3 * Many manager configuration settings that are only applicable to user
4 manager or system manager can be always set. It would be better to reject
5 them when parsing config.
9 * Fedora: add an rpmlint check that verifies that all unit files in the RPM are listed in %systemd_post macros.
11 * wiki: update journal format documentation for lz4 additions
15 * Rearrange tests so that the various test-xyz.c match a specific src/basic/xyz.c again
17 * rework mount.c and swap.c to follow proper state enumeration/deserialization
18 semantics, like we do for device.c now
22 * revisit SystemdOptions EFI variable. Find a better, systematic name and use
23 it for the env var, the bootctl verb and the EFI variable itself, clear up
26 * Figure out a better name for the global per-unit-type drop-in dirs, it's
27 confusing when it comes to -.mount.d/ (is it a drop-in-dir for all mount
28 units, or one for the root mount?). Also, OOM handling in
29 service_unit_name_is_valid() is borked.
33 * socket units: allow creating a udev monitor socket with ListenDevices= or so,
34 with matches, then actviate app thorugh that passing socket oveer
36 * coredump: maybe when coredumping read a new xattr from /proc/$PID/exe that
37 may be used to mark a whole binary as non-coredumpable. Would fix:
38 https://bugs.freedesktop.org/show_bug.cgi?id=69447
40 * teach parse_timestamp() timezones like the calendar spec already knows it
42 * beef up hibernation to optionally do swapon/swapoff immediately before/after
45 * beef up s2h to implement a battery watch loop: instead of entering
46 hibernation unconditionally after coming back from resume make a decision
47 based on the battery load level: if battery level is above a specific
48 threshold, go to suspend again, only hibernate if below it. This means we'd
49 stick to suspend usually, but fall back to hibernation only when battery runs
50 empty (well, subject to our sampling interval). Related to this, check if we
51 can make ACPI _BTP (i.e. /sys/class/power_supply/*/alarm) work for us too,
52 i.e. see if it can wake up machines from suspend, so that we could resume
53 automatically when the system is low on power and move automatically to
54 hibernation mode. (see
55 https://uefi.org/sites/default/files/resources/ACPI%206_2_A_Sept29.pdf
57 https://docs.microsoft.com/en-us/windows-hardware/design/device-experiences/modern-standby-wake-sources
60 * add an explicit "vertical" mode to format-table, so that "systemctl
61 status"-like outputs (i.e. with a series of field names left and values
62 right) become genuine first class citizens, and we gain automatic, sane JSON
65 * dissector: invoke fsck on the file systems we encounter, after all ext4 is
66 still pretty popular (and we mount the ESP too with it after all, which is
69 * systemd-firstboot: teach it dissector magic, so that you can point it to some
70 disk image and it will just set everything in it all behind the scenes.
72 * systemd-firstboot: add --force mode that replaces existing configuration.
74 * We should probably replace /var/log/README, /etc/rc.d/README with symlinks
75 that are linked to these places instead of copied. After all they are
78 * maybe add kernel cmdline params: 1) to force first-boot mode + 2) to force
81 * nspawn: on cgroupsv1 issue cgroup empty handler process based on host events,
82 so that we make cgroup agent logic safe
84 * nspawn/machined: add API to invoke binary in container, then use that as
85 fallback in "machinectl shell"
87 * logind: rework pam_logind to also do a bus call in case of invocation from
88 user@.service, which returns the XDG_RUNTIME_DIR value, and make this
89 behaviour selectable via pam module option.
91 * introduce a new per-process uuid, similar to the boot id, the machine id, the
92 invocation id, that is derived from process creds, specifically a hashed
93 combination of AT_RANDOM + getpid() + the starttime from
94 /proc/self/status. Then add these ids implicitly when logging. Deriving this
95 uuid from these three things has the benefit that it can be derived easily
96 from /proc/$PID/ in a stable, and unique way that changes on both fork() and
99 * let's not GC a unit while its ratelimits are still pending
101 * when killing due to service watchdog timeout maybe detect whether target
102 process is under ptracing and then log loudly and continue instead.
104 * introduce a new group to own TPM devices
106 * make rfkill uaccess controllable by default, i.e. steal rule from
107 gnome-bluetooth and friends
109 * tweak journald context caching. In addition to caching per-process attributes
110 keyed by PID, cache per-cgroup attributes (i.e. the various xattrs we read)
111 keyed by cgroup path, and guarded by ctime changes. This should provide us
112 with a nice speed-up on services that have many processes running in the same
115 * make MAINPID= message reception checks even stricter: if service uses User=,
116 then check sending UID and ignore message if it doesn't match the user or
119 * maybe trigger a uevent "change" on a device if "systemctl reload xyz.device"
122 * when importing an fs tree with machined, optionally apply userns-rec-chown
124 * when importing an fs tree with machined, complain if image is not an OS
126 * Maybe introduce a helper safe_exec() or so, which is to execve() which
127 safe_fork() is to fork(). And then make revert the RLIMIT_NOFILE soft limit
128 to 1K implicitly, unless explicitly opted-out.
130 * rework seccomp/nnp logic that even if User= is used in combination with
131 a seccomp option we don't have to set NNP. For that, change uid first whil
132 keeping CAP_SYS_ADMIN, then apply seccomp, the drop cap.
134 * add a concept for automatically loading per-unit secrets off disk and
135 inserting them into the kernel keyring. Maybe SecretsDirectory= similar to
136 ConfigurationDirectory=.
138 * when no locale is configured, default to UEFI's PlatformLang variable
140 * bootctl,sd-boot: actually honour the "architecture" key
142 * add a new syscall group "@esoteric" for more esoteric stuff such as bpf() and
143 usefaultd() and make systemd-analyze check for it.
145 * paranoia: whenever we process passwords, call mlock() on the memory
146 first. i.e. look for all places we use free_and_erasep() and
147 augment them with mlock(). Also use MADV_DONTDUMP.
149 * Move RestrictAddressFamily= to the new cgroup create socket
151 * support the bind/connect/sendmsg cgroup stuff for sandboxing, and possibly
154 * maybe implicitly attach monotonic+realtime timestamps to outgoing messages in
155 log.c and sd-journal-send
157 * optionally: turn on cgroup delegation for per-session scope units
159 * introduce per-unit (i.e. per-slice, per-service) journal log size limits.
161 * optionally, if a per-partition GPT flag is set for the root/home/… partitions
162 format the partition on next boot and unset the flag, in order to implement
163 factory reset. also, add a second flag that simply indicates whether such a
164 scheme is supported. then, add a tool (or maybe beef up systemd-dissect) to
165 show state of these flags, and optionally trigger such a factory reset on
166 next boot by setting the flag.
168 * sd-boot: automatically load EFI modules from some drop-in dir, so that people
169 can add in file system drivers and such
171 * sd-boot: optionally, show boot menu when previous default boot item has
172 non-zero "tries done" count
174 * maybe set a special xattr on cgroups that have delegate=yes set, to make it
175 easy to mark cut points, then use this information in "systemd-cgls" to show
176 them (e.g. color delegated subtrees in a different color)
178 * introduce an option (or replacement) for "systemctl show" that outputs all
179 properties as JSON, similar to busctl's new JSON output. In contrast to that
180 it should skip the variant type string though.
182 * augment CODE_FILE=, CODE_LINE= with something like CODE_BASE= or so which
183 contains some identifier for the project, which allows us to include
184 clickable links to source files generating these log messages. The identifier
185 could be some abberviated URL prefix or so (taking inspiration from Go
186 imports). For example, for systemd we could use
187 CODE_BASE=github.com/systemd/systemd/blob/98b0b1123cc or so which is
188 sufficient to build a link by prefixing "http://" and suffixing the
191 * Augment MESSAGE_ID with MESSAGE_BASE, in a similar fashion so that we can
192 make clickable links from log messages carrying a MESSAGE_ID, that lead to
193 some explanatory text online.
195 * maybe extend .path units to expose fanotify() per-mount change events
197 * Add a "systemctl list-units --by-slice" mode or so, which rearranges the
198 output of "systemctl list-units" slightly by showing the tree structure of
199 the slices, and the units attached to them.
201 * the a-posteriori stopping of units bound to units that disappeared logic
202 should be reworked: there should be a queue of units, and we should only
203 enqeue stop jobs from a defer event that processes queue instead of
204 right-away when we find a unit that is bound to one that doesn't exist
205 anymore. (similar to how the stop-unneeded queue has been reworked the same
208 * nspawn: make nspawn suitable for shell pipelines: instead of triggering a
209 hangup when input is finished, send ^D, which synthesizes an EOF. Then wait
210 for hangup or ^D before passing on the EOF.
212 * When reloading configuration PID 1 should reset all its properties to the
213 original defaults before calling parse_config()
215 * nspawn: greater control over selinux label?
217 * hibernate/s2h: make this robust and safe to enable in Fedora by default.
220 1. add resume_offset support to the resume code (i.e. support swap files
222 2. check if swap is on weird storage and refuse if so
223 3. add autodetection of hibernation images
225 * cgroups: use inotify to get notified when somebody else modifies cgroups
226 owned by us, then log a friendly warning.
228 * beef up log.c with support for stripping ANSI sequences from strings, so that
229 it is OK to include them in log strings. This would be particularly useful so
230 that our log messages could contain clickable links for example for unit
231 files and suchlike we operate on.
233 * add support for "portablectl attach http://foobar.com/waaa.raw (i.e. importd integration)
235 * add attach --enable and attach --now (for attach+enable+start)
237 * sync dynamic uids/gids between host+portable srvice (i.e. if DynamicUser=1 is set for a service, make sure that the
238 selected user is resolvable in the service even if it ships its own /etc/passwd)
240 * Fix DECIMAL_STR_MAX or DECIMAL_STR_WIDTH. One includes a trailing NUL, the
241 other doesn't. What a disaster. Probably to exclude it. Also
242 DECIMAL_STR_WIDTH should probably add an extra "-" into account for negative
245 * Check that users of inotify's IN_DELETE_SELF flag are using it properly, as
246 usually IN_ATTRIB is the right way to watch deleted files, as the former only
247 fires when a file is actually removed from disk, i.e. the link count drops to
248 zero and is not open anymore, while the latter happens when a file is
249 unlinked from any dir.
251 * port systemctl, busctl, … over to format-table.[ch]'s table formatters
253 * pid1: lock image configured with RootDirectory=/RootImage= using the usual nspawn semantics while the unit is up
255 * add --vacuum-xyz options to coredumpctl, matching those journalctl already has.
257 * introduce Ephemeral= unit file switch, that creates an ephemeral copy of all
258 files and directories that are left writable for a unit, and which are
259 removed after the unit goes down again. A bit like --ephemeral for
260 systemd-nspawn but for system services. If used together with RootImage= this
261 should reflink the image file itself.
263 Related: add Ephemeral=<path1> <path2> … which would allow marking
264 specific paths only like this.
266 * add CopyFile= or so as unit file setting that may be used to copy files or
267 directory trees from the host to the services RootImage= and RootDirectory=
268 environment. Which we can use for /etc/machine-id and in particular
269 /etc/resolv.conf. Should be smart and do something useful on read-only
270 images, for example fallback to read-only bind mounting the file instead.
272 * show invocation ID in systemd-run output
274 * bypass SIGTERM state in unit files if KillSignal is SIGKILL
276 * add proper dbus APIs for the various sd_notify() commands, such as MAINPID=1
277 and so on, which would mean we could report errors and such.
279 * teach tmpfiles.d q/Q logic something sensible in the context of XFS/ext4
282 * introduce DefaultSlice= or so in system.conf that allows changing where we
283 place our units by default, i.e. change system.slice to something
284 else. Similar, ManagerSlice= should exist so that PID1's own scope unit could
285 be moved somewhere else too. Finally machined and logind should get similar
286 options so that it is possible to move user session scopes and machines to a
287 different slice too by default. Usecase: people who want to put resources on
288 the entire system, with the exception of one specific service. See:
289 https://lists.freedesktop.org/archives/systemd-devel/2018-February/040369.html
291 * maybe rework get_user_creds() to query the user database if $SHELL is used
292 for root, but only then.
294 * be stricter with fds we receive for the fdstore: close them asynchronously
296 * calenderspec: add support for week numbers and day numbers within a
297 year. This would allow us to define "bi-weekly" triggers safely.
299 * sd-bus: add vtable flag, that may be used to request client creds implicitly
300 and asynchronously before dispatching the operation
302 * make use of ethtool veth peer info in machined, for automatically finding out
303 host-side interface pointing to the container.
305 * add some special mode to LogsDirectory=/StateDirectory=… that allows
306 declaring these directories without necessarily pulling in deps for them, or
307 creating them when starting up. That way, we could declare that
308 systemd-journald writes to /var/log/journal, which could be useful when we
309 doing disk usage calculations and so on.
311 * taint systemd if there are fewer than 65536 users assigned (userns) to the system.
313 * deprecate RootDirectoryStartOnly= in favour of a new ExecStart= prefix char
315 * add a new RuntimeDirectoryPreserve= mode that defines a similar lifecycle for
316 the runtime dir as we maintain for the fdstore: i.e. keep it around as long
317 as the unit is running or has a job queued.
319 * support projid-based quota in machinectl for containers
321 * maybe use SOURCE_DATE_EPOCH (i.e. the env var the reproducible builds folks
322 introduced) as the RTC epoch, instead of the mtime of NEWS.
324 * add a way to lock down cgroup migration: a boolean, which when set for a unit
325 makes sure the processes in it can never migrate out of it
327 * blog about fd store and restartable services
329 * document Environment=SYSTEMD_LOG_LEVEL=debug drop-in in debugging document
331 * rework ExecOutput and ExecInput enums so that EXEC_OUTPUT_NULL loses its
332 magic meaning and is no longer upgraded to something else if set explicitly.
334 * in the long run: permit a system with /etc/machine-id linked to /dev/null, to
335 make it lose its identity, i.e. be anonymous. For this we'd have to patch
336 through the whole tree to make all code deal with the case where no machine
339 * optionally, collect cgroup resource data, and store it in per-unit RRD files,
340 suitable for processing with rrdtool. Add bus API to access this data, and
341 possibly implement a CPULoad property based on it.
343 * beef up pam_systemd to take unit file settings such as cgroups properties as
346 * maybe hook of xfs/ext4 quotactl() with services? i.e. automatically manage
347 the quota of a the user indicated in User= via unit file settings, like the
348 other resource management concepts. Would mix nicely with DynamicUser=1. Or
349 alternatively, do this with projids, so that we can also cover services
350 running as root. Quota should probably cover all the special dirs such as
351 StateDirectory=, LogsDirectory=, CacheDirectory=, as well as RootDirectory= if it
352 is set, plus the whole disk space any image configured with RootImage=.
354 * In DynamicUser= mode: before selecting a UID, use disk quota APIs on relevant
355 disks to see if the UID is already in use.
357 * add "systemctl wait" or so, which does what "systemd-run --wait" does, but
358 for all units. It should be both a way to pin units into memory as well as a
359 wait to retrieve their exit data.
361 * expose IO accounting data on the bus, show it in systemd-run --wait and log
362 about it in the resource log message
364 * show whether a service has out-of-date configuration in "systemctl status" by
365 using mtime data of ConfigurationDirectory=.
367 * replace all remaining uses of fgets() + LINE_MAX by read_line()
369 * Add AddUser= setting to unit files, similar to DynamicUser=1 which however
370 creates a static, persistent user rather than a dynamic, transient user. We
371 can leverage code from sysusers.d for this.
373 * add some optional flag to ReadWritePaths= and friends, that has the effect
374 that we create the dir in question when the service is started. Example:
376 ReadWritePaths=:/var/lib/foobar
378 * maybe add call sd_journal_set_block_timeout() or so to set SO_SNDTIMEO for
379 the sd-journal logging socket, and, if the timeout is set to 0, sets
380 O_NONBLOCK on it. That way people can control if and when to block for
383 * hostnamed: populate form factor data from a new hwdb database, so that old
384 yogas can be recognized as "convertible" too, even if they predate the DMI
385 "convertible" form factor
387 * Maybe add a small tool invoked early at boot, that adds in or resizes
388 partitions automatically, to be used when the media used is actually larger
389 than the image written onto it is.
391 * Maybe add PrivatePIDs= as new unit setting, and do minimal PID namespacing
392 after all. Be strict however, only support the equivalent of nspawn's
393 --as-pid2 switch, and sanely proxy sd_notify() messages dropping stuff such
396 * Add ExecMonitor= setting. May be used multiple times. Forks off a process in
397 the service cgroup, which is supposed to monitor the service, and when it
398 exits the service is considered failed by its monitor.
400 * track the per-service PAM process properly (i.e. as an additional control
401 process), so that it may be queried on the bus and everything.
403 * add a new "debug" job mode, that is propagated to unit_start() and for
404 services results in two things: we raise SIGSTOP right before invoking
405 execve() and turn off watchdog support. Then, use that to implement
406 "systemd-gdb" for attaching to the start-up of any system service in its
409 * maybe introduce gpt auto discovery for /var/tmp?
411 * maybe add gpt-partition-based user management: each user gets his own
412 LUKS-encrypted GPT partition with a new GPT type. A small nss module
413 enumerates users via udev partition enumeration. UIDs are assigned in a fixed
414 way: the partition index is added as offset to some fixed base uid. User name
415 is stored in GPT partition name. A PAM module authenticates the user via the
416 LUKS partition password. Benefits: strong per-user security, compatibility
417 with stateless/read-only/verity-enabled root. (other idea: do this based on
418 loopback files in /home, without GPT involvement)
420 * gpt-auto logic: introduce support for discovering /var matching an image. For
421 that, use a partition type UUID that is hashed from the OS name (as encoded
422 in /etc/os-release), the architecture, and 4 new bits from the gpt flags
423 field of the root partition. This way can easily support multiple OS
424 installations on the same GPT partition table, without problems with
425 unmatched /var partitions.
427 * gpt-auto logic: related to the above, maybe support a "secondary" root
428 partition, that is mounted to / and is writable, and where the actual root's
429 /usr is mounted into.
431 * gpt-auto logic: support encrypted swap, add kernel cmdline option to force it, and honour a gpt bit about it, plus maybe a configuration file
433 * drop nss-myhostname in favour of nss-resolve?
435 * add a percentage syntax for TimeoutStopSec=, e.g. TimeoutStopSec=150%, and
436 then use that for the setting used in user@.service. It should be understood
437 relative to the configured default value.
439 * in networkd, when matching device types, fix up DEVTYPE rubbish the kernel passes to us
441 * enable LockMLOCK to take a percentage value relative to physical memory
443 * Permit masking specific netlink APIs with RestrictAddressFamily=
445 * nspawn: support that /proc, /sys/, /dev are pre-mounted
447 * define gpt header bits to select volatility mode
449 * ProtectKernelLogs= (drops CAP_SYSLOG, add seccomp for syslog() syscall, and DeviceAllow to /dev/kmsg) in service files
451 * ProtectClock= (drops CAP_SYS_TIMES, adds seecomp filters for settimeofday, adjtimex), sets DeviceAllow o /dev/rtc
453 * ProtectTracing= (drops CAP_SYS_PTRACE, blocks ptrace syscall, makes /sys/kernel/tracing go away)
455 * ProtectMount= (drop mount/umount/pivot_root from seccomp, disallow fuse via DeviceAllow, imply Mountflags=slave)
457 * ProtectKeyRing= to take keyring calls away
459 * RemoveKeyRing= to remove all keyring entries of the specified user
461 * ProtectReboot= that masks reboot() and kexec_load() syscalls, prohibits kill
462 on PID 1 with the relevant signals, and makes relevant files in /sys and
463 /proc (such as the sysrq stuff) unavailable
465 * make sure the ratelimit object can deal with USEC_INFINITY as way to turn off things
467 * journalctl: make sure -f ends when the container indicated by -M terminates
469 * mount: automatically search for "main" partition of an image has multiple
472 * in nss-systemd, if we run inside of RootDirectory= with PrivateUsers= set,
473 find a way to map the User=/Group= of the service to the right name. This way
474 a user/group for a service only has to exist on the host for the right
477 * add bus API for creating unit files in /etc, reusing the code for transient units
479 * add bus API to remove unit files from /etc
481 * add bus API to retrieve current unit file contents (i.e. implement "systemctl cat" on the bus only)
483 * rework fopen_temporary() to make use of open_tmpfile_linkable() (problem: the
484 kernel doesn't support linkat() that replaces existing files, currently)
486 * transient units: don't bother with actually setting unit properties, we
487 reload the unit file anyway
489 * journald: sigbus API via a signal-handler safe function that people may call
490 from the SIGBUS handler
492 * optionally, also require WATCHDOG=1 notifications during service start-up and shutdown
494 * delay activation of logind until somebody logs in, or when /dev/tty0 pulls it
495 in or lingering is on (so that containers don't bother with it until PAM is used). also exit-on-idle
497 * cache sd_event_now() result from before the first iteration...
499 * PID1: find a way how we can reload unit file configuration for
500 specific units only, without reloading the whole of systemd
502 * add an explicit parser for LimitRTPRIO= that verifies
503 the specified range and generates sane error messages for incorrect
506 * when we detect that there are waiting jobs but no running jobs, do something
508 * push CPUAffinity= also into the "cpuset" cgroup controller (only after the cpuset controller got ported to the unified hierarchy)
510 * PID 1 should send out sd_notify("WATCHDOG=1") messages (for usage in the --user mode, and when run via nspawn)
512 * there's probably something wrong with having user mounts below /sys,
513 as we have for debugfs. for example, src/core/mount.c handles mounts
514 prefixed with /sys generally special.
515 http://lists.freedesktop.org/archives/systemd-devel/2015-June/032962.html
517 * fstab-generator: default to tmpfs-as-root if only usr= is specified on the kernel cmdline
519 * docs: bring http://www.freedesktop.org/wiki/Software/systemd/MyServiceCantGetRealtime up to date
521 * add a job mode that will fail if a transaction would mean stopping
522 running units. Use this in timedated to manage the NTP service
524 http://lists.freedesktop.org/archives/systemd-devel/2015-April/030229.html
526 * The udev blkid built-in should expose a property that reflects
527 whether media was sensed in USB CF/SD card readers. This should then
528 be used to control SYSTEMD_READY=1/0 so that USB card readers aren't
529 picked up by systemd unless they contain a medium. This would mirror
530 the behaviour we already have for CD drives.
532 * networkd/udev: implement SR_IOV configuration in .link files:
533 http://lists.freedesktop.org/archives/systemd-devel/2015-January/027451.html
535 * hostnamectl: show root image uuid
537 * Find a solution for SMACK capabilities stuff:
538 http://lists.freedesktop.org/archives/systemd-devel/2014-December/026188.html
540 * "systemctl preset-all" should probably order the unit files it
541 operates on lexicographically before starting to work, in order to
542 ensure deterministic behaviour if two unit files conflict (like DMs
545 * synchronize console access with BSD locks:
546 http://lists.freedesktop.org/archives/systemd-devel/2014-October/024582.html
548 * as soon as we have sender timestamps, revisit coalescing multiple parallel daemon reloads:
549 http://lists.freedesktop.org/archives/systemd-devel/2014-December/025862.html
551 * in systemctl list-unit-files: show the install value the presets would suggest for a service in a third column
553 * figure out when we can use the coarse timers
555 * add "systemctl start -v foobar.service" that shows logs of a service
556 while the start command runs. This is non-trivial to do without
557 races though, since we should flush out all journal messages before
558 returning from the "systemctl stop".
560 * firstboot: make it useful to be run immediately after yum --installroot to set up a machine. (most specifically, make --copy-root-password work even if /etc/passwd already exists
562 * maybe add support for specifier expansion in user.conf, specifically DefaultEnvironment=
564 * maybe allow timer units with an empty Units= setting, so that they
565 can be used for resuming the system but nothing else.
567 * what to do about udev db binary stability for apps? (raw access is not an option)
569 * man: maybe use the word "inspect" rather than "introspect"?
571 * systemctl: if some operation fails, show log output?
573 * systemctl edit: use equivalent of cat() to insert existing config as a comment, prepended with #.
574 Upon editor exit, lines with one # are removed, lines with two # are left with one #, etc.
576 * exponential backoff in timesyncd when we cannot reach a server
578 * timesyncd: add ugly bus calls to set NTP servers per-interface, for usage by NM
580 * merge ~/.local/share and ~/.local/lib into one similar /usr/lib and /usr/share....
582 * systemd.show_status= should probably have a mode where only failed
585 * add systemd.abort_on_kill or some other such flag to send SIGABRT instead of SIGKILL
586 (throughout the codebase, not only PID1)
590 - service registration
591 - service/domain/types browsing
593 - DNS-SD service registration from socket units
594 - resolved should optionally register additional per-interface LLMNR
595 names, so that for the container case we can establish the same name
596 (maybe "host") for referencing the server, everywhere.
597 - allow clients to request DNSSEC for a single lookup even if DNSSEC is off (?)
598 - hook up resolved with machined-based address resolution
600 * refcounting in sd-resolve is borked
602 * Add a new verb "systemctl top"
604 * add new gpt type for btrfs volumes
606 * support empty /etc boots nicely:
607 - nspawn/gpt-generator: introduce new gpt partition type for /usr
609 * generator that automatically discovers btrfs subvolumes, identifies their purpose based on some xattr on them.
611 * a way for container managers to turn off getty starting via $container_headless= or so...
613 * figure out a nice way how we can let the admin know what child/sibling unit causes cgroup membership for a specific unit
615 * For timer units: add some mechanisms so that timer units that trigger immediately on boot do not have the services
616 they run added to the initial transaction and thus confuse Type=idle.
618 * add bus api to query unit file's X fields.
620 * gpt-auto-generator:
621 - Define new partition type for encrypted swap? Support probed LUKS for encrypted swap?
622 - Make /home automount rather than mount?
624 * add generator that pulls in systemd-network from containers when
625 CAP_NET_ADMIN is set, more than the loopback device is defined, even
626 when it is otherwise off
628 * MessageQueueMessageSize= (and suchlike) should use parse_iec_size().
630 * implement Distribute= in socket units to allow running multiple
631 service instances processing the listening socket, and open this up
634 * introduce bus call FreezeUnit(s, b), as well as "systemctl freeze
635 $UNIT" and "systemctl thaw $UNIT" as wrappers around this. The calls
636 should SIGSTOP all unit processes in a loop until all processes of
637 it are fully stopped. This can later be used for app management by
638 desktop UIs such as gnome-shell to freeze apps that are not visible
639 on screen, not unlike how job control works on the shell
642 - implement per-slice CPUFairScheduling=1 switch
643 - introduce high-level settings for RT budget, swappiness
644 - how to reset dynamically changed unit cgroup attributes sanely?
645 - when reloading configuration, apply new cgroup configuration
646 - when recursively showing the cgroup hierarchy, optionally also show
647 the hierarchies of child processes
650 - add field to transient units that indicate whether systemd or somebody else saves/restores its settings, for integration with libvirt
652 * when we detect low battery and no AC on boot, show pretty splash and refuse boot
654 * libsystemd-journal, libsystemd-login, libudev: add calls to easily attach these objects to sd-event event loops
656 * be more careful what we export on the bus as (usec_t) 0 and (usec_t) -1
658 * rfkill,backlight: we probably should run the load tools inside of the udev rules so that the state is properly initialized by the time other software sees it
660 * After coming back from hibernation reset hibernation swap partition using the /dev/snapshot ioctl APIs
662 * If we try to find a unit via a dangling symlink, generate a clean
663 error. Currently, we just ignore it and read the unit from the search
666 * refuse boot if /usr/lib/os-release is missing or /etc/machine-id cannot be set up
668 * man: the documentation of Restart= currently is very misleading and suggests the tools from ExecStartPre= might get restarted.
670 * load .d/*.conf dropins for device units
674 - GetAllProperties() on a non-existing object does not result in a failure currently
675 - port to sd-resolve for connecting to TCP dbus servers
676 - see if we can introduce a new sd_bus_get_owner_machine_id() call to retrieve the machine ID of the machine of the bus itself
677 - see if we can drop more message validation on the sending side
678 - add API to clone sd_bus_message objects
679 - longer term: priority inheritance
681 - NameLost/NameAcquired obsolete
684 - update systemd.special(7) to mention that dbus.socket is only about the compatibility socket now
687 - allow multiple signal handlers per signal?
688 - document chaining of signal handler for SIGCHLD and child handlers
689 - define more intervals where we will shift wakeup intervals around in, 1h, 6h, 24h, ...
690 - generate a failure of a default event loop is executed out-of-thread
692 * investigate endianness issues of UUID vs. GUID
694 * dbus: when a unit failed to load (i.e. is in UNIT_ERROR state), we
695 should be able to safely try another attempt when the bus call LoadUnit() is invoked.
697 * add a pam module that passes the hdd passphrase into the PAM stack and then expires it, for usage by gdm auto-login.
699 * add a pam module that on password changes updates any LUKS slot where the password matches
702 - add unit tests for config_parse_device_allow()
704 * seems that when we follow symlinks to units we prefer the symlink
705 destination path over /etc and /usr. We should not do that. Instead
706 /etc should always override /run+/usr and also any symlink
709 * when isolating, try to figure out a way how we implicitly can order
710 all units we stop before the isolating unit...
712 * teach ConditionKernelCommandLine= globs or regexes (in order to match foobar={no,0,off})
714 * BootLoaderSpec: Clarify that the kernel has to be in $BOOT. Clarify
715 that the boot loader should be installed to the ESP. Define a way
716 how an installer can figure out whether a BLS compliant boot loader
719 * think about requeuing jobs when daemon-reload is issued? usecase:
720 the initrd issues a reload after fstab from the host is accessible
721 and we might want to requeue the mounts local-fs acquired through
724 * systemd-inhibit: make taking delay locks useful: support sending SIGINT or SIGTERM on PrepareForSleep()
726 * remove any syslog support from log.c — we probably cannot do this before split-off udev is gone for good
728 * shutdown logging: store to EFI var, and store to USB stick?
730 * merge unit_kill_common() and unit_kill_context()
733 - honor language efi variables for default language selection (if there are any?)
734 - honor timezone efi variables for default timezone selection (if there are any?)
735 - change bootctl to be backed by systemd-bootd to control temporary and persistent default boot goal plus efi variables
737 - recognize the case when not booted on EFI
739 * maybe do not install getty@tty1.service symlink in /etc but in /usr?
741 * print a nicer explanation if people use variable/specifier expansion in ExecStart= for the first word
743 * mount: turn dependency information from /proc/self/mountinfo into dependency information between systemd units.
746 - logind: optionally, ignore idle-hint logic for autosuspend, block suspend as long as a session is around
747 - logind: wakelock/opportunistic suspend support
748 - Add pretty name for seats in logind
749 - logind: allow showing logout dialog from system?
750 - add Suspend() bus calls which take timestamps to fix double suspend issues when somebody hits suspend and closes laptop quickly.
751 - if pam_systemd is invoked by su from a process that is outside of a
752 any session we should probably just become a NOP, since that's
753 usually not a real user session but just some system code that just
755 - logind: make the Suspend()/Hibernate() bus calls wait for the for
756 the job to be completed. before returning, so that clients can wait
757 for "systemctl suspend" to finish to know when the suspending is
759 - logind: when the power button is pressed short, just popup a
760 logout dialog. If it is pressed for 1s, do the usual
761 shutdown. Inspiration are Macs here.
762 - expose "Locked" property on logind sesison objects
763 - maybe allow configuration of the StopTimeout for session scopes
764 - rename session scope so that it includes the UID. THat way
765 the session scope can be arranged freely in slices and we don't have
766 make assumptions about their slice anymore.
767 - follow PropertiesChanged state more closely, to deal with quick logouts and
771 - consider introducing implicit _TTY= + _PPID= + _EUID= + _EGID= + _FSUID= + _FSGID= fields
772 - journald: also get thread ID from client, plus thread name
773 - journal: when waiting for journal additions in the client always sleep at least 1s or so, in order to minimize wakeups
774 - add API to close/reopen/get fd for journal client fd in libsystemd-journal.
775 - fallback to /dev/log based logging in libsystemd-journal, if we cannot log natively?
776 - declare the local journal protocol stable in the wiki interface chart
777 - sd-journal: speed up sd_journal_get_data() with transparent hash table in bg
778 - journald: when dropping msgs due to ratelimit make sure to write
779 "dropped %u messages" not only when we are about to print the next
780 message that works, but already after a short timeout
781 - check if we can make journalctl by default use --follow mode inside of less if called without args?
782 - maybe add API to send pairs of iovecs via sd_journal_send
783 - journal: add a setgid "systemd-journal" utility to invoke from libsystemd-journal, which passes fds via STDOUT and does PK access
784 - journactl: support negative filtering, i.e. FOOBAR!="waldo",
785 and !FOOBAR for events without FOOBAR.
786 - journal: store timestamp of journal_file_set_offline() int he header,
787 so it is possible to display when the file was last synced.
788 - journal-send.c, log.c: when the log socket is clogged, and we drop, count this and write a message about this when it gets unclogged again.
789 - journal: find a way to allow dropping history early, based on priority, other rules
790 - journal: When used on NFS, check payload hashes
791 - journald: add kernel cmdline option to disable ratelimiting for debug purposes
792 - refuse taking lower-case variable names in sd_journal_send() and friends.
793 - journald: we currently rotate only after MaxUse+MaxFilesize has been reached.
794 - journal: deal nicely with byte-by-byte copied files, especially regards header
795 - journal: sanely deal with entries which are larger than the individual file size, but where the components would fit
796 - Replace utmp, wtmp, btmp, and lastlog completely with journal
797 - journalctl: instead --after-cursor= maybe have a --cursor=XYZ+1 syntax?
798 - when a kernel driver logs in a tight loop, we should ratelimit that too.
799 - journald: optionally, log debug messages to /run but everything else to /var
800 - journald: when we drop syslog messages because the syslog socket is
801 full, make sure to write how many messages are lost as first thing
802 to syslog when it works again.
803 - change systemd-journal-flush into a service that stays around during
804 boot, and causes the journal to be moved back to /run on shutdown,
805 so that we do not keep /var busy. This needs to happen synchronously,
806 hence doing this via signals is not going to work.
807 - optionally support running journald from the command line for testing purposes in external projects
808 - journald: allow per-priority and per-service retention times when rotating/vacuuming
809 - journald: make use of uid-range.h to managed uid ranges to split
811 - journalctl: add the ability to look for the most recent process of a binary. journalctl /usr/bin/X11 --pid=-1 or so...
812 - improve journalctl performance by loading journal files
813 lazily. Encode just enough information in the file name, so that we
814 do not have to open it to know that it is not interesting for us, for
815 the most common operations.
816 - man: document that corrupted journal files is nothing to act on
817 - rework journald sigbus stuff to use mutex
818 - Set RLIMIT_NPROC for systemd-journal-xyz, and all other of our
819 services that run under their own user ids, and use User= (but only
820 in a world where userns is ubiquitous since otherwise we cannot
821 invoke those daemons on the host AND in a container anymore). Also,
822 if LimitNPROC= is used without User= we should warn and refuse
824 - journalctl --verify: don't show files that are currently being
825 written to as FAIL, but instead show that their are being written to.
826 - add journalctl -H that talks via ssh to a remote peer and passes through
828 - add a version of --merge which also merges /var/log/journal/remote
829 - journalctl: -m should access container journals directly by enumerating
830 them via machined, and also watch containers coming and going.
831 Benefit: nspawn --ephemeral would start working nicely with the journal.
832 - assign MESSAGE_ID to log messages about failed services
834 * add a test if all entries in the catalog are properly formatted.
835 (Adding dashes in a catalog entry currently results in the catalog entry
836 being silently skipped. journalctl --update-catalog must warn about this,
837 and we should also have a unit test to check that all our message are OK.)
840 - document that deps in [Unit] sections ignore Alias= fields in
841 [Install] units of other units, unless those units are disabled
842 - man: clarify that time-sync.target is not only sysv compat but also useful otherwise. Same for similar targets
843 - document that service reload may be implemented as service reexec
844 - add a man page containing packaging guidelines and recommending usage of things like Documentation=, PrivateTmp=, PrivateNetwork= and ReadOnlyDirectories=/etc /usr.
845 - document systemd-journal-flush.service properly
846 - documentation: recommend to connect the timer units of a service to the service via Also= in [Install]
847 - man: document the very specific env the shutdown drop-in tools live in
848 - man: add more examples to man pages
849 - man: maybe sort directives in man pages, and take sections from --help and apply them to man too
850 - document root=gpt-auto properly
853 - add systemctl switch to dump transaction without executing it
854 - Add a verbose mode to "systemctl start" and friends that explains what is being done or not done
855 - "systemctl disable" on a static unit prints no message and does
856 nothing. "systemctl enable" does nothing, and gives a bad message
857 about it. Should fix both to print nice actionable messages.
858 - print nice message from systemctl --failed if there are no entries shown, and hook that into ExecStartPre of rescue.service/emergency.service
859 - add new command to systemctl: "systemctl system-reexec" which reexecs as many daemons as virtually possible
860 - systemctl enable: fail if target to alias into does not exist? maybe show how many units are enabled afterwards?
861 - systemctl: "Journal has been rotated since unit was started." message is misleading
862 - systemctl status output should include list of triggering units and their status
865 - "systemctl mask" should find all names by which a unit is accessible
866 (i.e. by scanning for symlinks to it) and link them all to /dev/null
869 - timer units should get the ability to trigger when:
871 - Modulate timer frequency based on battery state
873 * add libsystemd-password or so to query passwords during boot using the password agent logic
875 * clean up date formatting and parsing so that all absolute/relative timestamps we format can also be parsed
877 * on shutdown: move utmp, wall, audit logic all into PID 1 (or logind?), get rid of systemd-update-utmp-runlevel
879 * make repeated alt-ctrl-del presses printing a dump
881 * hostnamed: before returning information from /etc/machine-info.conf check the modification data and reread. Similar for localed, ...
883 * currently x-systemd.timeout is lost in the initrd, since crypttab is copied into dracut, but fstab is not
886 - emulate /dev/kmsg using CUSE and turn off the syslog syscall
887 with seccomp. That should provide us with a useful log buffer that
888 systemd can log to during early boot, and disconnect container logs
889 from the kernel's logs.
890 - as soon as networkd has a bus interface, hook up --network-interface=,
891 --network-bridge= with networkd, to trigger netdev creation should an
893 - a nice way to boot up without machine id set, so that it is set at boot
894 automatically for supporting --ephemeral. Maybe hash the host machine id
895 together with the machine name to generate the machine id for the container
896 - fix logic always print a final newline on output.
897 https://github.com/systemd/systemd/pull/272#issuecomment-113153176
898 - should optionally support receiving WATCHDOG=1 messages from its payload
900 - optionally automatically add FORWARD rules to iptables whenever nspawn is
901 running, remove them when shut down.
904 - refuse mounting over a mount point
905 - automatically discover .roothash files in dissect, similarly to nspawn
908 - add an API so that libvirt-lxc can inform us about network interfaces being
909 removed or added to an existing machine
910 - "machinectl migrate" or similar to copy a container from or to a
911 difference host, via ssh
912 - introduce systemd-nspawn-ephemeral@.service, and hook it into
913 "machinectl start" with a new --ephemeral switch
914 - "machinectl status" should also show internal logs of the container in
916 - "machinectl history"
918 - "machinectl commit" that takes a writable snapshot of a tree, invokes a
919 shell in it, and marks it read-only after use
922 - cryptsetup-generator: allow specification of passwords in crypttab itself
923 - support rd.luks.allow-discards= kernel cmdline params in cryptsetup generator
925 * hw watchdog: optionally try to use the preset watchdog timeout instead of always overriding it
926 https://bugs.freedesktop.org/show_bug.cgi?id=54712
928 * add a dependency on standard-conf.xml and other included files to man pages
930 * MountFlags=shared acts as MountFlags=slave right now.
932 * properly handle loop back mounts via fstab, especially regards to fsck/passno
934 * initialize the hostname from the fs label of /, if /etc/hostname does not exist?
939 - add trigger --subsystem-match=usb/usb_device device
940 - reimport udev db after MOVE events for devices without dev_t
942 * There's currently no way to cancel fsck (used to be possible via C-c or c on the console)
944 * add option to sockets to avoid activation. Instead just drop packets/connections, see http://cyberelk.net/tim/2012/02/15/portreserve-systemd-solution/
947 - save coredump in Windows/Mozilla minidump format
948 - when truncating coredumps, also log the full size that the process had, and make a metadata field so we can report truncated coredumps
950 * support crash reporting operation modes (https://live.gnome.org/GnomeOS/Design/Whiteboards/ProblemReporting)
952 * be able to specify a forced restart of service A where service B depends on, in case B
953 needs to be auto-respawned?
956 - apply "x" on "D" too (see patch from William Douglas)
957 - instead of ignoring unknown fields, reject them.
958 - creating new directories/subvolumes/fifos/device nodes
959 should not follow symlinks. None of the other adjustment or creation
960 calls follow symlinks.
962 * make sure systemd-ask-password-wall does not shutdown systemd-ask-password-console too early
964 * verify that the AF_UNIX sockets of a service in the fs still exist
965 when we start a service in order to avoid confusion when a user
966 assumes starting a service is enough to make it accessible
968 * Make it possible to set the keymap independently from the font on
969 the kernel cmdline. Right now setting one resets also the other.
971 * and a dbus call to generate target from current state
973 * write blog stories about:
974 - hwdb: what belongs into it, lsusb
975 - enabling dbus services
976 - how to make changes to sysctl and sysfs attributes
978 - how to pass throw-away units to systemd, or dynamically change properties of existing units
979 - testing with Harald's awesome test kit
981 - how to develop against journal browsing APIs
982 - the journal HTTP iface
983 - non-cgroup resource management
984 - dynamic resource management with cgroups
985 - refreshed, longer missions statement
986 - calendar time events
987 - init=/bin/sh vs. "emergency" mode, vs. "rescue" mode, vs. "multi-user" mode, vs. "graphical" mode, and the debug shell
988 - how to create your own target
989 - instantiated apache, dovecot and so on
990 - hooking a script into various stages of shutdown/rearly booot
992 * investigate whether the gnome pty helper should be moved into systemd, to provide cgroup support.
994 * dot output for --test showing the 'initial transaction'
997 - When logging about multiple units (stopping BoundTo units, conflicts, etc.),
998 log both units as UNIT=, so that journalctl -u triggers on both.
999 - generate better errors when people try to set transient properties
1000 that are not supported...
1001 http://lists.freedesktop.org/archives/systemd-devel/2015-February/028076.html
1002 - maybe introduce WantsMountsFor=? Usecase:
1003 http://lists.freedesktop.org/archives/systemd-devel/2015-January/027729.html
1004 - recreate systemd's D-Bus private socket file on SIGUSR2
1005 - move PAM code into its own binary
1006 - when we automatically restart a service, ensure we restart its rdeps, too.
1007 - hide PAM options in fragment parser when compile time disabled
1008 - Support --test based on current system state
1009 - If we show an error about a unit (such as not showing up) and it has no Description string, then show a description string generated form the reverse of unit_name_mangle().
1010 - after deserializing sockets in socket.c we should reapply sockopts and things
1011 - drop PID 1 reloading, only do reexecing (difficult: Reload()
1012 currently is properly synchronous, Reexec() is weird, because we
1013 cannot delay the response properly until we are back, so instead of
1014 being properly synchronous we just keep open the fd and close it
1015 when done. That means clients do not get a successful method reply,
1016 but much rather a disconnect on success.
1017 - when breaking cycles drop sysv services first, then services from /run, then from /etc, then from /usr
1018 - when a bus name of a service disappears from the bus make sure to queue further activation requests
1021 - allow port=0 in .socket units
1022 - maybe introduce ExecRestartPre=
1023 - add ReloadSignal= for configuring a reload signal to use
1024 - implement Register= switch in .socket units to enable registration
1025 in Avahi, RPC and other socket registration services.
1026 - allow Type=simple with PIDFile=
1027 https://bugzilla.redhat.com/show_bug.cgi?id=723942
1028 - allow writing multiple conditions in unit files on one line
1029 - load-fragment: when loading a unit file via a chain of symlinks
1030 verify that it is not masked via any of the names traversed.
1031 - introduce Type=pid-file
1032 - introduce mix of BindTo and Requisite
1033 - add a concept of RemainAfterExit= to scope units
1034 - Allow multiple ExecStart= for all Type= settings, so that we can cover rescue.service nicely
1037 - Make sure ID_PATH is always exported and complete for
1038 network devices where possible, so we can safely rely
1042 - add support for more attribute types
1043 - inbuilt piping support (essentially degenerate async)? see loopback-setup.c and other places
1046 - add more keys to [Route] and [Address] sections
1047 - add support for more DHCPv4 options (and, longer term, other kinds of dynamic config)
1048 - add proper initrd support (in particular generate .network/.link files based on /proc/cmdline)
1049 - add reduced [Link] support to .network files
1050 - add Scope= parsing option for [Network]
1051 - properly handle routerless dhcp leases
1052 - work with non-Ethernet devices
1053 - add support for more bond options
1054 - dhcp: do we allow configuring dhcp routes on interfaces that are not the one we got the dhcp info from?
1055 - the DHCP lease data (such as NTP/DNS) is still made available when
1056 a carrier is lost on a link. It should be removed instantly.
1057 - expose in the API the following bits:
1058 - option 15, domain name and/or option 119, search list
1059 - option 12, host name and/or option 81, fqdn
1060 - option 123, 144, geolocation
1061 - option 252, configure http proxy (PAC/wpad)
1062 - provide a way to define a per-network interface default metric value
1063 for all routes to it. possibly a second default for DHCP routes.
1064 - allow Name= to be specified repeatedly in the [Match] section. Maybe also
1065 support Name=foo*|bar*|baz ?
1066 - duplicate address check for static IPs (like ARPCHECK in network-scripts)
1067 - allow DUID/IAID to be customized, see issue #394.
1068 - whenever uplink info changes, make DHCP server send out FORCERENEW
1070 * networkd-wait-online:
1071 - make operstates to wait for configurable?
1074 - figure out how much we can increase Maximum Message Size
1077 - add functions to set previously stored IPv6 addresses on startup and get
1078 them at shutdown; store them in client->ia_na
1079 - write more test cases
1080 - implement reconfigure support, see 5.3., 15.11. and 22.20.
1081 - implement support for temporary adressess (IA_TA)
1082 - implement dhcpv6 authentication
1083 - investigate the usefulness of Confirm messages; i.e. are there any
1084 situations where the link changes without any loss in carrier detection
1086 - some servers don't do rapid commit without a filled in IA_NA, verify
1093 - natively watch for dbus-*.service symlinks (PENDING)
1094 - teach dbus to activate all services it finds in /etc/systemd/services/org-*.service
1096 * fix alsa mixer restore to not print error when no config is stored
1098 * make cryptsetup lower --iter-time
1100 * patch kernel for xattr support in /dev, /proc/, /sys?
1102 * kernel: add device_type = "fb", "fbcon" to class "graphics"
1104 * /usr/bin/service should actually show the new command line
1106 * fedora: suggest auto-restart on failure, but not on success and not on coredump. also, ask people to think about changing the start limit logic. Also point people to RestartPreventExitStatus=, SuccessExitStatus=
1108 * fedora: F20: go timer units all the way, leave cron.daily for cron
1110 * neither pkexec nor sudo initialize environ[] from the PAM environment?
1112 * fedora: update policy to declare access mode and ownership of unit files to root:root 0644, and add an rpmlint check for it
1114 * register catalog database signature as file magic
1116 * zsh shell completion:
1117 - <command> <verb> -<TAB> should complete options, but currently does not
1118 - systemctl add-wants,add-requires
1120 * systemctl status should know about 'systemd-analyze calendar ... --iterations='
1121 * If timer has just OnInactiveSec=..., it should fire after a specified time
1122 after being started.
1126 * look for close() vs. close_nointr() vs. close_nointr_nofail()
1128 * check for strerror(r) instead of strerror(-r)
1132 * set_put(), hashmap_put() return values check. i.e. == 0 does not free()!
1134 * use secure_getenv() instead of getenv() where appropriate
1136 * link up selected blog stories from man pages and unit files Documentation= fields
1137 String is not UTF-8 clean, ignoring assignment
1138 timedatex.service: Consumed 26ms CPU time.