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
28 * socket units: allow creating a udev monitor socket with ListenDevices= or so,
29 with matches, then actviate app thorugh that passing socket oveer
31 * move discoverable partitions spec into markdown and our tree
33 * sd-event: add native support for P_ALL waitid() watching, then move PID 1 to
34 it fo reaping assigned but unknown children. This needs to some special care
35 to operate somewhat sensibly in light of priorities: P_ALL will return
36 arbitrary processes, regardless of the priority we want to watch them with,
37 hence on each event loop iteration check all processes which we shall watch
38 with higher prio explicitly, and then watch the entire rest with P_ALL.
40 * tweak sd-event's child watching: keep a prioq of children to watch and use
41 waitid() only on the children with the highest priority until one is waitable
42 and ignore all lower-prio ones from that point on
44 * sd-event: drop stack allocated epoll_event buffer in sd_event_wait()
46 * maybe introduce xattrs that can be set on the root dir of the root fs
47 partition that declare the volatility mode to use the image in. Previously I
48 thought marking this via GPT partition flags but that's not ideal since
49 that's outside of the LUKS encryption/verity verification, and we probably
50 shouldn't operate in a volatile mode unless we got told so from a trusted
53 * look for /var/tmp automatically via gpt auto discovery
55 * figure out automatic partition discovery when combining writable root dir
58 * coredump: maybe when coredumping read a new xattr from /proc/$PID/exe that
59 may be used to mark a whole binary as non-coredumpable. Would fix:
60 https://bugs.freedesktop.org/show_bug.cgi?id=69447
62 * teach parse_timestamp() timezones like the calendar spec already knows it
64 * beef up hibernation to optionally do swapon/swapoff immediately before/after
67 * beef up s2h to implement a battery watch loop: instead of entering
68 hibernation unconditionally after coming back from resume make a decision
69 based on the battery load level: if battery level is above a specific
70 threshold, go to suspend again, only hibernate if below it. This means we'd
71 stick to suspend usually, but fall back to hibernation only when battery runs
72 empty (well, subject to our sampling interval). Related to this, check if we
73 can make ACPI _BTP (i.e. /sys/class/power_supply/*/alarm) work for us too,
74 i.e. see if it can wake up machines from suspend, so that we could resume
75 automatically when the system is low on power and move automatically to
76 hibernation mode. (see
77 https://uefi.org/sites/default/files/resources/ACPI%206_2_A_Sept29.pdf
79 https://docs.microsoft.com/en-us/windows-hardware/design/device-experiences/modern-standby-wake-sources
82 * add an explicit "vertical" mode to format-table, so that "systemctl
83 status"-like outputs (i.e. with a series of field names left and values
84 right) become genuine first class citizens, and we gain automatic, sane JSON
87 * dissector: invoke fsck on the file systems we encounter, after all ext4 is
88 still pretty popular (and we mount the ESP too with it after all, which is
91 * systemd-firstboot: teach it dissector magic, so that you can point it to some
92 disk image and it will just set everything in it all behind the scenes.
94 * systemd-firstboot: add --force mode that replaces existing configuration.
96 * We should probably replace /var/log/README, /etc/rc.d/README with symlinks
97 that are linked to these places instead of copied. After all they are
100 * maybe add kernel cmdline params: 1) to force first-boot mode + 2) to force
101 random seed crediting
103 * nspawn: on cgroupsv1 issue cgroup empty handler process based on host events,
104 so that we make cgroup agent logic safe
106 * nspawn/machined: add API to invoke binary in container, then use that as
107 fallback in "machinectl shell"
109 * logind: rework pam_logind to also do a bus call in case of invocation from
110 user@.service, which returns the XDG_RUNTIME_DIR value, and make this
111 behaviour selectable via pam module option.
113 * introduce a new per-process uuid, similar to the boot id, the machine id, the
114 invocation id, that is derived from process creds, specifically a hashed
115 combination of AT_RANDOM + getpid() + the starttime from
116 /proc/self/status. Then add these ids implicitly when logging. Deriving this
117 uuid from these three things has the benefit that it can be derived easily
118 from /proc/$PID/ in a stable, and unique way that changes on both fork() and
121 * let's not GC a unit while its ratelimits are still pending
123 * when killing due to service watchdog timeout maybe detect whether target
124 process is under ptracing and then log loudly and continue instead.
126 * introduce a new group to own TPM devices
128 * make rfkill uaccess controllable by default, i.e. steal rule from
129 gnome-bluetooth and friends
131 * tweak journald context caching. In addition to caching per-process attributes
132 keyed by PID, cache per-cgroup attributes (i.e. the various xattrs we read)
133 keyed by cgroup path, and guarded by ctime changes. This should provide us
134 with a nice speed-up on services that have many processes running in the same
137 * make MAINPID= message reception checks even stricter: if service uses User=,
138 then check sending UID and ignore message if it doesn't match the user or
141 * maybe trigger a uevent "change" on a device if "systemctl reload xyz.device"
144 * when importing an fs tree with machined, optionally apply userns-rec-chown
146 * when importing an fs tree with machined, complain if image is not an OS
148 * Maybe introduce a helper safe_exec() or so, which is to execve() which
149 safe_fork() is to fork(). And then make revert the RLIMIT_NOFILE soft limit
150 to 1K implicitly, unless explicitly opted-out.
152 * rework seccomp/nnp logic that even if User= is used in combination with
153 a seccomp option we don't have to set NNP. For that, change uid first whil
154 keeping CAP_SYS_ADMIN, then apply seccomp, the drop cap.
156 * add a concept for automatically loading per-unit secrets off disk and
157 inserting them into the kernel keyring. Maybe SecretsDirectory= similar to
158 ConfigurationDirectory=.
160 * when no locale is configured, default to UEFI's PlatformLang variable
162 * bootctl,sd-boot: actually honour the "architecture" key
164 * add a new syscall group "@esoteric" for more esoteric stuff such as bpf() and
165 usefaultd() and make systemd-analyze check for it.
167 * paranoia: whenever we process passwords, call mlock() on the memory
168 first. i.e. look for all places we use free_and_erasep() and
169 augment them with mlock(). Also use MADV_DONTDUMP.
171 * Move RestrictAddressFamily= to the new cgroup create socket
173 * support the bind/connect/sendmsg cgroup stuff for sandboxing, and possibly
176 * maybe implicitly attach monotonic+realtime timestamps to outgoing messages in
177 log.c and sd-journal-send
179 * optionally: turn on cgroup delegation for per-session scope units
181 * introduce per-unit (i.e. per-slice, per-service) journal log size limits.
183 * optionally, if a per-partition GPT flag is set for the root/home/… partitions
184 format the partition on next boot and unset the flag, in order to implement
185 factory reset. also, add a second flag that simply indicates whether such a
186 scheme is supported. then, add a tool (or maybe beef up systemd-dissect) to
187 show state of these flags, and optionally trigger such a factory reset on
188 next boot by setting the flag.
190 * sd-boot: automatically load EFI modules from some drop-in dir, so that people
191 can add in file system drivers and such
193 * sd-boot: optionally, show boot menu when previous default boot item has
194 non-zero "tries done" count
196 * introduce an option (or replacement) for "systemctl show" that outputs all
197 properties as JSON, similar to busctl's new JSON output. In contrast to that
198 it should skip the variant type string though.
200 * augment CODE_FILE=, CODE_LINE= with something like CODE_BASE= or so which
201 contains some identifier for the project, which allows us to include
202 clickable links to source files generating these log messages. The identifier
203 could be some abberviated URL prefix or so (taking inspiration from Go
204 imports). For example, for systemd we could use
205 CODE_BASE=github.com/systemd/systemd/blob/98b0b1123cc or so which is
206 sufficient to build a link by prefixing "http://" and suffixing the
209 * Augment MESSAGE_ID with MESSAGE_BASE, in a similar fashion so that we can
210 make clickable links from log messages carrying a MESSAGE_ID, that lead to
211 some explanatory text online.
213 * maybe extend .path units to expose fanotify() per-mount change events
215 * Add a "systemctl list-units --by-slice" mode or so, which rearranges the
216 output of "systemctl list-units" slightly by showing the tree structure of
217 the slices, and the units attached to them.
219 * the a-posteriori stopping of units bound to units that disappeared logic
220 should be reworked: there should be a queue of units, and we should only
221 enqeue stop jobs from a defer event that processes queue instead of
222 right-away when we find a unit that is bound to one that doesn't exist
223 anymore. (similar to how the stop-unneeded queue has been reworked the same
226 * nspawn: make nspawn suitable for shell pipelines: instead of triggering a
227 hangup when input is finished, send ^D, which synthesizes an EOF. Then wait
228 for hangup or ^D before passing on the EOF.
230 * When reloading configuration PID 1 should reset all its properties to the
231 original defaults before calling parse_config()
233 * nspawn: greater control over selinux label?
235 * hibernate/s2h: make this robust and safe to enable in Fedora by default.
238 1. add resume_offset support to the resume code (i.e. support swap files
240 2. check if swap is on weird storage and refuse if so
241 3. add autodetection of hibernation images
243 * cgroups: use inotify to get notified when somebody else modifies cgroups
244 owned by us, then log a friendly warning.
246 * beef up log.c with support for stripping ANSI sequences from strings, so that
247 it is OK to include them in log strings. This would be particularly useful so
248 that our log messages could contain clickable links for example for unit
249 files and suchlike we operate on.
251 * add support for "portablectl attach http://foobar.com/waaa.raw (i.e. importd integration)
253 * add attach --enable and attach --now (for attach+enable+start)
255 * sync dynamic uids/gids between host+portable srvice (i.e. if DynamicUser=1 is set for a service, make sure that the
256 selected user is resolvable in the service even if it ships its own /etc/passwd)
258 * Fix DECIMAL_STR_MAX or DECIMAL_STR_WIDTH. One includes a trailing NUL, the
259 other doesn't. What a disaster. Probably to exclude it. Also
260 DECIMAL_STR_WIDTH should probably add an extra "-" into account for negative
263 * Check that users of inotify's IN_DELETE_SELF flag are using it properly, as
264 usually IN_ATTRIB is the right way to watch deleted files, as the former only
265 fires when a file is actually removed from disk, i.e. the link count drops to
266 zero and is not open anymore, while the latter happens when a file is
267 unlinked from any dir.
269 * port systemctl, busctl, … over to format-table.[ch]'s table formatters
271 * pid1: lock image configured with RootDirectory=/RootImage= using the usual nspawn semantics while the unit is up
273 * add --vacuum-xyz options to coredumpctl, matching those journalctl already has.
275 * introduce Ephemeral= unit file switch, that creates an ephemeral copy of all
276 files and directories that are left writable for a unit, and which are
277 removed after the unit goes down again. A bit like --ephemeral for
278 systemd-nspawn but for system services. If used together with RootImage= this
279 should reflink the image file itself.
281 Related: add Ephemeral=<path1> <path2> … which would allow marking
282 specific paths only like this.
284 * add CopyFile= or so as unit file setting that may be used to copy files or
285 directory trees from the host to the services RootImage= and RootDirectory=
286 environment. Which we can use for /etc/machine-id and in particular
287 /etc/resolv.conf. Should be smart and do something useful on read-only
288 images, for example fallback to read-only bind mounting the file instead.
290 * show invocation ID in systemd-run output
292 * bypass SIGTERM state in unit files if KillSignal is SIGKILL
294 * add proper dbus APIs for the various sd_notify() commands, such as MAINPID=1
295 and so on, which would mean we could report errors and such.
297 * teach tmpfiles.d q/Q logic something sensible in the context of XFS/ext4
300 * introduce DefaultSlice= or so in system.conf that allows changing where we
301 place our units by default, i.e. change system.slice to something
302 else. Similar, ManagerSlice= should exist so that PID1's own scope unit could
303 be moved somewhere else too. Finally machined and logind should get similar
304 options so that it is possible to move user session scopes and machines to a
305 different slice too by default. Usecase: people who want to put resources on
306 the entire system, with the exception of one specific service. See:
307 https://lists.freedesktop.org/archives/systemd-devel/2018-February/040369.html
309 * maybe rework get_user_creds() to query the user database if $SHELL is used
310 for root, but only then.
312 * be stricter with fds we receive for the fdstore: close them asynchronously
314 * calenderspec: add support for week numbers and day numbers within a
315 year. This would allow us to define "bi-weekly" triggers safely.
317 * sd-bus: add vtable flag, that may be used to request client creds implicitly
318 and asynchronously before dispatching the operation
320 * make use of ethtool veth peer info in machined, for automatically finding out
321 host-side interface pointing to the container.
323 * add some special mode to LogsDirectory=/StateDirectory=… that allows
324 declaring these directories without necessarily pulling in deps for them, or
325 creating them when starting up. That way, we could declare that
326 systemd-journald writes to /var/log/journal, which could be useful when we
327 doing disk usage calculations and so on.
329 * taint systemd if there are fewer than 65536 users assigned (userns) to the system.
331 * deprecate RootDirectoryStartOnly= in favour of a new ExecStart= prefix char
333 * add a new RuntimeDirectoryPreserve= mode that defines a similar lifecycle for
334 the runtime dir as we maintain for the fdstore: i.e. keep it around as long
335 as the unit is running or has a job queued.
337 * support projid-based quota in machinectl for containers
339 * maybe use SOURCE_DATE_EPOCH (i.e. the env var the reproducible builds folks
340 introduced) as the RTC epoch, instead of the mtime of NEWS.
342 * add a way to lock down cgroup migration: a boolean, which when set for a unit
343 makes sure the processes in it can never migrate out of it
345 * blog about fd store and restartable services
347 * document Environment=SYSTEMD_LOG_LEVEL=debug drop-in in debugging document
349 * rework ExecOutput and ExecInput enums so that EXEC_OUTPUT_NULL loses its
350 magic meaning and is no longer upgraded to something else if set explicitly.
352 * in the long run: permit a system with /etc/machine-id linked to /dev/null, to
353 make it lose its identity, i.e. be anonymous. For this we'd have to patch
354 through the whole tree to make all code deal with the case where no machine
357 * optionally, collect cgroup resource data, and store it in per-unit RRD files,
358 suitable for processing with rrdtool. Add bus API to access this data, and
359 possibly implement a CPULoad property based on it.
361 * beef up pam_systemd to take unit file settings such as cgroups properties as
364 * maybe hook of xfs/ext4 quotactl() with services? i.e. automatically manage
365 the quota of a the user indicated in User= via unit file settings, like the
366 other resource management concepts. Would mix nicely with DynamicUser=1. Or
367 alternatively, do this with projids, so that we can also cover services
368 running as root. Quota should probably cover all the special dirs such as
369 StateDirectory=, LogsDirectory=, CacheDirectory=, as well as RootDirectory= if it
370 is set, plus the whole disk space any image configured with RootImage=.
372 * In DynamicUser= mode: before selecting a UID, use disk quota APIs on relevant
373 disks to see if the UID is already in use.
375 * add "systemctl wait" or so, which does what "systemd-run --wait" does, but
376 for all units. It should be both a way to pin units into memory as well as a
377 wait to retrieve their exit data.
379 * expose IO accounting data on the bus, show it in systemd-run --wait and log
380 about it in the resource log message
382 * show whether a service has out-of-date configuration in "systemctl status" by
383 using mtime data of ConfigurationDirectory=.
385 * replace all remaining uses of fgets() + LINE_MAX by read_line()
387 * Add AddUser= setting to unit files, similar to DynamicUser=1 which however
388 creates a static, persistent user rather than a dynamic, transient user. We
389 can leverage code from sysusers.d for this.
391 * add some optional flag to ReadWritePaths= and friends, that has the effect
392 that we create the dir in question when the service is started. Example:
394 ReadWritePaths=:/var/lib/foobar
396 * maybe add call sd_journal_set_block_timeout() or so to set SO_SNDTIMEO for
397 the sd-journal logging socket, and, if the timeout is set to 0, sets
398 O_NONBLOCK on it. That way people can control if and when to block for
401 * hostnamed: populate form factor data from a new hwdb database, so that old
402 yogas can be recognized as "convertible" too, even if they predate the DMI
403 "convertible" form factor
405 * Maybe add a small tool invoked early at boot, that adds in or resizes
406 partitions automatically, to be used when the media used is actually larger
407 than the image written onto it is.
409 * Maybe add PrivatePIDs= as new unit setting, and do minimal PID namespacing
410 after all. Be strict however, only support the equivalent of nspawn's
411 --as-pid2 switch, and sanely proxy sd_notify() messages dropping stuff such
414 * Add ExecMonitor= setting. May be used multiple times. Forks off a process in
415 the service cgroup, which is supposed to monitor the service, and when it
416 exits the service is considered failed by its monitor.
418 * track the per-service PAM process properly (i.e. as an additional control
419 process), so that it may be queried on the bus and everything.
421 * add a new "debug" job mode, that is propagated to unit_start() and for
422 services results in two things: we raise SIGSTOP right before invoking
423 execve() and turn off watchdog support. Then, use that to implement
424 "systemd-gdb" for attaching to the start-up of any system service in its
427 * maybe introduce gpt auto discovery for /var/tmp?
429 * maybe add gpt-partition-based user management: each user gets his own
430 LUKS-encrypted GPT partition with a new GPT type. A small nss module
431 enumerates users via udev partition enumeration. UIDs are assigned in a fixed
432 way: the partition index is added as offset to some fixed base uid. User name
433 is stored in GPT partition name. A PAM module authenticates the user via the
434 LUKS partition password. Benefits: strong per-user security, compatibility
435 with stateless/read-only/verity-enabled root. (other idea: do this based on
436 loopback files in /home, without GPT involvement)
438 * gpt-auto logic: introduce support for discovering /var matching an image. For
439 that, use a partition type UUID that is hashed from the OS name (as encoded
440 in /etc/os-release), the architecture, and 4 new bits from the gpt flags
441 field of the root partition. This way can easily support multiple OS
442 installations on the same GPT partition table, without problems with
443 unmatched /var partitions.
445 * gpt-auto logic: related to the above, maybe support a "secondary" root
446 partition, that is mounted to / and is writable, and where the actual root's
447 /usr is mounted into.
449 * 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
451 * drop nss-myhostname in favour of nss-resolve?
453 * add a percentage syntax for TimeoutStopSec=, e.g. TimeoutStopSec=150%, and
454 then use that for the setting used in user@.service. It should be understood
455 relative to the configured default value.
457 * in networkd, when matching device types, fix up DEVTYPE rubbish the kernel passes to us
459 * enable LockMLOCK to take a percentage value relative to physical memory
461 * Permit masking specific netlink APIs with RestrictAddressFamily=
463 * nspawn: support that /proc, /sys/, /dev are pre-mounted
465 * define gpt header bits to select volatility mode
467 * ProtectKernelLogs= (drops CAP_SYSLOG, add seccomp for syslog() syscall, and DeviceAllow to /dev/kmsg) in service files
469 * ProtectClock= (drops CAP_SYS_TIMES, adds seecomp filters for settimeofday, adjtimex), sets DeviceAllow o /dev/rtc
471 * ProtectTracing= (drops CAP_SYS_PTRACE, blocks ptrace syscall, makes /sys/kernel/tracing go away)
473 * ProtectMount= (drop mount/umount/pivot_root from seccomp, disallow fuse via DeviceAllow, imply Mountflags=slave)
475 * ProtectKeyRing= to take keyring calls away
477 * RemoveKeyRing= to remove all keyring entries of the specified user
479 * ProtectReboot= that masks reboot() and kexec_load() syscalls, prohibits kill
480 on PID 1 with the relevant signals, and makes relevant files in /sys and
481 /proc (such as the sysrq stuff) unavailable
483 * make sure the ratelimit object can deal with USEC_INFINITY as way to turn off things
485 * journalctl: make sure -f ends when the container indicated by -M terminates
487 * mount: automatically search for "main" partition of an image has multiple
490 * in nss-systemd, if we run inside of RootDirectory= with PrivateUsers= set,
491 find a way to map the User=/Group= of the service to the right name. This way
492 a user/group for a service only has to exist on the host for the right
495 * add bus API for creating unit files in /etc, reusing the code for transient units
497 * add bus API to remove unit files from /etc
499 * add bus API to retrieve current unit file contents (i.e. implement "systemctl cat" on the bus only)
501 * rework fopen_temporary() to make use of open_tmpfile_linkable() (problem: the
502 kernel doesn't support linkat() that replaces existing files, currently)
504 * transient units: don't bother with actually setting unit properties, we
505 reload the unit file anyway
507 * journald: sigbus API via a signal-handler safe function that people may call
508 from the SIGBUS handler
510 * optionally, also require WATCHDOG=1 notifications during service start-up and shutdown
512 * delay activation of logind until somebody logs in, or when /dev/tty0 pulls it
513 in or lingering is on (so that containers don't bother with it until PAM is used). also exit-on-idle
515 * cache sd_event_now() result from before the first iteration...
517 * PID1: find a way how we can reload unit file configuration for
518 specific units only, without reloading the whole of systemd
520 * add an explicit parser for LimitRTPRIO= that verifies
521 the specified range and generates sane error messages for incorrect
524 * when we detect that there are waiting jobs but no running jobs, do something
526 * push CPUAffinity= also into the "cpuset" cgroup controller (only after the cpuset controller got ported to the unified hierarchy)
528 * PID 1 should send out sd_notify("WATCHDOG=1") messages (for usage in the --user mode, and when run via nspawn)
530 * there's probably something wrong with having user mounts below /sys,
531 as we have for debugfs. for example, src/core/mount.c handles mounts
532 prefixed with /sys generally special.
533 http://lists.freedesktop.org/archives/systemd-devel/2015-June/032962.html
535 * fstab-generator: default to tmpfs-as-root if only usr= is specified on the kernel cmdline
537 * docs: bring http://www.freedesktop.org/wiki/Software/systemd/MyServiceCantGetRealtime up to date
539 * add a job mode that will fail if a transaction would mean stopping
540 running units. Use this in timedated to manage the NTP service
542 http://lists.freedesktop.org/archives/systemd-devel/2015-April/030229.html
544 * The udev blkid built-in should expose a property that reflects
545 whether media was sensed in USB CF/SD card readers. This should then
546 be used to control SYSTEMD_READY=1/0 so that USB card readers aren't
547 picked up by systemd unless they contain a medium. This would mirror
548 the behaviour we already have for CD drives.
550 * networkd/udev: implement SR_IOV configuration in .link files:
551 http://lists.freedesktop.org/archives/systemd-devel/2015-January/027451.html
553 * hostnamectl: show root image uuid
555 * Find a solution for SMACK capabilities stuff:
556 http://lists.freedesktop.org/archives/systemd-devel/2014-December/026188.html
558 * "systemctl preset-all" should probably order the unit files it
559 operates on lexicographically before starting to work, in order to
560 ensure deterministic behaviour if two unit files conflict (like DMs
563 * synchronize console access with BSD locks:
564 http://lists.freedesktop.org/archives/systemd-devel/2014-October/024582.html
566 * as soon as we have sender timestamps, revisit coalescing multiple parallel daemon reloads:
567 http://lists.freedesktop.org/archives/systemd-devel/2014-December/025862.html
569 * in systemctl list-unit-files: show the install value the presets would suggest for a service in a third column
571 * figure out when we can use the coarse timers
573 * add "systemctl start -v foobar.service" that shows logs of a service
574 while the start command runs. This is non-trivial to do without
575 races though, since we should flush out all journal messages before
576 returning from the "systemctl stop".
578 * 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
580 * maybe add support for specifier expansion in user.conf, specifically DefaultEnvironment=
582 * maybe allow timer units with an empty Units= setting, so that they
583 can be used for resuming the system but nothing else.
585 * what to do about udev db binary stability for apps? (raw access is not an option)
587 * man: maybe use the word "inspect" rather than "introspect"?
589 * systemctl: if some operation fails, show log output?
591 * systemctl edit: use equivalent of cat() to insert existing config as a comment, prepended with #.
592 Upon editor exit, lines with one # are removed, lines with two # are left with one #, etc.
594 * exponential backoff in timesyncd when we cannot reach a server
596 * timesyncd: add ugly bus calls to set NTP servers per-interface, for usage by NM
598 * merge ~/.local/share and ~/.local/lib into one similar /usr/lib and /usr/share....
600 * systemd.show_status= should probably have a mode where only failed
603 * add systemd.abort_on_kill or some other such flag to send SIGABRT instead of SIGKILL
604 (throughout the codebase, not only PID1)
608 - service registration
609 - service/domain/types browsing
611 - DNS-SD service registration from socket units
612 - resolved should optionally register additional per-interface LLMNR
613 names, so that for the container case we can establish the same name
614 (maybe "host") for referencing the server, everywhere.
615 - allow clients to request DNSSEC for a single lookup even if DNSSEC is off (?)
616 - hook up resolved with machined-based address resolution
618 * refcounting in sd-resolve is borked
620 * Add a new verb "systemctl top"
622 * add new gpt type for btrfs volumes
624 * support empty /etc boots nicely:
625 - nspawn/gpt-generator: introduce new gpt partition type for /usr
627 * generator that automatically discovers btrfs subvolumes, identifies their purpose based on some xattr on them.
629 * a way for container managers to turn off getty starting via $container_headless= or so...
631 * figure out a nice way how we can let the admin know what child/sibling unit causes cgroup membership for a specific unit
633 * For timer units: add some mechanisms so that timer units that trigger immediately on boot do not have the services
634 they run added to the initial transaction and thus confuse Type=idle.
636 * add bus api to query unit file's X fields.
638 * gpt-auto-generator:
639 - Define new partition type for encrypted swap? Support probed LUKS for encrypted swap?
640 - Make /home automount rather than mount?
642 * add generator that pulls in systemd-network from containers when
643 CAP_NET_ADMIN is set, more than the loopback device is defined, even
644 when it is otherwise off
646 * MessageQueueMessageSize= (and suchlike) should use parse_iec_size().
648 * implement Distribute= in socket units to allow running multiple
649 service instances processing the listening socket, and open this up
652 * introduce bus call FreezeUnit(s, b), as well as "systemctl freeze
653 $UNIT" and "systemctl thaw $UNIT" as wrappers around this. The calls
654 should SIGSTOP all unit processes in a loop until all processes of
655 it are fully stopped. This can later be used for app management by
656 desktop UIs such as gnome-shell to freeze apps that are not visible
657 on screen, not unlike how job control works on the shell
660 - implement per-slice CPUFairScheduling=1 switch
661 - introduce high-level settings for RT budget, swappiness
662 - how to reset dynamically changed unit cgroup attributes sanely?
663 - when reloading configuration, apply new cgroup configuration
664 - when recursively showing the cgroup hierarchy, optionally also show
665 the hierarchies of child processes
668 - add field to transient units that indicate whether systemd or somebody else saves/restores its settings, for integration with libvirt
670 * when we detect low battery and no AC on boot, show pretty splash and refuse boot
672 * libsystemd-journal, libsystemd-login, libudev: add calls to easily attach these objects to sd-event event loops
674 * be more careful what we export on the bus as (usec_t) 0 and (usec_t) -1
676 * 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
678 * After coming back from hibernation reset hibernation swap partition using the /dev/snapshot ioctl APIs
680 * If we try to find a unit via a dangling symlink, generate a clean
681 error. Currently, we just ignore it and read the unit from the search
684 * refuse boot if /usr/lib/os-release is missing or /etc/machine-id cannot be set up
686 * man: the documentation of Restart= currently is very misleading and suggests the tools from ExecStartPre= might get restarted.
688 * load .d/*.conf dropins for device units
692 - GetAllProperties() on a non-existing object does not result in a failure currently
693 - port to sd-resolve for connecting to TCP dbus servers
694 - 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
695 - see if we can drop more message validation on the sending side
696 - add API to clone sd_bus_message objects
697 - longer term: priority inheritance
699 - NameLost/NameAcquired obsolete
702 - update systemd.special(7) to mention that dbus.socket is only about the compatibility socket now
705 - allow multiple signal handlers per signal?
706 - document chaining of signal handler for SIGCHLD and child handlers
707 - define more intervals where we will shift wakeup intervals around in, 1h, 6h, 24h, ...
708 - generate a failure of a default event loop is executed out-of-thread
710 * investigate endianness issues of UUID vs. GUID
712 * dbus: when a unit failed to load (i.e. is in UNIT_ERROR state), we
713 should be able to safely try another attempt when the bus call LoadUnit() is invoked.
715 * add a pam module that passes the hdd passphrase into the PAM stack and then expires it, for usage by gdm auto-login.
717 * add a pam module that on password changes updates any LUKS slot where the password matches
720 - add unit tests for config_parse_device_allow()
722 * seems that when we follow symlinks to units we prefer the symlink
723 destination path over /etc and /usr. We should not do that. Instead
724 /etc should always override /run+/usr and also any symlink
727 * when isolating, try to figure out a way how we implicitly can order
728 all units we stop before the isolating unit...
730 * teach ConditionKernelCommandLine= globs or regexes (in order to match foobar={no,0,off})
732 * BootLoaderSpec: Clarify that the kernel has to be in $BOOT. Clarify
733 that the boot loader should be installed to the ESP. Define a way
734 how an installer can figure out whether a BLS compliant boot loader
737 * think about requeuing jobs when daemon-reload is issued? usecase:
738 the initrd issues a reload after fstab from the host is accessible
739 and we might want to requeue the mounts local-fs acquired through
742 * systemd-inhibit: make taking delay locks useful: support sending SIGINT or SIGTERM on PrepareForSleep()
744 * remove any syslog support from log.c — we probably cannot do this before split-off udev is gone for good
746 * shutdown logging: store to EFI var, and store to USB stick?
748 * merge unit_kill_common() and unit_kill_context()
751 - honor language efi variables for default language selection (if there are any?)
752 - honor timezone efi variables for default timezone selection (if there are any?)
753 - change bootctl to be backed by systemd-bootd to control temporary and persistent default boot goal plus efi variables
755 - recognize the case when not booted on EFI
757 * maybe do not install getty@tty1.service symlink in /etc but in /usr?
759 * print a nicer explanation if people use variable/specifier expansion in ExecStart= for the first word
761 * mount: turn dependency information from /proc/self/mountinfo into dependency information between systemd units.
764 - logind: optionally, ignore idle-hint logic for autosuspend, block suspend as long as a session is around
765 - logind: wakelock/opportunistic suspend support
766 - Add pretty name for seats in logind
767 - logind: allow showing logout dialog from system?
768 - add Suspend() bus calls which take timestamps to fix double suspend issues when somebody hits suspend and closes laptop quickly.
769 - if pam_systemd is invoked by su from a process that is outside of a
770 any session we should probably just become a NOP, since that's
771 usually not a real user session but just some system code that just
773 - logind: make the Suspend()/Hibernate() bus calls wait for the for
774 the job to be completed. before returning, so that clients can wait
775 for "systemctl suspend" to finish to know when the suspending is
777 - logind: when the power button is pressed short, just popup a
778 logout dialog. If it is pressed for 1s, do the usual
779 shutdown. Inspiration are Macs here.
780 - expose "Locked" property on logind sesison objects
781 - maybe allow configuration of the StopTimeout for session scopes
782 - rename session scope so that it includes the UID. THat way
783 the session scope can be arranged freely in slices and we don't have
784 make assumptions about their slice anymore.
785 - follow PropertiesChanged state more closely, to deal with quick logouts and
789 - consider introducing implicit _TTY= + _PPID= + _EUID= + _EGID= + _FSUID= + _FSGID= fields
790 - journald: also get thread ID from client, plus thread name
791 - journal: when waiting for journal additions in the client always sleep at least 1s or so, in order to minimize wakeups
792 - add API to close/reopen/get fd for journal client fd in libsystemd-journal.
793 - fallback to /dev/log based logging in libsystemd-journal, if we cannot log natively?
794 - declare the local journal protocol stable in the wiki interface chart
795 - sd-journal: speed up sd_journal_get_data() with transparent hash table in bg
796 - journald: when dropping msgs due to ratelimit make sure to write
797 "dropped %u messages" not only when we are about to print the next
798 message that works, but already after a short timeout
799 - check if we can make journalctl by default use --follow mode inside of less if called without args?
800 - maybe add API to send pairs of iovecs via sd_journal_send
801 - journal: add a setgid "systemd-journal" utility to invoke from libsystemd-journal, which passes fds via STDOUT and does PK access
802 - journactl: support negative filtering, i.e. FOOBAR!="waldo",
803 and !FOOBAR for events without FOOBAR.
804 - journal: store timestamp of journal_file_set_offline() int he header,
805 so it is possible to display when the file was last synced.
806 - 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.
807 - journal: find a way to allow dropping history early, based on priority, other rules
808 - journal: When used on NFS, check payload hashes
809 - journald: add kernel cmdline option to disable ratelimiting for debug purposes
810 - refuse taking lower-case variable names in sd_journal_send() and friends.
811 - journald: we currently rotate only after MaxUse+MaxFilesize has been reached.
812 - journal: deal nicely with byte-by-byte copied files, especially regards header
813 - journal: sanely deal with entries which are larger than the individual file size, but where the components would fit
814 - Replace utmp, wtmp, btmp, and lastlog completely with journal
815 - journalctl: instead --after-cursor= maybe have a --cursor=XYZ+1 syntax?
816 - when a kernel driver logs in a tight loop, we should ratelimit that too.
817 - journald: optionally, log debug messages to /run but everything else to /var
818 - journald: when we drop syslog messages because the syslog socket is
819 full, make sure to write how many messages are lost as first thing
820 to syslog when it works again.
821 - change systemd-journal-flush into a service that stays around during
822 boot, and causes the journal to be moved back to /run on shutdown,
823 so that we do not keep /var busy. This needs to happen synchronously,
824 hence doing this via signals is not going to work.
825 - optionally support running journald from the command line for testing purposes in external projects
826 - journald: allow per-priority and per-service retention times when rotating/vacuuming
827 - journald: make use of uid-range.h to managed uid ranges to split
829 - journalctl: add the ability to look for the most recent process of a binary. journalctl /usr/bin/X11 --pid=-1 or so...
830 - improve journalctl performance by loading journal files
831 lazily. Encode just enough information in the file name, so that we
832 do not have to open it to know that it is not interesting for us, for
833 the most common operations.
834 - man: document that corrupted journal files is nothing to act on
835 - rework journald sigbus stuff to use mutex
836 - Set RLIMIT_NPROC for systemd-journal-xyz, and all other of our
837 services that run under their own user ids, and use User= (but only
838 in a world where userns is ubiquitous since otherwise we cannot
839 invoke those daemons on the host AND in a container anymore). Also,
840 if LimitNPROC= is used without User= we should warn and refuse
842 - journalctl --verify: don't show files that are currently being
843 written to as FAIL, but instead show that their are being written to.
844 - add journalctl -H that talks via ssh to a remote peer and passes through
846 - add a version of --merge which also merges /var/log/journal/remote
847 - journalctl: -m should access container journals directly by enumerating
848 them via machined, and also watch containers coming and going.
849 Benefit: nspawn --ephemeral would start working nicely with the journal.
850 - assign MESSAGE_ID to log messages about failed services
852 * add a test if all entries in the catalog are properly formatted.
853 (Adding dashes in a catalog entry currently results in the catalog entry
854 being silently skipped. journalctl --update-catalog must warn about this,
855 and we should also have a unit test to check that all our message are OK.)
858 - document that deps in [Unit] sections ignore Alias= fields in
859 [Install] units of other units, unless those units are disabled
860 - man: clarify that time-sync.target is not only sysv compat but also useful otherwise. Same for similar targets
861 - document that service reload may be implemented as service reexec
862 - add a man page containing packaging guidelines and recommending usage of things like Documentation=, PrivateTmp=, PrivateNetwork= and ReadOnlyDirectories=/etc /usr.
863 - document systemd-journal-flush.service properly
864 - documentation: recommend to connect the timer units of a service to the service via Also= in [Install]
865 - man: document the very specific env the shutdown drop-in tools live in
866 - man: add more examples to man pages
867 - man: maybe sort directives in man pages, and take sections from --help and apply them to man too
868 - document root=gpt-auto properly
871 - add systemctl switch to dump transaction without executing it
872 - Add a verbose mode to "systemctl start" and friends that explains what is being done or not done
873 - "systemctl disable" on a static unit prints no message and does
874 nothing. "systemctl enable" does nothing, and gives a bad message
875 about it. Should fix both to print nice actionable messages.
876 - print nice message from systemctl --failed if there are no entries shown, and hook that into ExecStartPre of rescue.service/emergency.service
877 - add new command to systemctl: "systemctl system-reexec" which reexecs as many daemons as virtually possible
878 - systemctl enable: fail if target to alias into does not exist? maybe show how many units are enabled afterwards?
879 - systemctl: "Journal has been rotated since unit was started." message is misleading
880 - systemctl status output should include list of triggering units and their status
883 - "systemctl mask" should find all names by which a unit is accessible
884 (i.e. by scanning for symlinks to it) and link them all to /dev/null
887 - timer units should get the ability to trigger when:
889 - Modulate timer frequency based on battery state
891 * add libsystemd-password or so to query passwords during boot using the password agent logic
893 * clean up date formatting and parsing so that all absolute/relative timestamps we format can also be parsed
895 * on shutdown: move utmp, wall, audit logic all into PID 1 (or logind?), get rid of systemd-update-utmp-runlevel
897 * make repeated alt-ctrl-del presses printing a dump
899 * hostnamed: before returning information from /etc/machine-info.conf check the modification data and reread. Similar for localed, ...
901 * currently x-systemd.timeout is lost in the initrd, since crypttab is copied into dracut, but fstab is not
904 - emulate /dev/kmsg using CUSE and turn off the syslog syscall
905 with seccomp. That should provide us with a useful log buffer that
906 systemd can log to during early boot, and disconnect container logs
907 from the kernel's logs.
908 - as soon as networkd has a bus interface, hook up --network-interface=,
909 --network-bridge= with networkd, to trigger netdev creation should an
911 - a nice way to boot up without machine id set, so that it is set at boot
912 automatically for supporting --ephemeral. Maybe hash the host machine id
913 together with the machine name to generate the machine id for the container
914 - fix logic always print a final newline on output.
915 https://github.com/systemd/systemd/pull/272#issuecomment-113153176
916 - should optionally support receiving WATCHDOG=1 messages from its payload
918 - optionally automatically add FORWARD rules to iptables whenever nspawn is
919 running, remove them when shut down.
922 - refuse mounting over a mount point
923 - automatically discover .roothash files in dissect, similarly to nspawn
926 - add an API so that libvirt-lxc can inform us about network interfaces being
927 removed or added to an existing machine
928 - "machinectl migrate" or similar to copy a container from or to a
929 difference host, via ssh
930 - introduce systemd-nspawn-ephemeral@.service, and hook it into
931 "machinectl start" with a new --ephemeral switch
932 - "machinectl status" should also show internal logs of the container in
934 - "machinectl history"
936 - "machinectl commit" that takes a writable snapshot of a tree, invokes a
937 shell in it, and marks it read-only after use
940 - cryptsetup-generator: allow specification of passwords in crypttab itself
941 - support rd.luks.allow-discards= kernel cmdline params in cryptsetup generator
943 * hw watchdog: optionally try to use the preset watchdog timeout instead of always overriding it
944 https://bugs.freedesktop.org/show_bug.cgi?id=54712
946 * add a dependency on standard-conf.xml and other included files to man pages
948 * MountFlags=shared acts as MountFlags=slave right now.
950 * properly handle loop back mounts via fstab, especially regards to fsck/passno
952 * initialize the hostname from the fs label of /, if /etc/hostname does not exist?
957 - add trigger --subsystem-match=usb/usb_device device
958 - reimport udev db after MOVE events for devices without dev_t
960 * There's currently no way to cancel fsck (used to be possible via C-c or c on the console)
962 * add option to sockets to avoid activation. Instead just drop packets/connections, see http://cyberelk.net/tim/2012/02/15/portreserve-systemd-solution/
965 - save coredump in Windows/Mozilla minidump format
966 - when truncating coredumps, also log the full size that the process had, and make a metadata field so we can report truncated coredumps
968 * support crash reporting operation modes (https://live.gnome.org/GnomeOS/Design/Whiteboards/ProblemReporting)
970 * be able to specify a forced restart of service A where service B depends on, in case B
971 needs to be auto-respawned?
974 - apply "x" on "D" too (see patch from William Douglas)
975 - instead of ignoring unknown fields, reject them.
976 - creating new directories/subvolumes/fifos/device nodes
977 should not follow symlinks. None of the other adjustment or creation
978 calls follow symlinks.
980 * make sure systemd-ask-password-wall does not shutdown systemd-ask-password-console too early
982 * verify that the AF_UNIX sockets of a service in the fs still exist
983 when we start a service in order to avoid confusion when a user
984 assumes starting a service is enough to make it accessible
986 * Make it possible to set the keymap independently from the font on
987 the kernel cmdline. Right now setting one resets also the other.
989 * and a dbus call to generate target from current state
991 * write blog stories about:
992 - hwdb: what belongs into it, lsusb
993 - enabling dbus services
994 - how to make changes to sysctl and sysfs attributes
996 - how to pass throw-away units to systemd, or dynamically change properties of existing units
997 - testing with Harald's awesome test kit
999 - how to develop against journal browsing APIs
1000 - the journal HTTP iface
1001 - non-cgroup resource management
1002 - dynamic resource management with cgroups
1003 - refreshed, longer missions statement
1004 - calendar time events
1005 - init=/bin/sh vs. "emergency" mode, vs. "rescue" mode, vs. "multi-user" mode, vs. "graphical" mode, and the debug shell
1006 - how to create your own target
1007 - instantiated apache, dovecot and so on
1008 - hooking a script into various stages of shutdown/rearly booot
1010 * investigate whether the gnome pty helper should be moved into systemd, to provide cgroup support.
1012 * dot output for --test showing the 'initial transaction'
1015 - When logging about multiple units (stopping BoundTo units, conflicts, etc.),
1016 log both units as UNIT=, so that journalctl -u triggers on both.
1017 - generate better errors when people try to set transient properties
1018 that are not supported...
1019 http://lists.freedesktop.org/archives/systemd-devel/2015-February/028076.html
1020 - maybe introduce WantsMountsFor=? Usecase:
1021 http://lists.freedesktop.org/archives/systemd-devel/2015-January/027729.html
1022 - recreate systemd's D-Bus private socket file on SIGUSR2
1023 - move PAM code into its own binary
1024 - when we automatically restart a service, ensure we restart its rdeps, too.
1025 - hide PAM options in fragment parser when compile time disabled
1026 - Support --test based on current system state
1027 - 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().
1028 - after deserializing sockets in socket.c we should reapply sockopts and things
1029 - drop PID 1 reloading, only do reexecing (difficult: Reload()
1030 currently is properly synchronous, Reexec() is weird, because we
1031 cannot delay the response properly until we are back, so instead of
1032 being properly synchronous we just keep open the fd and close it
1033 when done. That means clients do not get a successful method reply,
1034 but much rather a disconnect on success.
1035 - when breaking cycles drop sysv services first, then services from /run, then from /etc, then from /usr
1036 - when a bus name of a service disappears from the bus make sure to queue further activation requests
1039 - allow port=0 in .socket units
1040 - maybe introduce ExecRestartPre=
1041 - add ReloadSignal= for configuring a reload signal to use
1042 - implement Register= switch in .socket units to enable registration
1043 in Avahi, RPC and other socket registration services.
1044 - allow Type=simple with PIDFile=
1045 https://bugzilla.redhat.com/show_bug.cgi?id=723942
1046 - allow writing multiple conditions in unit files on one line
1047 - load-fragment: when loading a unit file via a chain of symlinks
1048 verify that it is not masked via any of the names traversed.
1049 - introduce Type=pid-file
1050 - introduce mix of BindTo and Requisite
1051 - add a concept of RemainAfterExit= to scope units
1052 - Allow multiple ExecStart= for all Type= settings, so that we can cover rescue.service nicely
1055 - Make sure ID_PATH is always exported and complete for
1056 network devices where possible, so we can safely rely
1060 - add support for more attribute types
1061 - inbuilt piping support (essentially degenerate async)? see loopback-setup.c and other places
1064 - add more keys to [Route] and [Address] sections
1065 - add support for more DHCPv4 options (and, longer term, other kinds of dynamic config)
1066 - add proper initrd support (in particular generate .network/.link files based on /proc/cmdline)
1067 - add reduced [Link] support to .network files
1068 - add Scope= parsing option for [Network]
1069 - properly handle routerless dhcp leases
1070 - work with non-Ethernet devices
1071 - add support for more bond options
1072 - dhcp: do we allow configuring dhcp routes on interfaces that are not the one we got the dhcp info from?
1073 - the DHCP lease data (such as NTP/DNS) is still made available when
1074 a carrier is lost on a link. It should be removed instantly.
1075 - expose in the API the following bits:
1076 - option 15, domain name and/or option 119, search list
1077 - option 12, host name and/or option 81, fqdn
1078 - option 123, 144, geolocation
1079 - option 252, configure http proxy (PAC/wpad)
1080 - provide a way to define a per-network interface default metric value
1081 for all routes to it. possibly a second default for DHCP routes.
1082 - allow Name= to be specified repeatedly in the [Match] section. Maybe also
1083 support Name=foo*|bar*|baz ?
1084 - duplicate address check for static IPs (like ARPCHECK in network-scripts)
1085 - allow DUID/IAID to be customized, see issue #394.
1086 - whenever uplink info changes, make DHCP server send out FORCERENEW
1088 * networkd-wait-online:
1089 - make operstates to wait for configurable?
1092 - figure out how much we can increase Maximum Message Size
1095 - add functions to set previously stored IPv6 addresses on startup and get
1096 them at shutdown; store them in client->ia_na
1097 - write more test cases
1098 - implement reconfigure support, see 5.3., 15.11. and 22.20.
1099 - implement support for temporary adressess (IA_TA)
1100 - implement dhcpv6 authentication
1101 - investigate the usefulness of Confirm messages; i.e. are there any
1102 situations where the link changes without any loss in carrier detection
1104 - some servers don't do rapid commit without a filled in IA_NA, verify
1111 - natively watch for dbus-*.service symlinks (PENDING)
1112 - teach dbus to activate all services it finds in /etc/systemd/services/org-*.service
1114 * fix alsa mixer restore to not print error when no config is stored
1116 * make cryptsetup lower --iter-time
1118 * patch kernel for xattr support in /dev, /proc/, /sys?
1120 * kernel: add device_type = "fb", "fbcon" to class "graphics"
1122 * /usr/bin/service should actually show the new command line
1124 * 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=
1126 * fedora: F20: go timer units all the way, leave cron.daily for cron
1128 * neither pkexec nor sudo initialize environ[] from the PAM environment?
1130 * fedora: update policy to declare access mode and ownership of unit files to root:root 0644, and add an rpmlint check for it
1132 * register catalog database signature as file magic
1134 * zsh shell completion:
1135 - <command> <verb> -<TAB> should complete options, but currently does not
1136 - systemctl add-wants,add-requires
1138 * systemctl status should know about 'systemd-analyze calendar ... --iterations='
1139 * If timer has just OnInactiveSec=..., it should fire after a specified time
1140 after being started.
1144 * look for close() vs. close_nointr() vs. close_nointr_nofail()
1146 * check for strerror(r) instead of strerror(-r)
1150 * set_put(), hashmap_put() return values check. i.e. == 0 does not free()!
1152 * use secure_getenv() instead of getenv() where appropriate
1154 * link up selected blog stories from man pages and unit files Documentation= fields
1155 String is not UTF-8 clean, ignoring assignment
1156 timedatex.service: Consumed 26ms CPU time.