]> git.ipfire.org Git - thirdparty/systemd.git/blob - TODO
update TODO
[thirdparty/systemd.git] / TODO
1 Bugfixes:
2
3 * Should systemctl status \* work on all unit types, not just .service?
4
5 * Dangling symlinks of .automount unit files in .wants/ directories, set up
6 automount points even when the original .automount file did not exist
7 anymore. Only the .mount unit was still around.
8
9 * ExecStart with unicode characters fails in strv_split_extract:
10
11 [Service]
12 Environment=ONE='one' "TWO='two two' too" THREE=
13 ExecStart=/bin/python3 -c 'import sys;print(sys.argv)' $ONE $TWO $THREE
14
15 External:
16
17 * Fedora: add an rpmlint check that verifies that all unit files in the RPM are listed in %systemd_post macros.
18
19 * wiki: update journal format documentation for lz4 additions
20
21 Janitorial Clean-ups:
22
23 * Rearrange tests so that the various test-xyz.c match a specific src/basic/xyz.c again
24
25 Features:
26
27 * rework pid watching logic, allow any number of units watch the same pid
28
29 * be stricter with fds we receive for the fdstore: close them asynchronously
30
31 * be stricter with pid file and notify pids: don't allow them to be outside of cgroup, except if sender is privileged
32
33 * calenderspec: add support for week numbers and day numbers within a
34 year. This would allow us to define "bi-weekly" triggers safely.
35
36 * add support for recursive bpf firewalling as supported by the newest kernel
37
38 * add bpf-based implementation of devices cgroup controller logic for compat with cgroupsv2 as supported by newest kernel
39
40 * emulate properties of the root cgroup on controllers that don't support such
41 properties natively on cpu/io/memory, the way we already do it for
42 "pids". Also, add the same logic to cgtop.
43
44 * set TasksAccounting=1 on the root slice if we are running on the root cgroup,
45 and similar for the others, as soon as we emulate them properly. After all,
46 Linux keeps these system-wide stats anyway, and it costs nothing to expose
47 them.
48
49 * sd-bus: add vtable flag, that may be used to request client creds implicitly
50 and asynchronously before dispatching the operation
51
52 * implement transient socket unit.
53
54 * make systemd-run create transient path and socket unit.
55
56 * make use of ethtool veth peer info in machined, for automatically finding out
57 host-side interface pointing to the container.
58
59 * add some special mode to LogsDirectory=/StateDirectory=… that allows
60 declaring these directories without necessarily pulling in deps for them, or
61 creating them when starting up. That way, we could declare that
62 systemd-journald writes to /var/log/journal, which could be useful when we
63 doing disk usage calculations and so on.
64
65 * taint systemd if there are fewer than 65536 users assigned to the system.
66
67 * deprecate PermissionsStartOnly= and RootDirectoryStartOnly= in favour of the ExecStart= prefix chars
68
69 * add a new RuntimeDirectoryPreserve= mode that defines a similar lifecycle for
70 the runtime dir as we maintain for the fdstore: i.e. keep it around as long
71 as the unit is running or has a job queued.
72
73 * support projid-based quota in machinectl for containers, and then drop
74 implicit btrfs loopback magic in machined
75
76 * Add NetworkNamespacePath= to specify a path to a network namespace
77
78 * maybe use SOURCE_DATE_EPOCH (i.e. the env var the reproducible builds folks
79 introduced) as the RTC epoch, instead of the mtime of NEWS.
80
81 * add a way to lock down cgroup migration: a boolean, which when set for a unit
82 makes sure the processes in it can never migrate out of it
83
84 * blog about fd store and restartable services
85
86 * document Environment=SYSTEMD_LOG_LEVEL=debug drop-in in debugging document
87
88 * rework ExecOutput and ExecInput enums so that EXEC_OUTPUT_NULL loses its
89 magic meaning and is no longer upgraded to something else if set explicitly.
90
91 * in the long run: permit a system with /etc/machine-id linked to /dev/null, to
92 make it lose its identity, i.e. be anonymous. For this we'd have to patch
93 through the whole tree to make all code deal with the case where no machine
94 ID is available.
95
96 * optionally, collect cgroup resource data, and store it in per-unit RRD files,
97 suitable for processing with rrdtool. Add bus API to access this data, and
98 possibly implement a CPULoad property based on it.
99
100 * beef up pam_systemd to take unit file settings such as cgroups properties as
101 parameters
102
103 * a new "systemd-analyze security" tool outputting a checklist of security
104 features a service does and does not implement
105
106 * maybe hook of xfs/ext4 quotactl() with services? i.e. automatically manage
107 the quota of a the user indicated in User= via unit file settings, like the
108 other resource management concepts. Would mix nicely with DynamicUser=1. Or
109 alternatively, do this with projids, so that we can also cover services
110 running as root. Quota should probably cover all the special dirs such as
111 StateDirectory=, LogsDirectory=, CacheDirectory=, as well as RootDirectory= if it
112 is set, plus the whole disk space any image configured with RootImage=.
113
114 * Introduce "exit" as an EmergencyAction value, and allow to configure a
115 per-unit success/failure exit code to configure. This would be useful for
116 running commands inside of services inside of containers, which could then
117 propagate their failure state all the way up.
118
119 * In DynamicUser= mode: before selecting a UID, use disk quota APIs on relevant
120 disks to see if the UID is already in use.
121
122 * add dissect_image_warn() as a wrapper around dissect_image() that prints
123 friendly log messages for the returned errors, so that we don't have to
124 duplicate that in nspawn, systemd-dissect and PID 1.
125
126 * add "systemctl wait" or so, which does what "systemd-run --wait" does, but
127 for all units. It should be both a way to pin units into memory as well as a
128 wait to retrieve their exit data.
129
130 * maybe set a new set of env vars for services, based on RuntimeDirectory=,
131 StateDirectory=, LogsDirectory=, CacheDirectory= and ConfigurationDirectory=
132 automatically. For example, there could be $RUNTIME_DIRECTORY,
133 $STATE_DIRECTORY, $LOGS_DIRECTORY=, $CACHE_DIRECTORY and
134 $CONFIGURATION_DIRECTORY or so. This could be useful to write services that
135 can adapt to varying directories for these purposes. Special care has to be
136 taken if multiple dirs are configured. Maybe avoid setting the env vars in
137 that case?
138
139 * expose IO accounting data on the bus, show it in systemd-run --wait and log
140 about it in the resource log message
141
142 * add "systemctl purge" for flushing out configuration, state, logs, ... of a
143 unit when it is stopped
144
145 * show whether a service has out-of-date configuration in "systemctl status" by
146 using mtime data of ConfigurationDirectory=.
147
148 * replace all uses of fgets() + LINE_MAX by read_line()
149
150 * Add AddUser= setting to unit files, similar to DynamicUser=1 which however
151 creates a static, persistent user rather than a dynamic, transient user. We
152 can leverage code from sysusers.d for this.
153
154 * add some optional flag to ReadWritePaths= and friends, that has the effect
155 that we create the dir in question when the service is started. Example:
156
157 ReadWritePaths=:/var/lib/foobar
158
159 * maybe add call sd_journal_set_block_timeout() or so to set SO_SNDTIMEO for
160 the sd-journal logging socket, and, if the timeout is set to 0, sets
161 O_NONBLOCK on it. That way people can control if and when to block for
162 logging.
163
164 * tighten sd_notify() MAINPID= checks a bit: don't accept foreign PIDs (i.e.
165 PIDs not managed by the service manager)
166
167 * hostnamed: populate form factor data from a new hwdb database, so that old
168 yogas can be recognized as "convertible" too, even if they predate the DMI
169 "convertible" form factor
170
171 * Maybe add a small tool invoked early at boot, that adds in or resizes
172 partitions automatically, to be used when the media used is actually larger
173 than the image written onto it is.
174
175 * Maybe add PrivatePIDs= as new unit setting, and do minimal PID namespacing
176 after all. Be strict however, only support the equivalent of nspawn's
177 --as-pid2 switch, and sanely proxy sd_notify() messages dropping stuff such
178 as MAINPID.
179
180 * Add ExecMonitor= setting. May be used multiple times. Forks off a process in
181 the service cgroup, which is supposed to monitor the service, and when it
182 exits the service is considered failed by its monitor.
183
184 * track the per-service PAM process properly (i.e. as an additional control
185 process), so that it may be queried on the bus and everything.
186
187 * add a new "debug" job mode, that is propagated to unit_start() and for
188 services results in two things: we raise SIGSTOP right before invoking
189 execve() and turn off watchdog support. Then, use that to implement
190 "systemd-gdb" for attaching to the start-up of any system service in its
191 natural habitat.
192
193 * maybe introduce gpt auto discovery for /var/tmp?
194
195 * maybe add gpt-partition-based user management: each user gets his own
196 LUKS-encrypted GPT partition with a new GPT type. A small nss module
197 enumerates users via udev partition enumeration. UIDs are assigned in a fixed
198 way: the partition index is added as offset to some fixed base uid. User name
199 is stored in GPT partition name. A PAM module authenticates the user via the
200 LUKS partition password. Benefits: strong per-user security, compatibility
201 with stateless/read-only/verity-enabled root. (other idea: do this based on
202 loopback files in /home, without GPT involvement)
203
204 * gpt-auto logic: introduce support for discovering /var matching an image. For
205 that, use a partition type UUID that is hashed from the OS name (as encoded
206 in /etc/os-release), the architecture, and 4 new bits from the gpt flags
207 field of the root partition. This way can easily support multiple OS
208 installations on the same GPT partition table, without problems with
209 unmatched /var partitions.
210
211 * gpt-auto logic: related to the above, maybe support a "secondary" root
212 partition, that is mounted to / and is writable, and where the actual root's
213 /usr is mounted into.
214
215 * 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
216
217 * drop nss-myhostname in favour of nss-resolve?
218
219 * add a percentage syntax for TimeoutStopSec=, e.g. TimeoutStopSec=150%, and
220 then use that for the setting used in user@.service. It should be understood
221 relative to the configured default value.
222
223 * in networkd, when matching device types, fix up DEVTYPE rubbish the kernel passes to us
224
225 * enable LockMLOCK to take a percentage value relative to physical memory
226
227 * Permit masking specific netlink APIs with RestrictAddressFamily=
228
229 * nspawn: support that /proc, /sys/, /dev are pre-mounted
230
231 * define gpt header bits to select volatility mode
232
233 * ProtectKernelLogs= (drops CAP_SYSLOG, add seccomp for syslog() syscall, and DeviceAllow to /dev/kmsg) in service files
234
235 * ProtectClock= (drops CAP_SYS_TIMES, adds seecomp filters for settimeofday, adjtimex), sets DeviceAllow o /dev/rtc
236
237 * ProtectTracing= (drops CAP_SYS_PTRACE, blocks ptrace syscall, makes /sys/kernel/tracing go away)
238
239 * ProtectMount= (drop mount/umount/pivot_root from seccomp, disallow fuse via DeviceAllow, imply Mountflags=slave)
240
241 * ProtectKeyRing= to take keyring calls away
242
243 * RemoveKeyRing= to remove all keyring entries of the specified user
244
245 * ProtectReboot= that masks reboot() and kexec_load() syscalls, prohibits kill
246 on PID 1 with the relevant signals, and makes relevant files in /sys and
247 /proc (such as the sysrq stuff) unavailable
248
249 * DeviceAllow= should also generate seccomp filters for mknod()
250
251 * make sure the ratelimit object can deal with USEC_INFINITY as way to turn off things
252
253 * journalctl: make sure -f ends when the container indicated by -M terminates
254
255 * mount: automatically search for "main" partition of an image has multiple
256 partitions
257
258 * expose the "privileged" flag of ExecCommand on the bus, and open it up to
259 transient units
260
261 * in nss-systemd, if we run inside of RootDirectory= with PrivateUsers= set,
262 find a way to map the User=/Group= of the service to the right name. This way
263 a user/group for a service only has to exist on the host for the right
264 mapping to work.
265
266 * add bus API for creating unit files in /etc, reusing the code for transient units
267
268 * add bus API to remove unit files from /etc
269
270 * add bus API to retrieve current unit file contents (i.e. implement "systemctl cat" on the bus only)
271
272 * rework fopen_temporary() to make use of open_tmpfile_linkable() (problem: the
273 kernel doesn't support linkat() that replaces existing files, currently)
274
275 * check if DeviceAllow= should split first, resolve specifiers later
276
277 * transient units: don't bother with actually setting unit properties, we
278 reload the unit file anyway
279
280 * journald: sigbus API via a signal-handler safe function that people may call
281 from the SIGBUS handler
282
283 * optionally, also require WATCHDOG=1 notifications during service start-up and shutdown
284
285 * resolved: when routing queries, make sure only look for the *longest* suffix...
286
287 * delay activation of logind until somebody logs in, or when /dev/tty0 pulls it
288 in or lingering is on (so that containers don't bother with it until PAM is used). also exit-on-idle
289
290 * cache sd_event_now() result from before the first iteration...
291
292 * add systemctl stop --job-mode=triggering that follows TRIGGERED_BY deps and adds them to the same transaction
293
294 * PID1: find a way how we can reload unit file configuration for
295 specific units only, without reloading the whole of systemd
296
297 * add an explicit parser for LimitRTPRIO= that verifies
298 the specified range and generates sane error messages for incorrect
299 specifications.
300
301 * when we detect that there are waiting jobs but no running jobs, do something
302
303 * push CPUAffinity= also into the "cpuset" cgroup controller (only after the cpuset controller got ported to the unified hierarchy)
304
305 * PID 1 should send out sd_notify("WATCHDOG=1") messages (for usage in the --user mode, and when run via nspawn)
306
307 * there's probably something wrong with having user mounts below /sys,
308 as we have for debugfs. for exmaple, src/core/mount.c handles mounts
309 prefixed with /sys generally special.
310 http://lists.freedesktop.org/archives/systemd-devel/2015-June/032962.html
311
312 * fstab-generator: default to tmpfs-as-root if only usr= is specified on the kernel cmdline
313
314 * docs: bring http://www.freedesktop.org/wiki/Software/systemd/MyServiceCantGetRealtime up to date
315
316 * mounting and unmounting mount points manually with different source
317 devices will result in collected on all devices used.
318 http://lists.freedesktop.org/archives/systemd-devel/2015-April/030225.html
319
320 * add a job mode that will fail if a transaction would mean stopping
321 running units. Use this in timedated to manage the NTP service
322 state.
323 http://lists.freedesktop.org/archives/systemd-devel/2015-April/030229.html
324
325 * Maybe add support for the equivalent of "ethtool advertise" to .link files?
326 http://lists.freedesktop.org/archives/systemd-devel/2015-April/030112.html
327
328 * The udev blkid built-in should expose a property that reflects
329 whether media was sensed in USB CF/SD card readers. This should then
330 be used to control SYSTEMD_READY=1/0 so that USB card readers aren't
331 picked up by systemd unless they contain a medium. This would mirror
332 the behaviour we already have for CD drives.
333
334 * networkd/udev: implement SR_IOV configuration in .link files:
335 http://lists.freedesktop.org/archives/systemd-devel/2015-January/027451.html
336
337 * Rework systemctl's GetAll property parsing to use the generic bus_map_all_properties() API
338
339 * Port various tools to make use of verbs.[ch], where applicable: busctl,
340 coredumpctl, hostnamectl, localectl, timedatectl
341
342 * hostnamectl: show root image uuid
343
344 * sysfs set api in libudev is not const
345
346 * Find a solution for SMACK capabilities stuff:
347 http://lists.freedesktop.org/archives/systemd-devel/2014-December/026188.html
348
349 * "systemctl preset-all" should probably order the unit files it
350 operates on lexicographically before starting to work, in order to
351 ensure deterministic behaviour if two unit files conflict (like DMs
352 do, for example)
353
354 * synchronize console access with BSD locks:
355 http://lists.freedesktop.org/archives/systemd-devel/2014-October/024582.html
356
357 * as soon as we have sender timestamps, revisit coalescing multiple parallel daemon reloads:
358 http://lists.freedesktop.org/archives/systemd-devel/2014-December/025862.html
359
360 * in systemctl list-unit-files: show the install value the presets would suggest for a service in a third column
361
362 * figure out when we can use the coarse timers
363
364 * add "systemctl start -v foobar.service" that shows logs of a service
365 while the start command runs. This is non-trivial to do without
366 races though, since we should flush out all journal messages before
367 returning from the "systemctl stop".
368
369 * 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
370
371 * maybe add support for specifier expansion in user.conf, specifically DefaultEnvironment=
372
373 * introduce systemd-timesync-wait.service or so to sync on an NTP fix?
374
375 * consider showing the unit names during boot up in the status output, not just the unit descriptions
376
377 * maybe allow timer units with an empty Units= setting, so that they
378 can be used for resuming the system but nothing else.
379
380 * what to do about udev db binary stability for apps? (raw access is not an option)
381
382 * maybe provide an API to allow migration of foreign PIDs into existing scopes.
383
384 * man: maybe use the word "inspect" rather than "introspect"?
385
386 * systemctl: if some operation fails, show log output?
387
388 * systemctl edit: use equvalent of cat() to insert existing config as a comment, prepended with #.
389 Upon editor exit, lines with one # are removed, lines with two # are left with one #, etc.
390
391 * exponential backoff in timesyncd when we cannot reach a server
392
393 * timesyncd: add ugly bus calls to set NTP servers per-interface, for usage by NM
394
395 * merge ~/.local/share and ~/.local/lib into one similar /usr/lib and /usr/share....
396
397 * systemd.show_status= should probably have a mode where only failed
398 units are shown.
399
400 * add systemd.abort_on_kill or some other such flag to send SIGABRT instead of SIGKILL
401 (throughout the codebase, not only PID1)
402
403 * resolved:
404 - mDNS/DNS-SD
405 - service registration
406 - service/domain/types browsing
407 - avahi compat
408 - DNS-SD service registration from socket units
409 - resolved should optionally register additional per-interface LLMNR
410 names, so that for the container case we can establish the same name
411 (maybe "host") for referencing the server, everywhere.
412 - allow clients to request DNSSEC for a single lookup even if DNSSEC is off (?)
413 - hook up resolved with machined-based address resolution
414
415 * refcounting in sd-resolve is borked
416
417 * Add a new verb "systemctl top"
418
419 * add new gpt type for btrfs volumes
420
421 * support empty /etc boots nicely:
422 - nspawn/gpt-generator: introduce new gpt partition type for /usr
423
424 * generator that automatically discovers btrfs subvolumes, identifies their purpose based on some xattr on them.
425
426 * a way for container managers to turn off getty starting via $container_headless= or so...
427
428 * figure out a nice way how we can let the admin know what child/sibling unit causes cgroup membership for a specific unit
429
430 * For timer units: add some mechanisms so that timer units that trigger immediately on boot do not have the services
431 they run added to the initial transaction and thus confuse Type=idle.
432
433 * add bus api to query unit file's X fields.
434
435 * gpt-auto-generator:
436 - Define new partition type for encrypted swap? Support probed LUKS for encrypted swap?
437 - Make /home automount rather than mount?
438
439 * add generator that pulls in systemd-network from containers when
440 CAP_NET_ADMIN is set, more than the loopback device is defined, even
441 when it is otherwise off
442
443 * MessageQueueMessageSize= (and suchlike) should use parse_iec_size().
444
445 * implement Distribute= in socket units to allow running multiple
446 service instances processing the listening socket, and open this up
447 for ReusePort=
448
449 * socket units: support creating sockets in different namespace,
450 opening it up for JoinsNamespaceOf=. This would require to fork off
451 a tiny process that joins the namespace and creates/binds the socket
452 and passes this back to PID1 via SCM_RIGHTS. This also could be used
453 to allow Chown/chgrp on sockets without requiring NSS in PID 1.
454
455 * introduce bus call FreezeUnit(s, b), as well as "systemctl freeze
456 $UNIT" and "systemctl thaw $UNIT" as wrappers around this. The calls
457 should SIGSTOP all unit processes in a loop until all processes of
458 it are fully stopped. This can later be used for app management by
459 desktop UIs such as gnome-shell to freeze apps that are not visible
460 on screen, not unlike how job control works on the shell
461
462 * cgroups:
463 - implement per-slice CPUFairScheduling=1 switch
464 - handle jointly mounted controllers correctly
465 - introduce high-level settings for RT budget, swappiness
466 - how to reset dynamically changed unit cgroup attributes sanely?
467 - when reloading configuration, apply new cgroup configuration
468 - when recursively showing the cgroup hierarchy, optionally also show
469 the hierarchies of child processes
470
471 * transient units:
472 - add field to transient units that indicate whether systemd or somebody else saves/restores its settings, for integration with libvirt
473 - ensure scope units may be started only a single time
474
475 * Automatically configure swap partition to use for hibernation by looking for largest swap partition on the root disk?
476
477 * when we detect low battery and no AC on boot, show pretty splash and refuse boot
478
479 * libsystemd-journal, libsystemd-login, libudev: add calls to easily attach these objects to sd-event event loops
480
481 * be more careful what we export on the bus as (usec_t) 0 and (usec_t) -1
482
483 * 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
484
485 * After coming back from hibernation reset hibernation swap partition using the /dev/snapshot ioctl APIs
486
487 * If we try to find a unit via a dangling symlink, generate a clean
488 error. Currently, we just ignore it and read the unit from the search
489 path anyway.
490
491 * refuse boot if /usr/lib/os-release is missing or /etc/machine-id cannot be set up
492
493 * man: the documentation of Restart= currently is very misleading and suggests the tools from ExecStartPre= might get restarted.
494
495 * load .d/*.conf dropins for device units
496
497 * allow implementation of InaccessibleDirectories=/ plus
498 ReadOnlyDirectories=... for whitelisting files for a service.
499
500 * sd-bus:
501 - EBADSLT handling
502 - GetAllProperties() on a non-existing object does not result in a failure currently
503 - port to sd-resolve for connecting to TCP dbus servers
504 - 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
505 - see if we can drop more message validation on the sending side
506 - add API to clone sd_bus_message objects
507 - longer term: priority inheritance
508 - dbus spec updates:
509 - NameLost/NameAcquired obsolete
510 - GVariant
511 - path escaping
512 - update systemd.special(7) to mention that dbus.socket is only about the compatibility socket now
513
514 * sd-event
515 - allow multiple signal handlers per signal?
516 - document chaining of signal handler for SIGCHLD and child handlers
517 - define more intervals where we will shift wakeup intervals around in, 1h, 6h, 24h, ...
518 - generate a failure of a default event loop is executed out-of-thread
519 - maybe add support for inotify events
520
521 * investigate endianness issues of UUID vs. GUID
522
523 * dbus: when a unit failed to load (i.e. is in UNIT_ERROR state), we
524 should be able to safely try another attempt when the bus call LoadUnit() is invoked.
525
526 * add a pam module that passes the hdd passphrase into the PAM stack and then expires it, for usage by gdm auto-login.
527
528 * add a pam module that on password changes updates any LUKS slot where the password matches
529
530 * maybe add a generator that looks for "systemd.run=" on the kernel cmdline for container usercases...
531
532 * test/:
533 - add 'set -e' to scripts in test/
534 - make stuff in test/ work with separate output dir
535
536 * seems that when we follow symlinks to units we prefer the symlink
537 destination path over /etc and /usr. We should not do that. Instead
538 /etc should always override /run+/usr and also any symlink
539 destination.
540
541 * when isolating, try to figure out a way how we implicitly can order
542 all units we stop before the isolating unit...
543
544 * teach ConditionKernelCommandLine= globs or regexes (in order to match foobar={no,0,off})
545
546 * BootLoaderSpec: Clarify that the kernel has to be in $BOOT. Clarify
547 that the boot loader should be installed to the ESP. Define a way
548 how an installer can figure out whether a BLS compliant boot loader
549 is installed.
550
551 * think about requeuing jobs when daemon-reload is issued? usecase:
552 the initrd issues a reload after fstab from the host is accessible
553 and we might want to requeue the mounts local-fs acquired through
554 that automatically.
555
556 * systemd-inhibit: make taking delay locks useful: support sending SIGINT or SIGTERM on PrepareForSleep()
557
558 * remove any syslog support from log.c — we probably cannot do this before split-off udev is gone for good
559
560 * shutdown logging: store to EFI var, and store to USB stick?
561
562 * merge unit_kill_common() and unit_kill_context()
563
564 * introduce ExecCondition= in services
565
566 * EFI:
567 - honor language efi variables for default language selection (if there are any?)
568 - honor timezone efi variables for default timezone selection (if there are any?)
569 - change bootctl to be backed by systemd-bootd to control temporary and persistent default boot goal plus efi variables
570
571 * maybe do not install getty@tty1.service symlink in /etc but in /usr?
572
573 * print a nicer explanation if people use variable/specifier expansion in ExecStart= for the first word
574
575 * mount: turn dependency information from /proc/self/mountinfo into dependency information between systemd units.
576
577 * logind:
578 - logind: optionally, ignore idle-hint logic for autosuspend, block suspend as long as a session is around
579 - When we update the kernel all kind of hibernation should be prohibited until shutdown/reboot
580 - logind: wakelock/opportunistic suspend support
581 - Add pretty name for seats in logind
582 - logind: allow showing logout dialog from system?
583 - we should probably handle SIGTERM/SIGINT to not leave dot files around, just in case
584 - session scopes/user unit: add RequiresMountsFor for the home directory of the user
585 - add Suspend() bus calls which take timestamps to fix double suspend issues when somebody hits suspend and closes laptop quickly.
586 - if pam_systemd is invoked by su from a process that is outside of a
587 any session we should probably just become a NOP, since that's
588 usually not a real user session but just some system code that just
589 needs setuid().
590 - logind: make the Suspend()/Hibernate() bus calls wait for the for
591 the job to be completed. before returning, so that clients can wait
592 for "systemctl suspend" to finish to know when the suspending is
593 complete.
594 - logind: when the power button is pressed short, just popup a
595 logout dialog. If it is pressed for 1s, do the usual
596 shutdown. Inspiration are Macs here.
597 - expose "Locked" property on logind sesison objects
598 - given that logind now lets PID 1 do all nasty work, we can
599 probably reduce the capability set it retains substantially.
600 (we need CAP_SYS_ADMIN for drmSetMaster(), so maybe not worth it)
601 - expose orientation sensors and tablet mode through logind
602 - maybe allow configuration of the StopTimeout for session scopes
603 - rename session scope so that it includes the UID. THat way
604 the session scope can be arranged freely in slices and we don't have
605 make assumptions about their slice anymore.
606 - follow PropertiesChanged state more closely, to deal with quick logouts and
607 relogins
608
609 * exec: when deinitializating a tty device fix the perms and group, too, not only when initializing. Set access mode/gid to 0620/tty.
610
611 * service: watchdog logic: for testing purposes allow ping, but do not require pong
612
613 * journal:
614 - consider introducing implicit _TTY= + _PPID= + _EUID= + _EGID= + _FSUID= + _FSGID= fields
615 - import and delete pstore filesystem content at startup
616 - journald: also get thread ID from client, plus thread name
617 - journal: when waiting for journal additions in the client always sleep at least 1s or so, in order to minimize wakeups
618 - add API to close/reopen/get fd for journal client fd in libsystemd-journal.
619 - fallback to /dev/log based logging in libsystemd-journal, if we cannot log natively?
620 - declare the local journal protocol stable in the wiki interface chart
621 - sd-journal: speed up sd_journal_get_data() with transparent hash table in bg
622 - journald: when dropping msgs due to ratelimit make sure to write
623 "dropped %u messages" not only when we are about to print the next
624 message that works, but alraedy after a short tiemout
625 - check if we can make journalctl by default use --follow mode inside of less if called without args?
626 - maybe add API to send pairs of iovecs via sd_journal_send
627 - journal: add a setgid "systemd-journal" utility to invoke from libsystemd-journal, which passes fds via STDOUT and does PK access
628 - journactl: support negative filtering, i.e. FOOBAR!="waldo",
629 and !FOOBAR for events without FOOBAR.
630 - journal: store timestamp of journal_file_set_offline() int he header,
631 so it is possible to display when the file was last synced.
632 - 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.
633 - journal: find a way to allow dropping history early, based on priority, other rules
634 - journal: When used on NFS, check payload hashes
635 - journald: add kernel cmdline option to disable ratelimiting for debug purposes
636 - refuse taking lower-case variable names in sd_journal_send() and friends.
637 - journald: we currently rotate only after MaxUse+MaxFilesize has been reached.
638 - journal: deal nicely with byte-by-byte copied files, especially regards header
639 - journal: sanely deal with entries which are larger than the individual file size, but where the components would fit
640 - Replace utmp, wtmp, btmp, and lastlog completely with journal
641 - journalctl: instead --after-cursor= maybe have a --cursor=XYZ+1 syntax?
642 - when a kernel driver logs in a tight loop, we should ratelimit that too.
643 - journald: optionally, log debug messages to /run but everything else to /var
644 - journald: when we drop syslog messages because the syslog socket is
645 full, make sure to write how many messages are lost as first thing
646 to syslog when it works again.
647 - change systemd-journal-flush into a service that stays around during
648 boot, and causes the journal to be moved back to /run on shutdown,
649 so that we do not keep /var busy. This needs to happen synchronously,
650 hence doing this via signals is not going to work.
651 - optionally support running journald from the command line for testing purposes in external projects
652 - journald: allow per-priority and per-service retention times when rotating/vacuuming
653 - journald: make use of uid-range.h to managed uid ranges to split
654 journals in.
655 - journalctl: add the ability to look for the most recent process of a binary. journalctl /usr/bin/X11 --pid=-1 or so...
656 - improve journalctl performance by loading journal files
657 lazily. Encode just enough information in the file name, so that we
658 do not have to open it to know that it is not interesting for us, for
659 the most common operations.
660 - journal-or-kmsg is currently broken? See reverted
661 commit 4a01181e460686d8b4a543b1dfa7f77c9e3c5ab8.
662 - man: document that corrupted journal files is nothing to act on
663 - rework journald sigbus stuff to use mutex
664 - Set RLIMIT_NPROC for systemd-journal-xyz, and all other of our
665 services that run under their own user ids, and use User= (but only
666 in a world where userns is ubiquitous since otherwise we cannot
667 invoke those daemons on the host AND in a container anymore). Also,
668 if LimitNPROC= is used without User= we should warn and refuse
669 operation.
670 - journalctl --verify: don't show files that are currently being
671 written to as FAIL, but instead show that their are being written to.
672 - add journalctl -H that talks via ssh to a remote peer and passes through
673 binary logs data
674 - add a version of --merge which also merges /var/log/journal/remote
675 - journalctl: -m should access container journals directly by enumerating
676 them via machined, and also watch containers coming and going.
677 Benefit: nspawn --ephemeral would start working nicely with the journal.
678 - assign MESSAGE_ID to log messages about failed services
679
680 * add a test if all entries in the catalog are properly formatted.
681 (Adding dashes in a catalog entry currently results in the catalog entry
682 being silently skipped. journalctl --update-catalog must warn about this,
683 and we should also have a unit test to check that all our message are OK.)
684
685 * document:
686 - document that deps in [Unit] sections ignore Alias= fields in
687 [Install] units of other units, unless those units are disabled
688 - man: clarify that time-sync.target is not only sysv compat but also useful otherwise. Same for similar targets
689 - document that service reload may be implemented as service reexec
690 - add a man page containing packaging guidelines and recommending usage of things like Documentation=, PrivateTmp=, PrivateNetwork= and ReadOnlyDirectories=/etc /usr.
691 - document systemd-journal-flush.service properly
692 - documentation: recommend to connect the timer units of a service to the service via Also= in [Install]
693 - man: document the very specific env the shutdown drop-in tools live in
694 - man: add more examples to man pages
695 - man: maybe sort directives in man pages, and take sections from --help and apply them to man too
696
697 * systemctl:
698 - add systemctl switch to dump transaction without executing it
699 - Add a verbose mode to "systemctl start" and friends that explains what is being done or not done
700 - "systemctl disable" on a static unit prints no message and does
701 nothing. "systemctl enable" does nothing, and gives a bad message
702 about it. Should fix both to print nice actionable messages.
703 - print nice message from systemctl --failed if there are no entries shown, and hook that into ExecStartPre of rescue.service/emergency.service
704 - add new command to systemctl: "systemctl system-reexec" which reexecs as many daemons as virtually possible
705 - systemctl enable: fail if target to alias into does not exist? maybe show how many units are enabled afterwards?
706 - systemctl: "Journal has been rotated since unit was started." message is misleading
707 - systemctl status output should include list of triggering units and their status
708
709 * unit install:
710 - "systemctl mask" should find all names by which a unit is accessible
711 (i.e. by scanning for symlinks to it) and link them all to /dev/null
712
713 * timer units:
714 - timer units should get the ability to trigger when:
715 o CLOCK_REALTIME makes jumps (TFD_TIMER_CANCEL_ON_SET)
716 o DST changes
717 - Modulate timer frequency based on battery state
718
719 * add libsystemd-password or so to query passwords during boot using the password agent logic
720
721 * clean up date formatting and parsing so that all absolute/relative timestamps we format can also be parsed
722
723 * on shutdown: move utmp, wall, audit logic all into PID 1 (or logind?), get rid of systemd-update-utmp-runlevel
724
725 * make repeated alt-ctrl-del presses printing a dump
726
727 * hostnamed: before returning information from /etc/machine-info.conf check the modification data and reread. Similar for localed, ...
728
729 * currently x-systemd.timeout is lost in the initrd, since crypttab is copied into dracut, but fstab is not
730
731 * nspawn:
732 - emulate /dev/kmsg using CUSE and turn off the syslog syscall
733 with seccomp. That should provide us with a useful log buffer that
734 systemd can log to during early boot, and disconnect container logs
735 from the kernel's logs.
736 - as soon as networkd has a bus interface, hook up --network-interface=,
737 --network-bridge= with networkd, to trigger netdev creation should an
738 interface be missing
739 - a nice way to boot up without machine id set, so that it is set at boot
740 automatically for supporting --ephemeral. Maybe hash the host machine id
741 together with the machine name to generate the machine id for the container
742 - fix logic always print a final newline on output.
743 https://github.com/systemd/systemd/pull/272#issuecomment-113153176
744 - should optionally support receiving WATCHDOG=1 messages from its payload
745 PID 1...
746 - optionally automatically add FORWARD rules to iptables whenever nspawn is
747 running, remove them when shut down.
748 - maybe make copying of /etc/resolv.conf optional, and skip it if --read-only
749 is used
750
751 * dissect
752 - refuse mounting over a mount point
753 - automatically discover .roothash files in dissect, similarly to nspawn
754
755 * machined:
756 - add an API so that libvirt-lxc can inform us about network interfaces being
757 removed or added to an existing machine
758 - "machinectl migrate" or similar to copy a container from or to a
759 difference host, via ssh
760 - introduce systemd-nspawn-ephemeral@.service, and hook it into
761 "machinectl start" with a new --ephemeral switch
762 - "machinectl status" should also show internal logs of the container in
763 question
764 - "machinectl list-images" should show os-release data, as well as
765 machine-info data (including deployment level)
766 - "machinectl history"
767 - "machinectl diff"
768 - "machinectl commit" that takes a writable snapshot of a tree, invokes a
769 shell in it, and marks it read-only after use
770
771 * importd:
772 - generate a nice warning if mkfs.btrfs is missing
773
774 * cryptsetup:
775 - cryptsetup-generator: allow specification of passwords in crypttab itself
776 - support rd.luks.allow-discards= kernel cmdline params in cryptsetup generator
777
778 * hw watchdog: optionally try to use the preset watchdog timeout instead of always overriding it
779 https://bugs.freedesktop.org/show_bug.cgi?id=54712
780
781 * create /sbin/init symlinks from the build system
782
783 * add a dependency on standard-conf.xml and other included files to man pages
784
785 * MountFlags=shared acts as MountFlags=slave right now.
786
787 * properly handle loop back mounts via fstab, especially regards to fsck/passno
788
789 * initialize the hostname from the fs label of /, if /etc/hostname does not exist?
790
791 * udev:
792 - move to LGPL
793 - kill scsi_id
794 - add trigger --subsystem-match=usb/usb_device device
795 - reimport udev db after MOVE events for devices without dev_t
796
797 * There's currently no way to cancel fsck (used to be possible via C-c or c on the console)
798
799 * add option to sockets to avoid activation. Instead just drop packets/connections, see http://cyberelk.net/tim/2012/02/15/portreserve-systemd-solution/
800
801 * coredump:
802 - save coredump in Windows/Mozilla minidump format
803 - when truncating coredumps, also log the full size that the process had, and make a metadata field so we can report truncated coredumps
804
805 * support crash reporting operation modes (https://live.gnome.org/GnomeOS/Design/Whiteboards/ProblemReporting)
806
807 * default to actual 32-bit PIDs, via /proc/sys/kernel/pid_max
808
809 * be able to specify a forced restart of service A where service B depends on, in case B
810 needs to be auto-respawned?
811
812 * tmpfiles:
813 - apply "x" on "D" too (see patch from William Douglas)
814 - replace F with f+.
815 - instead of ignoring unknown fields, reject them.
816 - creating new directories/subvolumes/fifos/device nodes
817 should not follow symlinks. None of the other adjustment or creation
818 calls follow symlinks.
819
820 * make sure systemd-ask-password-wall does not shutdown systemd-ask-password-console too early
821
822 * verify that the AF_UNIX sockets of a service in the fs still exist
823 when we start a service in order to avoid confusion when a user
824 assumes starting a service is enough to make it accessible
825
826 * Make it possible to set the keymap independently from the font on
827 the kernel cmdline. Right now setting one resets also the other.
828
829 * and a dbus call to generate target from current state
830
831 * write blog stories about:
832 - hwdb: what belongs into it, lsusb
833 - enabling dbus services
834 - how to make changes to sysctl and sysfs attributes
835 - remote access
836 - how to pass throw-away units to systemd, or dynamically change properties of existing units
837 - testing with Harald's awesome test kit
838 - auto-restart
839 - how to develop against journal browsing APIs
840 - the journal HTTP iface
841 - non-cgroup resource management
842 - dynamic resource management with cgroups
843 - refreshed, longer missions statement
844 - calendar time events
845 - init=/bin/sh vs. "emergency" mode, vs. "rescue" mode, vs. "multi-user" mode, vs. "graphical" mode, and the debug shell
846 - how to create your own target
847 - instantiated apache, dovecot and so on
848 - hooking a script into various stages of shutdown/rearly booot
849
850 * investigate whether the gnome pty helper should be moved into systemd, to provide cgroup support.
851
852 * dot output for --test showing the 'initial transaction'
853
854 * pid1:
855 - When logging about multiple units (stopping BoundTo units, conflicts, etc.),
856 log both units as UNIT=, so that journalctl -u triggers on both.
857 - generate better errors when people try to set transient properties
858 that are not supported...
859 http://lists.freedesktop.org/archives/systemd-devel/2015-February/028076.html
860 - maybe introduce WantsMountsFor=? Usecase:
861 http://lists.freedesktop.org/archives/systemd-devel/2015-January/027729.html
862 - recreate systemd's D-Bus private socket file on SIGUSR2
863 - move PAM code into its own binary
864 - when we automatically restart a service, ensure we restart its rdeps, too.
865 - hide PAM options in fragment parser when compile time disabled
866 - Support --test based on current system state
867 - 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().
868 - after deserializing sockets in socket.c we should reapply sockopts and things
869 - drop PID 1 reloading, only do reexecing (difficult: Reload()
870 currently is properly synchronous, Reexec() is weird, because we
871 cannot delay the response properly until we are back, so instead of
872 being properly synchronous we just keep open the fd and close it
873 when done. That means clients do not get a successful method reply,
874 but much rather a disconnect on success.
875 - when breaking cycles drop sysv services first, then services from /run, then from /etc, then from /usr
876 - when a bus name of a service disappears from the bus make sure to queue further activation requests
877
878 * unit files:
879 - allow port=0 in .socket units
880 - maybe introduce ExecRestartPre=
881 - add ReloadSignal= for configuring a reload signal to use
882 - implement Register= switch in .socket units to enable registration
883 in Avahi, RPC and other socket registration services.
884 - allow Type=simple with PIDFile=
885 https://bugzilla.redhat.com/show_bug.cgi?id=723942
886 - allow writing multiple conditions in unit files on one line
887 - load-fragment: when loading a unit file via a chain of symlinks
888 verify that it is not masked via any of the names traversed.
889 - introduce Type=pid-file
890 - introduce mix of BindTo and Requisite
891 - add a concept of RemainAfterExit= to scope units
892 - Allow multiple ExecStart= for all Type= settings, so that we can cover rescue.service nicely
893
894 * udev-link-config:
895 - Make sure ID_PATH is always exported and complete for
896 network devices where possible, so we can safely rely
897 on Path= matching
898 - check MTUBytes parsing (expecting size_t but we are using unsigned)
899
900 * sd-rtnl:
901 - add support for more attribute types
902 - inbuilt piping support (essentially degenerate async)? see loopback-setup.c and other places
903
904 * networkd:
905 - add more keys to [Route] and [Address] sections
906 - add support for more DHCPv4 options (and, longer term, other kinds of dynamic config)
907 - add proper initrd support (in particular generate .network/.link files based on /proc/cmdline)
908 - add reduced [Link] support to .network files
909 - add Scope= parsing option for [Network]
910 - properly handle routerless dhcp leases
911 - work with non-Ethernet devices
912 - add support for more bond options
913 - dhcp: do we allow configuring dhcp routes on interfaces that are not the one we got the dhcp info from?
914 - the DHCP lease data (such as NTP/DNS) is still made available when
915 a carrier is lost on a link. It should be removed instantly.
916 - expose in the API the following bits:
917 - option 15, domain name and/or option 119, search list
918 - option 12, host name and/or option 81, fqdn
919 - option 123, 144, geolocation
920 - option 252, configure http proxy (PAC/wpad)
921 - provide a way to define a per-network interface default metric value
922 for all routes to it. possibly a second default for DHCP routes.
923 - allow Name= to be specified repeatedly in the [Match] section. Maybe also
924 support Name=foo*|bar*|baz ?
925 - duplicate address check for static IPs (like ARPCHECK in network-scripts)
926 - allow DUID/IAID to be customized, see issue #394.
927 - whenever uplink info changes, make DHCP server send out FORCERENEW
928
929 * networkd-wait-online:
930 - make operstates to wait for configurable?
931
932 * dhcp:
933 - figure out how much we can increase Maximum Message Size
934
935 * dhcp6:
936 - add functions to set previously stored IPv6 addresses on startup and get
937 them at shutdown; store them in client->ia_na
938 - write more test cases
939 - implement reconfigure support, see 5.3., 15.11. and 22.20.
940 - implement support for temporary adressess (IA_TA)
941 - implement dhcpv6 authentication
942 - investigate the usefulness of Confirm messages; i.e. are there any
943 situations where the link changes without any loss in carrier detection
944 or interface down
945 - some servers don't do rapid commit without a filled in IA_NA, verify
946 this behavior
947 - RouteTable= ?
948
949 External:
950
951 * dbus:
952 - natively watch for dbus-*.service symlinks (PENDING)
953 - teach dbus to activate all services it finds in /etc/systemd/services/org-*.service
954
955 * fix alsa mixer restore to not print error when no config is stored
956
957 * make cryptsetup lower --iter-time
958
959 * patch kernel for xattr support in /dev, /proc/, /sys?
960
961 * kernel: add device_type = "fb", "fbcon" to class "graphics"
962
963 * drop accountsservice's StandardOutput=syslog and Type=dbus fields
964
965 * /usr/bin/service should actually show the new command line
966
967 * 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=
968
969 * fedora: F20: go timer units all the way, leave cron.daily for cron
970
971 * neither pkexec nor sudo initialize environ[] from the PAM environment?
972
973 * fedora: update policy to declare access mode and ownership of unit files to root:root 0644, and add an rpmlint check for it
974
975 * register catalog database signature as file magic
976
977 * zsh shell completion:
978 - <command> <verb> -<TAB> should complete options, but currently does not
979 - systemctl add-wants,add-requires
980
981
982 Regularly:
983
984 * look for close() vs. close_nointr() vs. close_nointr_nofail()
985
986 * check for strerror(r) instead of strerror(-r)
987
988 * pahole
989
990 * set_put(), hashmap_put() return values check. i.e. == 0 does not free()!
991
992 * use secure_getenv() instead of getenv() where appropriate
993
994 * link up selected blog stories from man pages and unit files Documentation= fields