]> git.ipfire.org Git - thirdparty/systemd.git/blob - TODO
Merge pull request #7631 from cgwalters/systemctl-offline
[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 * make use of ethtool veth peer info in machined, for automatically finding out
28 host-side interface pointing to the container.
29
30 * add some special mode to LogsDirectory=/StateDirectory=… that allows
31 declaring these directories without necessarily pulling in deps for them, or
32 creating them when starting up. That way, we could declare that
33 systemd-journald writes to /var/log/journal, which could be useful when we
34 doing disk usage calculations and so on.
35
36 * taint systemd if there are fewer than 65536 users assigned to the system.
37
38 * deprecate PermissionsStartOnly= and RootDirectoryStartOnly= in favour of the ExecStart= prefix chars
39
40 * add a new RuntimeDirectoryPreserve= mode that defines a similar lifecycle for
41 the runtime dir as we maintain for the fdstore: i.e. keep it around as long
42 as the unit is running or has a job queued.
43
44 * hook up sd-bus' creds stuff with SO_PEERGROUPS
45
46 * add async version of sd_bus_add_match and make use of that
47
48 * support projid-based quota in machinectl for containers, and then drop
49 implicit btrfs loopback magic in machined
50
51 * Add NetworkNamespacePath= to specify a path to a network namespace
52
53 * maybe use SOURCE_DATE_EPOCH (i.e. the env var the reproducible builds folks
54 introduced) as the RTC epoch, instead of the mtime of NEWS.
55
56 * add a way to lock down cgroup migration: a boolean, which when set for a unit
57 makes sure the processes in it can never migrate out of it
58
59 * blog about fd store and restartable services
60
61 * document Environment=SYSTEMD_LOG_LEVEL=debug drop-in in debugging document
62
63 * rework ExecOutput and ExecInput enums so that EXEC_OUTPUT_NULL loses its
64 magic meaning and is no longer upgraded to something else if set explicitly.
65
66 * in the long run: permit a system with /etc/machine-id linked to /dev/null, to
67 make it lose its identity, i.e. be anonymous. For this we'd have to patch
68 through the whole tree to make all code deal with the case where no machine
69 ID is available.
70
71 * optionally, collect cgroup resource data, and store it in per-unit RRD files,
72 suitable for processing with rrdtool. Add bus API to access this data, and
73 possibly implement a CPULoad property based on it.
74
75 * beef up pam_systemd to take unit file settings such as cgroups properties as
76 parameters
77
78 * a new "systemd-analyze security" tool outputting a checklist of security
79 features a service does and does not implement
80
81 * maybe hook of xfs/ext4 quotactl() with services? i.e. automatically manage
82 the quota of a the user indicated in User= via unit file settings, like the
83 other resource management concepts. Would mix nicely with DynamicUser=1. Or
84 alternatively, do this with projids, so that we can also cover services
85 running as root. Quota should probably cover all the special dirs such as
86 StateDirectory=, LogsDirectory=, CacheDirectory=, as well as RootDirectory= if it
87 is set, plus the whole disk space any image configured with RootImage=.
88
89 * Introduce "exit" as an EmergencyAction value, and allow to configure a
90 per-unit success/failure exit code to configure. This would be useful for
91 running commands inside of services inside of containers, which could then
92 propagate their failure state all the way up.
93
94 * In DynamicUser= mode: before selecting a UID, use disk quota APIs on relevant
95 disks to see if the UID is already in use.
96
97 * add dissect_image_warn() as a wrapper around dissect_image() that prints
98 friendly log messages for the returned errors, so that we don't have to
99 duplicate that in nspawn, systemd-dissect and PID 1.
100
101 * add "systemctl wait" or so, which does what "systemd-run --wait" does, but
102 for all units. It should be both a way to pin units into memory as well as a
103 wait to retrieve their exit data.
104
105 * maybe set a new set of env vars for services, based on RuntimeDirectory=,
106 StateDirectory=, LogsDirectory=, CacheDirectory= and ConfigurationDirectory=
107 automatically. For example, there could be $RUNTIME_DIRECTORY,
108 $STATE_DIRECTORY, $LOGS_DIRECTORY=, $CACHE_DIRECTORY and
109 $CONFIGURATION_DIRECTORY or so. This could be useful to write services that
110 can adapt to varying directories for these purposes. Special care has to be
111 taken if multiple dirs are configured. Maybe avoid setting the env vars in
112 that case?
113
114 * expose IO accounting data on the bus, show it in systemd-run --wait and log
115 about it in the resource log message
116
117 * add "systemctl purge" for flushing out configuration, state, logs, ... of a
118 unit when it is stopped
119
120 * show whether a service has out-of-date configuration in "systemctl status" by
121 using mtime data of ConfigurationDirectory=.
122
123 * replace all uses of fgets() + LINE_MAX by read_line()
124
125 * Add AddUser= setting to unit files, similar to DynamicUser=1 which however
126 creates a static, persistent user rather than a dynamic, transient user. We
127 can leverage code from sysusers.d for this.
128
129 * add some optional flag to ReadWritePaths= and friends, that has the effect
130 that we create the dir in question when the service is started. Example:
131
132 ReadWritePaths=:/var/lib/foobar
133
134 * maybe add call sd_journal_set_block_timeout() or so to set SO_SNDTIMEO for
135 the sd-journal logging socket, and, if the timeout is set to 0, sets
136 O_NONBLOCK on it. That way people can control if and when to block for
137 logging.
138
139 * tighten sd_notify() MAINPID= checks a bit: don't accept foreign PIDs (i.e.
140 PIDs not managed by the service manager)
141
142 * hostnamed: populate form factor data from a new hwdb database, so that old
143 yogas can be recognized as "convertible" too, even if they predate the DMI
144 "convertible" form factor
145
146 * Maybe add a small tool invoked early at boot, that adds in or resizes
147 partitions automatically, to be used when the media used is actually larger
148 than the image written onto it is.
149
150 * Maybe add PrivatePIDs= as new unit setting, and do minimal PID namespacing
151 after all. Be strict however, only support the equivalent of nspawn's
152 --as-pid2 switch, and sanely proxy sd_notify() messages dropping stuff such
153 as MAINPID.
154
155 * Add ExecMonitor= setting. May be used multiple times. Forks off a process in
156 the service cgroup, which is supposed to monitor the service, and when it
157 exits the service is considered failed by its monitor.
158
159 * track the per-service PAM process properly (i.e. as an additional control
160 process), so that it may be queried on the bus and everything.
161
162 * add a new "debug" job mode, that is propagated to unit_start() and for
163 services results in two things: we raise SIGSTOP right before invoking
164 execve() and turn off watchdog support. Then, use that to implement
165 "systemd-gdb" for attaching to the start-up of any system service in its
166 natural habitat.
167
168 * replace all canonicalize_file_name() invocations by chase_symlinks(), in
169 particulr those where a rootdir is relevant.
170
171 * maybe introduce gpt auto discovery for /var/tmp?
172
173 * maybe add gpt-partition-based user management: each user gets his own
174 LUKS-encrypted GPT partition with a new GPT type. A small nss module
175 enumerates users via udev partition enumeration. UIDs are assigned in a fixed
176 way: the partition index is added as offset to some fixed base uid. User name
177 is stored in GPT partition name. A PAM module authenticates the user via the
178 LUKS partition password. Benefits: strong per-user security, compatibility
179 with stateless/read-only/verity-enabled root. (other idea: do this based on
180 loopback files in /home, without GPT involvement)
181
182 * gpt-auto logic: introduce support for discovering /var matching an image. For
183 that, use a partition type UUID that is hashed from the OS name (as encoded
184 in /etc/os-release), the architecture, and 4 new bits from the gpt flags
185 field of the root partition. This way can easily support multiple OS
186 installations on the same GPT partition table, without problems with
187 unmatched /var partitions.
188
189 * gpt-auto logic: related to the above, maybe support a "secondary" root
190 partition, that is mounted to / and is writable, and where the actual root's
191 /usr is mounted into.
192
193 * 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
194
195 * drop nss-myhostname in favour of nss-resolve?
196
197 * add a percentage syntax for TimeoutStopSec=, e.g. TimeoutStopSec=150%, and
198 then use that for the setting used in user@.service. It should be understood
199 relative to the configured default value.
200
201 * in networkd, when matching device types, fix up DEVTYPE rubbish the kernel passes to us
202
203 * enable LockMLOCK to take a percentage value relative to physical memory
204
205 * Permit masking specific netlink APIs with RestrictAddressFamily=
206
207 * nspawn: support that /proc, /sys/, /dev are pre-mounted
208
209 * define gpt header bits to select volatility mode
210
211 * ProtectKernelLogs= (drops CAP_SYSLOG, add seccomp for syslog() syscall, and DeviceAllow to /dev/kmsg) in service files
212
213 * ProtectClock= (drops CAP_SYS_TIMES, adds seecomp filters for settimeofday, adjtimex), sets DeviceAllow o /dev/rtc
214
215 * ProtectTracing= (drops CAP_SYS_PTRACE, blocks ptrace syscall, makes /sys/kernel/tracing go away)
216
217 * ProtectMount= (drop mount/umount/pivot_root from seccomp, disallow fuse via DeviceAllow, imply Mountflags=slave)
218
219 * ProtectKeyRing= to take keyring calls away
220
221 * RemoveKeyRing= to remove all keyring entries of the specified user
222
223 * ProtectReboot= that masks reboot() and kexec_load() syscalls, prohibits kill
224 on PID 1 with the relevant signals, and makes relevant files in /sys and
225 /proc (such as the sysrq stuff) unavailable
226
227 * DeviceAllow= should also generate seccomp filters for mknod()
228
229 * make sure the ratelimit object can deal with USEC_INFINITY as way to turn off things
230
231 * journalctl: make sure -f ends when the container indicated by -M terminates
232
233 * mount: automatically search for "main" partition of an image has multiple
234 partitions
235
236 * expose the "privileged" flag of ExecCommand on the bus, and open it up to
237 transient units
238
239 * in nss-systemd, if we run inside of RootDirectory= with PrivateUsers= set,
240 find a way to map the User=/Group= of the service to the right name. This way
241 a user/group for a service only has to exist on the host for the right
242 mapping to work.
243
244 * add bus API for creating unit files in /etc, reusing the code for transient units
245
246 * add bus API to remove unit files from /etc
247
248 * add bus API to retrieve current unit file contents (i.e. implement "systemctl cat" on the bus only)
249
250 * rework fopen_temporary() to make use of open_tmpfile_linkable() (problem: the
251 kernel doesn't support linkat() that replaces existing files, currently)
252
253 * check if DeviceAllow= should split first, resolve specifiers later
254
255 * transient units: don't bother with actually setting unit properties, we
256 reload the unit file anyway
257
258 * journald: sigbus API via a signal-handler safe function that people may call
259 from the SIGBUS handler
260
261 * optionally, also require WATCHDOG=1 notifications during service start-up and shutdown
262
263 * resolved: when routing queries, make sure only look for the *longest* suffix...
264
265 * delay activation of logind until somebody logs in, or when /dev/tty0 pulls it
266 in or lingering is on (so that containers don't bother with it until PAM is used). also exit-on-idle
267
268 * cache sd_event_now() result from before the first iteration...
269
270 * add systemctl stop --job-mode=triggering that follows TRIGGERED_BY deps and adds them to the same transaction
271
272 * PID1: find a way how we can reload unit file configuration for
273 specific units only, without reloading the whole of systemd
274
275 * add an explicit parser for LimitRTPRIO= that verifies
276 the specified range and generates sane error messages for incorrect
277 specifications.
278
279 * when we detect that there are waiting jobs but no running jobs, do something
280
281 * push CPUAffinity= also into the "cpuset" cgroup controller (only after the cpuset controller got ported to the unified hierarchy)
282
283 * PID 1 should send out sd_notify("WATCHDOG=1") messages (for usage in the --user mode, and when run via nspawn)
284
285 * there's probably something wrong with having user mounts below /sys,
286 as we have for debugfs. for exmaple, src/core/mount.c handles mounts
287 prefixed with /sys generally special.
288 http://lists.freedesktop.org/archives/systemd-devel/2015-June/032962.html
289
290 * fstab-generator: default to tmpfs-as-root if only usr= is specified on the kernel cmdline
291
292 * docs: bring http://www.freedesktop.org/wiki/Software/systemd/MyServiceCantGetRealtime up to date
293
294 * mounting and unmounting mount points manually with different source
295 devices will result in collected on all devices used.
296 http://lists.freedesktop.org/archives/systemd-devel/2015-April/030225.html
297
298 * add a job mode that will fail if a transaction would mean stopping
299 running units. Use this in timedated to manage the NTP service
300 state.
301 http://lists.freedesktop.org/archives/systemd-devel/2015-April/030229.html
302
303 * Maybe add support for the equivalent of "ethtool advertise" to .link files?
304 http://lists.freedesktop.org/archives/systemd-devel/2015-April/030112.html
305
306 * The udev blkid built-in should expose a property that reflects
307 whether media was sensed in USB CF/SD card readers. This should then
308 be used to control SYSTEMD_READY=1/0 so that USB card readers aren't
309 picked up by systemd unless they contain a medium. This would mirror
310 the behaviour we already have for CD drives.
311
312 * networkd/udev: implement SR_IOV configuration in .link files:
313 http://lists.freedesktop.org/archives/systemd-devel/2015-January/027451.html
314
315 * Rework systemctl's GetAll property parsing to use the generic bus_map_all_properties() API
316
317 * Port various tools to make use of verbs.[ch], where applicable: busctl,
318 coredumpctl, hostnamectl, localectl, systemd-analyze, timedatectl
319
320 * hostnamectl: show root image uuid
321
322 * sysfs set api in libudev is not const
323
324 * Find a solution for SMACK capabilities stuff:
325 http://lists.freedesktop.org/archives/systemd-devel/2014-December/026188.html
326
327 * "systemctl preset-all" should probably order the unit files it
328 operates on lexicographically before starting to work, in order to
329 ensure deterministic behaviour if two unit files conflict (like DMs
330 do, for example)
331
332 * synchronize console access with BSD locks:
333 http://lists.freedesktop.org/archives/systemd-devel/2014-October/024582.html
334
335 * as soon as we have sender timestamps, revisit coalescing multiple parallel daemon reloads:
336 http://lists.freedesktop.org/archives/systemd-devel/2014-December/025862.html
337
338 * in systemctl list-unit-files: show the install value the presets would suggest for a service in a third column
339
340 * figure out when we can use the coarse timers
341
342 * add "systemctl start -v foobar.service" that shows logs of a service
343 while the start command runs. This is non-trivial to do without
344 races though, since we should flush out all journal messages before
345 returning from the "systemctl stop".
346
347 * 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
348
349 * maybe add support for specifier expansion in user.conf, specifically DefaultEnvironment=
350
351 * introduce systemd-timesync-wait.service or so to sync on an NTP fix?
352
353 * consider showing the unit names during boot up in the status output, not just the unit descriptions
354
355 * maybe allow timer units with an empty Units= setting, so that they
356 can be used for resuming the system but nothing else.
357
358 * what to do about udev db binary stability for apps? (raw access is not an option)
359
360 * maybe provide an API to allow migration of foreign PIDs into existing scopes.
361
362 * man: maybe use the word "inspect" rather than "introspect"?
363
364 * systemctl: if some operation fails, show log output?
365
366 * systemctl edit: use equvalent of cat() to insert existing config as a comment, prepended with #.
367 Upon editor exit, lines with one # are removed, lines with two # are left with one #, etc.
368
369 * exponential backoff in timesyncd when we cannot reach a server
370
371 * timesyncd: add ugly bus calls to set NTP servers per-interface, for usage by NM
372
373 * merge ~/.local/share and ~/.local/lib into one similar /usr/lib and /usr/share....
374
375 * systemd.show_status= should probably have a mode where only failed
376 units are shown.
377
378 * add systemd.abort_on_kill or some other such flag to send SIGABRT instead of SIGKILL
379 (throughout the codebase, not only PID1)
380
381 * resolved:
382 - mDNS/DNS-SD
383 - service registration
384 - service/domain/types browsing
385 - avahi compat
386 - DNS-SD service registration from socket units
387 - resolved should optionally register additional per-interface LLMNR
388 names, so that for the container case we can establish the same name
389 (maybe "host") for referencing the server, everywhere.
390 - allow clients to request DNSSEC for a single lookup even if DNSSEC is off (?)
391 - hook up resolved with machined-based address resolution
392
393 * refcounting in sd-resolve is borked
394
395 * Add a new verb "systemctl top"
396
397 * add new gpt type for btrfs volumes
398
399 * support empty /etc boots nicely:
400 - nspawn/gpt-generator: introduce new gpt partition type for /usr
401
402 * generator that automatically discovers btrfs subvolumes, identifies their purpose based on some xattr on them.
403
404 * a way for container managers to turn off getty starting via $container_headless= or so...
405
406 * figure out a nice way how we can let the admin know what child/sibling unit causes cgroup membership for a specific unit
407
408 * For timer units: add some mechanisms so that timer units that trigger immediately on boot do not have the services
409 they run added to the initial transaction and thus confuse Type=idle.
410
411 * add bus api to query unit file's X fields.
412
413 * gpt-auto-generator:
414 - Define new partition type for encrypted swap? Support probed LUKS for encrypted swap?
415 - Make /home automount rather than mount?
416
417 * add generator that pulls in systemd-network from containers when
418 CAP_NET_ADMIN is set, more than the loopback device is defined, even
419 when it is otherwise off
420
421 * MessageQueueMessageSize= (and suchlike) should use parse_iec_size().
422
423 * implement Distribute= in socket units to allow running multiple
424 service instances processing the listening socket, and open this up
425 for ReusePort=
426
427 * socket units: support creating sockets in different namespace,
428 opening it up for JoinsNamespaceOf=. This would require to fork off
429 a tiny process that joins the namespace and creates/binds the socket
430 and passes this back to PID1 via SCM_RIGHTS. This also could be used
431 to allow Chown/chgrp on sockets without requiring NSS in PID 1.
432
433 * introduce bus call FreezeUnit(s, b), as well as "systemctl freeze
434 $UNIT" and "systemctl thaw $UNIT" as wrappers around this. The calls
435 should SIGSTOP all unit processes in a loop until all processes of
436 it are fully stopped. This can later be used for app management by
437 desktop UIs such as gnome-shell to freeze apps that are not visible
438 on screen, not unlike how job control works on the shell
439
440 * cgroups:
441 - implement per-slice CPUFairScheduling=1 switch
442 - handle jointly mounted controllers correctly
443 - introduce high-level settings for RT budget, swappiness
444 - how to reset dynamically changed unit cgroup attributes sanely?
445 - when reloading configuration, apply new cgroup configuration
446 - when recursively showing the cgroup hierarchy, optionally also show
447 the hierarchies of child processes
448
449 * transient units:
450 - add field to transient units that indicate whether systemd or somebody else saves/restores its settings, for integration with libvirt
451 - ensure scope units may be started only a single time
452
453 * Automatically configure swap partition to use for hibernation by looking for largest swap partition on the root disk?
454
455 * when we detect low battery and no AC on boot, show pretty splash and refuse boot
456
457 * libsystemd-journal, libsystemd-login, libudev: add calls to easily attach these objects to sd-event event loops
458
459 * be more careful what we export on the bus as (usec_t) 0 and (usec_t) -1
460
461 * 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
462
463 * After coming back from hibernation reset hibernation swap partition using the /dev/snapshot ioctl APIs
464
465 * If we try to find a unit via a dangling symlink, generate a clean
466 error. Currently, we just ignore it and read the unit from the search
467 path anyway.
468
469 * refuse boot if /usr/lib/os-release is missing or /etc/machine-id cannot be set up
470
471 * man: the documentation of Restart= currently is very misleading and suggests the tools from ExecStartPre= might get restarted.
472
473 * load .d/*.conf dropins for device units
474
475 * allow implementation of InaccessibleDirectories=/ plus
476 ReadOnlyDirectories=... for whitelisting files for a service.
477
478 * sd-bus:
479 - EBADSLT handling
480 - GetAllProperties() on a non-existing object does not result in a failure currently
481 - port to sd-resolve for connecting to TCP dbus servers
482 - 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
483 - see if we can drop more message validation on the sending side
484 - add API to clone sd_bus_message objects
485 - make AddMatch calls on dbus1 transports async?
486 - longer term: priority inheritance
487 - dbus spec updates:
488 - NameLost/NameAcquired obsolete
489 - GVariant
490 - path escaping
491 - update systemd.special(7) to mention that dbus.socket is only about the compatibility socket now
492 - test bloom filter generation indexes
493
494 * sd-event
495 - allow multiple signal handlers per signal?
496 - document chaining of signal handler for SIGCHLD and child handlers
497 - define more intervals where we will shift wakeup intervals around in, 1h, 6h, 24h, ...
498 - generate a failure of a default event loop is executed out-of-thread
499 - maybe add support for inotify events
500
501 * investigate endianness issues of UUID vs. GUID
502
503 * dbus: when a unit failed to load (i.e. is in UNIT_ERROR state), we
504 should be able to safely try another attempt when the bus call LoadUnit() is invoked.
505
506 * add a pam module that passes the hdd passphrase into the PAM stack and then expires it, for usage by gdm auto-login.
507
508 * add a pam module that on password changes updates any LUKS slot where the password matches
509
510 * maybe add a generator that looks for "systemd.run=" on the kernel cmdline for container usercases...
511
512 * test/:
513 - add 'set -e' to scripts in test/
514 - make stuff in test/ work with separate output dir
515
516 * seems that when we follow symlinks to units we prefer the symlink
517 destination path over /etc and /usr. We should not do that. Instead
518 /etc should always override /run+/usr and also any symlink
519 destination.
520
521 * when isolating, try to figure out a way how we implicitly can order
522 all units we stop before the isolating unit...
523
524 * teach ConditionKernelCommandLine= globs or regexes (in order to match foobar={no,0,off})
525
526 * BootLoaderSpec: Clarify that the kernel has to be in $BOOT. Clarify
527 that the boot loader should be installed to the ESP. Define a way
528 how an installer can figure out whether a BLS compliant boot loader
529 is installed.
530
531 * think about requeuing jobs when daemon-reload is issued? usecase:
532 the initrd issues a reload after fstab from the host is accessible
533 and we might want to requeue the mounts local-fs acquired through
534 that automatically.
535
536 * systemd-inhibit: make taking delay locks useful: support sending SIGINT or SIGTERM on PrepareForSleep()
537
538 * remove any syslog support from log.c — we probably cannot do this before split-off udev is gone for good
539
540 * shutdown logging: store to EFI var, and store to USB stick?
541
542 * merge unit_kill_common() and unit_kill_context()
543
544 * introduce ExecCondition= in services
545
546 * EFI:
547 - honor language efi variables for default language selection (if there are any?)
548 - honor timezone efi variables for default timezone selection (if there are any?)
549 - change bootctl to be backed by systemd-bootd to control temporary and persistent default boot goal plus efi variables
550
551 * maybe do not install getty@tty1.service symlink in /etc but in /usr?
552
553 * print a nicer explanation if people use variable/specifier expansion in ExecStart= for the first word
554
555 * mount: turn dependency information from /proc/self/mountinfo into dependency information between systemd units.
556
557 * logind:
558 - logind: optionally, ignore idle-hint logic for autosuspend, block suspend as long as a session is around
559 - When we update the kernel all kind of hibernation should be prohibited until shutdown/reboot
560 - logind: wakelock/opportunistic suspend support
561 - Add pretty name for seats in logind
562 - logind: allow showing logout dialog from system?
563 - we should probably handle SIGTERM/SIGINT to not leave dot files around, just in case
564 - session scopes/user unit: add RequiresMountsFor for the home directory of the user
565 - add Suspend() bus calls which take timestamps to fix double suspend issues when somebody hits suspend and closes laptop quickly.
566 - if pam_systemd is invoked by su from a process that is outside of a
567 any session we should probably just become a NOP, since that's
568 usually not a real user session but just some system code that just
569 needs setuid().
570 - logind: make the Suspend()/Hibernate() bus calls wait for the for
571 the job to be completed. before returning, so that clients can wait
572 for "systemctl suspend" to finish to know when the suspending is
573 complete.
574 - logind: when the power button is pressed short, just popup a
575 logout dialog. If it is pressed for 1s, do the usual
576 shutdown. Inspiration are Macs here.
577 - expose "Locked" property on logind sesison objects
578 - given that logind now lets PID 1 do all nasty work, we can
579 probably reduce the capability set it retains substantially.
580 (we need CAP_SYS_ADMIN for drmSetMaster(), so maybe not worth it)
581 - expose orientation sensors and tablet mode through logind
582 - maybe allow configuration of the StopTimeout for session scopes
583 - rename session scope so that it includes the UID. THat way
584 the session scope can be arranged freely in slices and we don't have
585 make assumptions about their slice anymore.
586 - follow PropertiesChanged state more closely, to deal with quick logouts and
587 relogins
588
589 * exec: when deinitializating a tty device fix the perms and group, too, not only when initializing. Set access mode/gid to 0620/tty.
590
591 * service: watchdog logic: for testing purposes allow ping, but do not require pong
592
593 * journal:
594 - consider introducing implicit _TTY= + _PPID= + _EUID= + _EGID= + _FSUID= + _FSGID= fields
595 - import and delete pstore filesystem content at startup
596 - journald: also get thread ID from client, plus thread name
597 - journal: when waiting for journal additions in the client always sleep at least 1s or so, in order to minimize wakeups
598 - add API to close/reopen/get fd for journal client fd in libsystemd-journal.
599 - fallback to /dev/log based logging in libsystemd-journal, if we cannot log natively?
600 - declare the local journal protocol stable in the wiki interface chart
601 - sd-journal: speed up sd_journal_get_data() with transparent hash table in bg
602 - journald: when dropping msgs due to ratelimit make sure to write
603 "dropped %u messages" not only when we are about to print the next
604 message that works, but alraedy after a short tiemout
605 - check if we can make journalctl by default use --follow mode inside of less if called without args?
606 - maybe add API to send pairs of iovecs via sd_journal_send
607 - journal: add a setgid "systemd-journal" utility to invoke from libsystemd-journal, which passes fds via STDOUT and does PK access
608 - journactl: support negative filtering, i.e. FOOBAR!="waldo",
609 and !FOOBAR for events without FOOBAR.
610 - journal: store timestamp of journal_file_set_offline() int he header,
611 so it is possible to display when the file was last synced.
612 - 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.
613 - journal: find a way to allow dropping history early, based on priority, other rules
614 - journal: When used on NFS, check payload hashes
615 - journald: add kernel cmdline option to disable ratelimiting for debug purposes
616 - refuse taking lower-case variable names in sd_journal_send() and friends.
617 - journald: we currently rotate only after MaxUse+MaxFilesize has been reached.
618 - journal: deal nicely with byte-by-byte copied files, especially regards header
619 - journal: sanely deal with entries which are larger than the individual file size, but where the components would fit
620 - Replace utmp, wtmp, btmp, and lastlog completely with journal
621 - journalctl: instead --after-cursor= maybe have a --cursor=XYZ+1 syntax?
622 - when a kernel driver logs in a tight loop, we should ratelimit that too.
623 - journald: optionally, log debug messages to /run but everything else to /var
624 - journald: when we drop syslog messages because the syslog socket is
625 full, make sure to write how many messages are lost as first thing
626 to syslog when it works again.
627 - change systemd-journal-flush into a service that stays around during
628 boot, and causes the journal to be moved back to /run on shutdown,
629 so that we do not keep /var busy. This needs to happen synchronously,
630 hence doing this via signals is not going to work.
631 - optionally support running journald from the command line for testing purposes in external projects
632 - journald: allow per-priority and per-service retention times when rotating/vacuuming
633 - journald: make use of uid-range.h to managed uid ranges to split
634 journals in.
635 - journalctl: add the ability to look for the most recent process of a binary. journalctl /usr/bin/X11 --pid=-1 or so...
636 - improve journalctl performance by loading journal files
637 lazily. Encode just enough information in the file name, so that we
638 do not have to open it to know that it is not interesting for us, for
639 the most common operations.
640 - journal-or-kmsg is currently broken? See reverted
641 commit 4a01181e460686d8b4a543b1dfa7f77c9e3c5ab8.
642 - man: document that corrupted journal files is nothing to act on
643 - rework journald sigbus stuff to use mutex
644 - Set RLIMIT_NPROC for systemd-journal-xyz, and all other of our
645 services that run under their own user ids, and use User= (but only
646 in a world where userns is ubiquitous since otherwise we cannot
647 invoke those daemons on the host AND in a container anymore). Also,
648 if LimitNPROC= is used without User= we should warn and refuse
649 operation.
650 - journalctl --verify: don't show files that are currently being
651 written to as FAIL, but instead show that their are being written to.
652 - add journalctl -H that talks via ssh to a remote peer and passes through
653 binary logs data
654 - add a version of --merge which also merges /var/log/journal/remote
655 - journalctl: -m should access container journals directly by enumerating
656 them via machined, and also watch containers coming and going.
657 Benefit: nspawn --ephemeral would start working nicely with the journal.
658 - assign MESSAGE_ID to log messages about failed services
659
660 * add a test if all entries in the catalog are properly formatted.
661 (Adding dashes in a catalog entry currently results in the catalog entry
662 being silently skipped. journalctl --update-catalog must warn about this,
663 and we should also have a unit test to check that all our message are OK.)
664
665 * document:
666 - document that deps in [Unit] sections ignore Alias= fields in
667 [Install] units of other units, unless those units are disabled
668 - man: clarify that time-sync.target is not only sysv compat but also useful otherwise. Same for similar targets
669 - document that service reload may be implemented as service reexec
670 - add a man page containing packaging guidelines and recommending usage of things like Documentation=, PrivateTmp=, PrivateNetwork= and ReadOnlyDirectories=/etc /usr.
671 - document systemd-journal-flush.service properly
672 - documentation: recommend to connect the timer units of a service to the service via Also= in [Install]
673 - man: document the very specific env the shutdown drop-in tools live in
674 - man: add more examples to man pages
675 - man: maybe sort directives in man pages, and take sections from --help and apply them to man too
676
677 * systemctl:
678 - add systemctl switch to dump transaction without executing it
679 - Add a verbose mode to "systemctl start" and friends that explains what is being done or not done
680 - "systemctl disable" on a static unit prints no message and does
681 nothing. "systemctl enable" does nothing, and gives a bad message
682 about it. Should fix both to print nice actionable messages.
683 - print nice message from systemctl --failed if there are no entries shown, and hook that into ExecStartPre of rescue.service/emergency.service
684 - add new command to systemctl: "systemctl system-reexec" which reexecs as many daemons as virtually possible
685 - systemctl enable: fail if target to alias into does not exist? maybe show how many units are enabled afterwards?
686 - systemctl: "Journal has been rotated since unit was started." message is misleading
687 - systemctl status output should include list of triggering units and their status
688
689 * unit install:
690 - "systemctl mask" should find all names by which a unit is accessible
691 (i.e. by scanning for symlinks to it) and link them all to /dev/null
692
693 * timer units:
694 - timer units should get the ability to trigger when:
695 o CLOCK_REALTIME makes jumps (TFD_TIMER_CANCEL_ON_SET)
696 o DST changes
697 - Modulate timer frequency based on battery state
698
699 * add libsystemd-password or so to query passwords during boot using the password agent logic
700
701 * clean up date formatting and parsing so that all absolute/relative timestamps we format can also be parsed
702
703 * on shutdown: move utmp, wall, audit logic all into PID 1 (or logind?), get rid of systemd-update-utmp-runlevel
704
705 * make repeated alt-ctrl-del presses printing a dump
706
707 * hostnamed: before returning information from /etc/machine-info.conf check the modification data and reread. Similar for localed, ...
708
709 * currently x-systemd.timeout is lost in the initrd, since crypttab is copied into dracut, but fstab is not
710
711 * nspawn:
712 - emulate /dev/kmsg using CUSE and turn off the syslog syscall
713 with seccomp. That should provide us with a useful log buffer that
714 systemd can log to during early boot, and disconnect container logs
715 from the kernel's logs.
716 - as soon as networkd has a bus interface, hook up --network-interface=,
717 --network-bridge= with networkd, to trigger netdev creation should an
718 interface be missing
719 - a nice way to boot up without machine id set, so that it is set at boot
720 automatically for supporting --ephemeral. Maybe hash the host machine id
721 together with the machine name to generate the machine id for the container
722 - fix logic always print a final newline on output.
723 https://github.com/systemd/systemd/pull/272#issuecomment-113153176
724 - should optionally support receiving WATCHDOG=1 messages from its payload
725 PID 1...
726 - optionally automatically add FORWARD rules to iptables whenever nspawn is
727 running, remove them when shut down.
728 - maybe make copying of /etc/resolv.conf optional, and skip it if --read-only
729 is used
730
731 * dissect
732 - refuse mounting over a mount point
733 - automatically discover .roothash files in dissect, similarly to nspawn
734
735 * machined:
736 - add an API so that libvirt-lxc can inform us about network interfaces being
737 removed or added to an existing machine
738 - "machinectl migrate" or similar to copy a container from or to a
739 difference host, via ssh
740 - introduce systemd-nspawn-ephemeral@.service, and hook it into
741 "machinectl start" with a new --ephemeral switch
742 - "machinectl status" should also show internal logs of the container in
743 question
744 - "machinectl list-images" should show os-release data, as well as
745 machine-info data (including deployment level)
746 - "machinectl history"
747 - "machinectl diff"
748 - "machinectl commit" that takes a writable snapshot of a tree, invokes a
749 shell in it, and marks it read-only after use
750
751 * importd:
752 - generate a nice warning if mkfs.btrfs is missing
753
754 * cryptsetup:
755 - cryptsetup-generator: allow specification of passwords in crypttab itself
756 - support rd.luks.allow-discards= kernel cmdline params in cryptsetup generator
757
758 * hw watchdog: optionally try to use the preset watchdog timeout instead of always overriding it
759 https://bugs.freedesktop.org/show_bug.cgi?id=54712
760
761 * create /sbin/init symlinks from the build system
762
763 * add a dependency on standard-conf.xml and other included files to man pages
764
765 * MountFlags=shared acts as MountFlags=slave right now.
766
767 * properly handle loop back mounts via fstab, especially regards to fsck/passno
768
769 * initialize the hostname from the fs label of /, if /etc/hostname does not exist?
770
771 * udev:
772 - move to LGPL
773 - kill scsi_id
774 - add trigger --subsystem-match=usb/usb_device device
775 - reimport udev db after MOVE events for devices without dev_t
776
777 * There's currently no way to cancel fsck (used to be possible via C-c or c on the console)
778
779 * add option to sockets to avoid activation. Instead just drop packets/connections, see http://cyberelk.net/tim/2012/02/15/portreserve-systemd-solution/
780
781 * coredump:
782 - save coredump in Windows/Mozilla minidump format
783 - when truncating coredumps, also log the full size that the process had, and make a metadata field so we can report truncated coredumps
784
785 * support crash reporting operation modes (https://live.gnome.org/GnomeOS/Design/Whiteboards/ProblemReporting)
786
787 * default to actual 32-bit PIDs, via /proc/sys/kernel/pid_max
788
789 * be able to specify a forced restart of service A where service B depends on, in case B
790 needs to be auto-respawned?
791
792 * tmpfiles:
793 - apply "x" on "D" too (see patch from William Douglas)
794 - replace F with f+.
795 - instead of ignoring unknown fields, reject them.
796 - creating new directories/subvolumes/fifos/device nodes
797 should not follow symlinks. None of the other adjustment or creation
798 calls follow symlinks.
799
800 * make sure systemd-ask-password-wall does not shutdown systemd-ask-password-console too early
801
802 * verify that the AF_UNIX sockets of a service in the fs still exist
803 when we start a service in order to avoid confusion when a user
804 assumes starting a service is enough to make it accessible
805
806 * Make it possible to set the keymap independently from the font on
807 the kernel cmdline. Right now setting one resets also the other.
808
809 * and a dbus call to generate target from current state
810
811 * write blog stories about:
812 - hwdb: what belongs into it, lsusb
813 - enabling dbus services
814 - how to make changes to sysctl and sysfs attributes
815 - remote access
816 - how to pass throw-away units to systemd, or dynamically change properties of existing units
817 - testing with Harald's awesome test kit
818 - auto-restart
819 - how to develop against journal browsing APIs
820 - the journal HTTP iface
821 - non-cgroup resource management
822 - dynamic resource management with cgroups
823 - refreshed, longer missions statement
824 - calendar time events
825 - init=/bin/sh vs. "emergency" mode, vs. "rescue" mode, vs. "multi-user" mode, vs. "graphical" mode, and the debug shell
826 - how to create your own target
827 - instantiated apache, dovecot and so on
828 - hooking a script into various stages of shutdown/rearly booot
829
830 * investigate whether the gnome pty helper should be moved into systemd, to provide cgroup support.
831
832 * dot output for --test showing the 'initial transaction'
833
834 * pid1:
835 - When logging about multiple units (stopping BoundTo units, conflicts, etc.),
836 log both units as UNIT=, so that journalctl -u triggers on both.
837 - generate better errors when people try to set transient properties
838 that are not supported...
839 http://lists.freedesktop.org/archives/systemd-devel/2015-February/028076.html
840 - maybe introduce WantsMountsFor=? Usecase:
841 http://lists.freedesktop.org/archives/systemd-devel/2015-January/027729.html
842 - recreate systemd's D-Bus private socket file on SIGUSR2
843 - move PAM code into its own binary
844 - when we automatically restart a service, ensure we restart its rdeps, too.
845 - hide PAM options in fragment parser when compile time disabled
846 - Support --test based on current system state
847 - 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().
848 - after deserializing sockets in socket.c we should reapply sockopts and things
849 - drop PID 1 reloading, only do reexecing (difficult: Reload()
850 currently is properly synchronous, Reexec() is weird, because we
851 cannot delay the response properly until we are back, so instead of
852 being properly synchronous we just keep open the fd and close it
853 when done. That means clients do not get a successful method reply,
854 but much rather a disconnect on success.
855 - when breaking cycles drop sysv services first, then services from /run, then from /etc, then from /usr
856 - when a bus name of a service disappears from the bus make sure to queue further activation requests
857
858 * unit files:
859 - allow port=0 in .socket units
860 - maybe introduce ExecRestartPre=
861 - add ReloadSignal= for configuring a reload signal to use
862 - implement Register= switch in .socket units to enable registration
863 in Avahi, RPC and other socket registration services.
864 - allow Type=simple with PIDFile=
865 https://bugzilla.redhat.com/show_bug.cgi?id=723942
866 - allow writing multiple conditions in unit files on one line
867 - load-fragment: when loading a unit file via a chain of symlinks
868 verify that it is not masked via any of the names traversed.
869 - introduce Type=pid-file
870 - introduce mix of BindTo and Requisite
871 - add a concept of RemainAfterExit= to scope units
872 - Allow multiple ExecStart= for all Type= settings, so that we can cover rescue.service nicely
873
874 * udev-link-config:
875 - Make sure ID_PATH is always exported and complete for
876 network devices where possible, so we can safely rely
877 on Path= matching
878 - check MTUBytes parsing (expecting size_t but we are using unsigned)
879
880 * sd-rtnl:
881 - add support for more attribute types
882 - inbuilt piping support (essentially degenerate async)? see loopback-setup.c and other places
883
884 * networkd:
885 - add more keys to [Route] and [Address] sections
886 - add support for more DHCPv4 options (and, longer term, other kinds of dynamic config)
887 - add proper initrd support (in particular generate .network/.link files based on /proc/cmdline)
888 - add reduced [Link] support to .network files
889 - add Scope= parsing option for [Network]
890 - properly handle routerless dhcp leases
891 - work with non-Ethernet devices
892 - add support for more bond options
893 - dhcp: do we allow configuring dhcp routes on interfaces that are not the one we got the dhcp info from?
894 - the DHCP lease data (such as NTP/DNS) is still made available when
895 a carrier is lost on a link. It should be removed instantly.
896 - expose in the API the following bits:
897 - option 15, domain name and/or option 119, search list
898 - option 12, host name and/or option 81, fqdn
899 - option 123, 144, geolocation
900 - option 252, configure http proxy (PAC/wpad)
901 - provide a way to define a per-network interface default metric value
902 for all routes to it. possibly a second default for DHCP routes.
903 - allow Name= to be specified repeatedly in the [Match] section. Maybe also
904 support Name=foo*|bar*|baz ?
905 - duplicate address check for static IPs (like ARPCHECK in network-scripts)
906 - allow DUID/IAID to be customized, see issue #394.
907 - whenever uplink info changes, make DHCP server send out FORCERENEW
908
909 * networkd-wait-online:
910 - make operstates to wait for configurable?
911
912 * dhcp:
913 - figure out how much we can increase Maximum Message Size
914
915 * dhcp6:
916 - add functions to set previously stored IPv6 addresses on startup and get
917 them at shutdown; store them in client->ia_na
918 - write more test cases
919 - implement reconfigure support, see 5.3., 15.11. and 22.20.
920 - implement support for temporary adressess (IA_TA)
921 - implement dhcpv6 authentication
922 - investigate the usefulness of Confirm messages; i.e. are there any
923 situations where the link changes without any loss in carrier detection
924 or interface down
925 - some servers don't do rapid commit without a filled in IA_NA, verify
926 this behavior
927 - RouteTable= ?
928
929 External:
930
931 * dbus:
932 - natively watch for dbus-*.service symlinks (PENDING)
933 - teach dbus to activate all services it finds in /etc/systemd/services/org-*.service
934
935 * fix alsa mixer restore to not print error when no config is stored
936
937 * make cryptsetup lower --iter-time
938
939 * patch kernel for xattr support in /dev, /proc/, /sys?
940
941 * kernel: add device_type = "fb", "fbcon" to class "graphics"
942
943 * drop accountsservice's StandardOutput=syslog and Type=dbus fields
944
945 * /usr/bin/service should actually show the new command line
946
947 * 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=
948
949 * fedora: F20: go timer units all the way, leave cron.daily for cron
950
951 * neither pkexec nor sudo initialize environ[] from the PAM environment?
952
953 * fedora: update policy to declare access mode and ownership of unit files to root:root 0644, and add an rpmlint check for it
954
955 * register catalog database signature as file magic
956
957 * zsh shell completion:
958 - <command> <verb> -<TAB> should complete options, but currently does not
959 - systemctl add-wants,add-requires
960
961
962 Regularly:
963
964 * look for close() vs. close_nointr() vs. close_nointr_nofail()
965
966 * check for strerror(r) instead of strerror(-r)
967
968 * pahole
969
970 * set_put(), hashmap_put() return values check. i.e. == 0 does not free()!
971
972 * use secure_getenv() instead of getenv() where appropriate
973
974 * link up selected blog stories from man pages and unit files Documentation= fields