systemd-cat tool for executing arbitrary programs | |
16925 | with STDERR/STDOUT connected to the journal. Can also act as | |
16926 | BSD logger replacement, and does so by default. | |
16927 | ||
16928 | * Optionally store all locally generated coredumps in the | |
16929 | journal along with meta data. | |
16930 | ||
16931 | * systemd-tmpfiles learnt four new commands: n, L, c, b, for | |
16932 | writing short strings to files (for usage for /sys), and for | |
16933 | creating symlinks, character and block device nodes. | |
16934 | ||
16935 | * New unit file option ControlGroupPersistent= to make cgroups | |
16936 | persistent, following the mechanisms outlined in | |
56cadcb6 | 16937 | https://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups |
220a21d3 LP |
16938 | |
16939 | * Support multiple local RTCs in a sane way | |
16940 | ||
16941 | * No longer monopolize IO when replaying readahead data on | |
16942 | rotating disks, since we might starve non-file-system IO to | |
16943 | death, since fanotify() will not see accesses done by blkid, | |
16944 | or fsck. | |
16945 | ||
d28315e4 | 16946 | * Do not show kernel threads in systemd-cgls anymore, unless |
220a21d3 LP |
16947 | requested with new -k switch. |
16948 | ||
16949 | Contributions from: Dan Horák, Kay Sievers, Lennart | |
16950 | Poettering, Michal Schmidt | |
16951 | ||
16952 | CHANGES WITH 38: | |
b6a86739 | 16953 | |
220a21d3 LP |
16954 | * This is mostly a test release, but incorporates many |
16955 | bugfixes. | |
16956 | ||
16957 | * The git repository moved to: | |
16958 | git://anongit.freedesktop.org/systemd/systemd | |
16959 | ssh://git.freedesktop.org/git/systemd/systemd | |
16960 | ||
16961 | * First release with the journal | |
dc7e580e | 16962 | https://0pointer.de/blog/projects/the-journal.html |
220a21d3 LP |
16963 | |
16964 | * The journal replaces both systemd-kmsg-syslogd and | |
16965 | systemd-stdout-bridge. | |
16966 | ||
16967 | * New sd_pid_get_unit() API call in libsystemd-logind | |
16968 | ||
16969 | * Many systemadm clean-ups | |
16970 | ||
16971 | * Introduce remote-fs-pre.target which is ordered before all | |
16972 | remote mounts and may be used to start services before all | |
16973 | remote mounts. | |
16974 | ||
16975 | * Added Mageia support | |
16976 | ||
16977 | * Add bash completion for systemd-loginctl | |
16978 | ||
16979 | * Actively monitor PID file creation for daemons which exit in | |
16980 | the parent process before having finished writing the PID | |
16981 | file in the daemon process. Daemons which do this need to be | |
16982 | fixed (i.e. PID file creation must have finished before the | |
16983 | parent exits), but we now react a bit more gracefully to them. | |
16984 | ||
16985 | * Add colourful boot output, mimicking the well-known output | |
16986 | of existing distributions. | |
16987 | ||
16988 | * New option PassCredentials= for socket units, for | |
16989 | compatibility with a recent kernel ABI breakage. | |
16990 | ||
16991 | * /etc/rc.local is now hooked in via a generator binary, and | |
16992 | thus will no longer act as synchronization point during | |
16993 | boot. | |
16994 | ||
16995 | * systemctl list-unit-files now supports --root=. | |
16996 | ||
16997 | * systemd-tmpfiles now understands two new commands: z, Z for | |
16998 | relabelling files according to the SELinux database. This is | |
16999 | useful to apply SELinux labels to specific files in /sys, | |
17000 | among other things. | |
17001 | ||
17002 | * Output of SysV services is now forwarded to both the console | |
17003 | and the journal by default, not only just the console. | |
17004 | ||
17005 | * New man pages for all APIs from libsystemd-login. | |
17006 | ||
ce830873 | 17007 | * The build tree got reorganized and the build system is a |
220a21d3 LP |
17008 | lot more modular allowing embedded setups to specifically |
17009 | select the components of systemd they are interested in. | |
17010 | ||
17011 | * Support for Linux systems lacking the kernel VT subsystem is | |
17012 | restored. | |
17013 | ||
17014 | * configure's --with-rootdir= got renamed to | |
17015 | --with-rootprefix= to follow the naming used by udev and | |
17016 | kmod | |
17017 | ||
d28315e4 | 17018 | * Unless specified otherwise we will now install to /usr instead |
220a21d3 LP |
17019 | of /usr/local by default. |
17020 | ||
17021 | * Processes with '@' in argv[0][0] are now excluded from the | |
17022 | final shut-down killing spree, following the logic explained | |
17023 | in: | |
a794a4d8 | 17024 | https://systemd.io/ROOT_STORAGE_DAEMONS/ |
220a21d3 LP |
17025 | |
17026 | * All processes remaining in a service cgroup when we enter | |
17027 | the START or START_PRE states are now killed with | |
17028 | SIGKILL. That means it is no longer possible to spawn | |
17029 | background processes from ExecStart= lines (which was never | |
17030 | supported anyway, and bad style). | |
17031 | ||
17032 | * New PropagateReloadTo=/PropagateReloadFrom= options to bind | |
17033 | reloading of units together. | |
17034 | ||
4c8cd173 | 17035 | Contributions from: Bill Nottingham, Daniel J. Walsh, Dave |
220a21d3 LP |
17036 | Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay |
17037 | Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt, | |
17038 | Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef | |
17039 | Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek |