]> git.ipfire.org Git - thirdparty/systemd.git/blob - TODO
Merge pull request #7860 from poettering/watch-pids-rework
[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 * there should be path_hash_ops and we should use it in tmpfiles' hashmap object to deal with identical but differently spelt paths
28
29 * rework pid watching logic, allow any number of units watch the same pid
30
31 * be stricter with fds we receive for the fdstore: close them asynchronously
32
33 * be stricter with pid file and notify pids: don't allow them to be outside of cgroup, except if sender is privileged
34
35 * calenderspec: add support for week numbers and day numbers within a
36 year. This would allow us to define "bi-weekly" triggers safely.
37
38 * add support for recursive bpf firewalling as supported by the newest kernel
39
40 * add bpf-based implementation of devices cgroup controller logic for compat with cgroupsv2 as supported by newest kernel
41
42 * introduce sd_id128_get_boot_app_specific() which is like
43 sd_id128_get_machine_app_specific(). After all on long-running systems both
44 IDs have similar properties.
45
46 * sd-bus: add vtable flag, that may be used to request client creds implicitly
47 and asynchronously before dispatching the operation
48
49 * implement transient socket unit.
50
51 * make systemd-run create transient path and socket unit.
52
53 * make use of ethtool veth peer info in machined, for automatically finding out
54 host-side interface pointing to the container.
55
56 * add some special mode to LogsDirectory=/StateDirectory=… that allows
57 declaring these directories without necessarily pulling in deps for them, or
58 creating them when starting up. That way, we could declare that
59 systemd-journald writes to /var/log/journal, which could be useful when we
60 doing disk usage calculations and so on.
61
62 * taint systemd if there are fewer than 65536 users assigned to the system.
63
64 * deprecate PermissionsStartOnly= and RootDirectoryStartOnly= in favour of the ExecStart= prefix chars
65
66 * add a new RuntimeDirectoryPreserve= mode that defines a similar lifecycle for
67 the runtime dir as we maintain for the fdstore: i.e. keep it around as long
68 as the unit is running or has a job queued.
69
70 * support projid-based quota in machinectl for containers, and then drop
71 implicit btrfs loopback magic in machined
72
73 * Add NetworkNamespacePath= to specify a path to a network namespace
74
75 * maybe use SOURCE_DATE_EPOCH (i.e. the env var the reproducible builds folks
76 introduced) as the RTC epoch, instead of the mtime of NEWS.
77
78 * add a way to lock down cgroup migration: a boolean, which when set for a unit
79 makes sure the processes in it can never migrate out of it
80
81 * blog about fd store and restartable services
82
83 * document Environment=SYSTEMD_LOG_LEVEL=debug drop-in in debugging document
84
85 * rework ExecOutput and ExecInput enums so that EXEC_OUTPUT_NULL loses its
86 magic meaning and is no longer upgraded to something else if set explicitly.
87
88 * in the long run: permit a system with /etc/machine-id linked to /dev/null, to
89 make it lose its identity, i.e. be anonymous. For this we'd have to patch
90 through the whole tree to make all code deal with the case where no machine
91 ID is available.
92
93 * optionally, collect cgroup resource data, and store it in per-unit RRD files,
94 suitable for processing with rrdtool. Add bus API to access this data, and
95 possibly implement a CPULoad property based on it.
96
97 * beef up pam_systemd to take unit file settings such as cgroups properties as
98 parameters
99
100 * a new "systemd-analyze security" tool outputting a checklist of security
101 features a service does and does not implement
102
103 * maybe hook of xfs/ext4 quotactl() with services? i.e. automatically manage
104 the quota of a the user indicated in User= via unit file settings, like the
105 other resource management concepts. Would mix nicely with DynamicUser=1. Or
106 alternatively, do this with projids, so that we can also cover services
107 running as root. Quota should probably cover all the special dirs such as
108 StateDirectory=, LogsDirectory=, CacheDirectory=, as well as RootDirectory= if it
109 is set, plus the whole disk space any image configured with RootImage=.
110
111 * Introduce "exit" as an EmergencyAction value, and allow to configure a
112 per-unit success/failure exit code to configure. This would be useful for
113 running commands inside of services inside of containers, which could then
114 propagate their failure state all the way up.
115
116 * In DynamicUser= mode: before selecting a UID, use disk quota APIs on relevant
117 disks to see if the UID is already in use.
118
119 * add dissect_image_warn() as a wrapper around dissect_image() that prints
120 friendly log messages for the returned errors, so that we don't have to
121 duplicate that in nspawn, systemd-dissect and PID 1.
122
123 * add "systemctl wait" or so, which does what "systemd-run --wait" does, but
124 for all units. It should be both a way to pin units into memory as well as a
125 wait to retrieve their exit data.
126
127 * maybe set a new set of env vars for services, based on RuntimeDirectory=,
128 StateDirectory=, LogsDirectory=, CacheDirectory= and ConfigurationDirectory=
129 automatically. For example, there could be $RUNTIME_DIRECTORY,
130 $STATE_DIRECTORY, $LOGS_DIRECTORY=, $CACHE_DIRECTORY and
131 $CONFIGURATION_DIRECTORY or so. This could be useful to write services that
132 can adapt to varying directories for these purposes. Special care has to be
133 taken if multiple dirs are configured. Maybe avoid setting the env vars in
134 that case?
135
136 * expose IO accounting data on the bus, show it in systemd-run --wait and log
137 about it in the resource log message
138
139 * add "systemctl purge" for flushing out configuration, state, logs, ... of a
140 unit when it is stopped
141
142 * show whether a service has out-of-date configuration in "systemctl status" by
143 using mtime data of ConfigurationDirectory=.
144
145 * replace all uses of fgets() + LINE_MAX by read_line()
146
147 * Add AddUser= setting to unit files, similar to DynamicUser=1 which however
148 creates a static, persistent user rather than a dynamic, transient user. We
149 can leverage code from sysusers.d for this.
150
151 * add some optional flag to ReadWritePaths= and friends, that has the effect
152 that we create the dir in question when the service is started. Example:
153
154 ReadWritePaths=:/var/lib/foobar
155
156 * maybe add call sd_journal_set_block_timeout() or so to set SO_SNDTIMEO for
157 the sd-journal logging socket, and, if the timeout is set to 0, sets
158 O_NONBLOCK on it. That way people can control if and when to block for
159 logging.
160
161 * tighten sd_notify() MAINPID= checks a bit: don't accept foreign PIDs (i.e.
162 PIDs not managed by the service manager)
163
164 * hostnamed: populate form factor data from a new hwdb database, so that old
165 yogas can be recognized as "convertible" too, even if they predate the DMI
166 "convertible" form factor
167
168 * Maybe add a small tool invoked early at boot, that adds in or resizes
169 partitions automatically, to be used when the media used is actually larger
170 than the image written onto it is.
171
172 * Maybe add PrivatePIDs= as new unit setting, and do minimal PID namespacing
173 after all. Be strict however, only support the equivalent of nspawn's
174 --as-pid2 switch, and sanely proxy sd_notify() messages dropping stuff such
175 as MAINPID.
176
177 * Add ExecMonitor= setting. May be used multiple times. Forks off a process in
178 the service cgroup, which is supposed to monitor the service, and when it
179 exits the service is considered failed by its monitor.
180
181 * track the per-service PAM process properly (i.e. as an additional control
182 process), so that it may be queried on the bus and everything.
183
184 * add a new "debug" job mode, that is propagated to unit_start() and for
185 services results in two things: we raise SIGSTOP right before invoking
186 execve() and turn off watchdog support. Then, use that to implement
187 "systemd-gdb" for attaching to the start-up of any system service in its
188 natural habitat.
189
190 * maybe introduce gpt auto discovery for /var/tmp?
191
192 * maybe add gpt-partition-based user management: each user gets his own
193 LUKS-encrypted GPT partition with a new GPT type. A small nss module
194 enumerates users via udev partition enumeration. UIDs are assigned in a fixed
195 way: the partition index is added as offset to some fixed base uid. User name
196 is stored in GPT partition name. A PAM module authenticates the user via the
197 LUKS partition password. Benefits: strong per-user security, compatibility
198 with stateless/read-only/verity-enabled root. (other idea: do this based on
199 loopback files in /home, without GPT involvement)
200
201 * gpt-auto logic: introduce support for discovering /var matching an image. For
202 that, use a partition type UUID that is hashed from the OS name (as encoded
203 in /etc/os-release), the architecture, and 4 new bits from the gpt flags
204 field of the root partition. This way can easily support multiple OS
205 installations on the same GPT partition table, without problems with
206 unmatched /var partitions.
207
208 * gpt-auto logic: related to the above, maybe support a "secondary" root
209 partition, that is mounted to / and is writable, and where the actual root's
210 /usr is mounted into.
211
212 * 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
213
214 * drop nss-myhostname in favour of nss-resolve?
215
216 * add a percentage syntax for TimeoutStopSec=, e.g. TimeoutStopSec=150%, and
217 then use that for the setting used in user@.service. It should be understood
218 relative to the configured default value.
219
220 * in networkd, when matching device types, fix up DEVTYPE rubbish the kernel passes to us
221
222 * enable LockMLOCK to take a percentage value relative to physical memory
223
224 * Permit masking specific netlink APIs with RestrictAddressFamily=
225
226 * nspawn: support that /proc, /sys/, /dev are pre-mounted
227
228 * define gpt header bits to select volatility mode
229
230 * ProtectKernelLogs= (drops CAP_SYSLOG, add seccomp for syslog() syscall, and DeviceAllow to /dev/kmsg) in service files
231
232 * ProtectClock= (drops CAP_SYS_TIMES, adds seecomp filters for settimeofday, adjtimex), sets DeviceAllow o /dev/rtc
233
234 * ProtectTracing= (drops CAP_SYS_PTRACE, blocks ptrace syscall, makes /sys/kernel/tracing go away)
235
236 * ProtectMount= (drop mount/umount/pivot_root from seccomp, disallow fuse via DeviceAllow, imply Mountflags=slave)
237
238 * ProtectKeyRing= to take keyring calls away
239
240 * RemoveKeyRing= to remove all keyring entries of the specified user
241
242 * ProtectReboot= that masks reboot() and kexec_load() syscalls, prohibits kill
243 on PID 1 with the relevant signals, and makes relevant files in /sys and
244 /proc (such as the sysrq stuff) unavailable
245
246 * DeviceAllow= should also generate seccomp filters for mknod()
247
248 * make sure the ratelimit object can deal with USEC_INFINITY as way to turn off things
249
250 * journalctl: make sure -f ends when the container indicated by -M terminates
251
252 * mount: automatically search for "main" partition of an image has multiple
253 partitions
254
255 * expose the "privileged" flag of ExecCommand on the bus, and open it up to
256 transient units
257
258 * in nss-systemd, if we run inside of RootDirectory= with PrivateUsers= set,
259 find a way to map the User=/Group= of the service to the right name. This way
260 a user/group for a service only has to exist on the host for the right
261 mapping to work.
262
263 * add bus API for creating unit files in /etc, reusing the code for transient units
264
265 * add bus API to remove unit files from /etc
266
267 * add bus API to retrieve current unit file contents (i.e. implement "systemctl cat" on the bus only)
268
269 * rework fopen_temporary() to make use of open_tmpfile_linkable() (problem: the
270 kernel doesn't support linkat() that replaces existing files, currently)
271
272 * check if DeviceAllow= should split first, resolve specifiers later
273
274 * transient units: don't bother with actually setting unit properties, we
275 reload the unit file anyway
276
277 * journald: sigbus API via a signal-handler safe function that people may call
278 from the SIGBUS handler
279
280 * optionally, also require WATCHDOG=1 notifications during service start-up and shutdown
281
282 * resolved: when routing queries, make sure only look for the *longest* suffix...
283
284 * delay activation of logind until somebody logs in, or when /dev/tty0 pulls it
285 in or lingering is on (so that containers don't bother with it until PAM is used). also exit-on-idle
286
287 * cache sd_event_now() result from before the first iteration...
288
289 * add systemctl stop --job-mode=triggering that follows TRIGGERED_BY deps and adds them to the same transaction
290
291 * PID1: find a way how we can reload unit file configuration for
292 specific units only, without reloading the whole of systemd
293
294 * add an explicit parser for LimitRTPRIO= that verifies
295 the specified range and generates sane error messages for incorrect
296 specifications.
297
298 * when we detect that there are waiting jobs but no running jobs, do something
299
300 * push CPUAffinity= also into the "cpuset" cgroup controller (only after the cpuset controller got ported to the unified hierarchy)
301
302 * PID 1 should send out sd_notify("WATCHDOG=1") messages (for usage in the --user mode, and when run via nspawn)
303
304 * there's probably something wrong with having user mounts below /sys,
305 as we have for debugfs. for exmaple, src/core/mount.c handles mounts
306 prefixed with /sys generally special.
307 http://lists.freedesktop.org/archives/systemd-devel/2015-June/032962.html
308
309 * fstab-generator: default to tmpfs-as-root if only usr= is specified on the kernel cmdline
310
311 * docs: bring http://www.freedesktop.org/wiki/Software/systemd/MyServiceCantGetRealtime up to date
312
313 * mounting and unmounting mount points manually with different source
314 devices will result in collected on all devices used.
315 http://lists.freedesktop.org/archives/systemd-devel/2015-April/030225.html
316
317 * add a job mode that will fail if a transaction would mean stopping
318 running units. Use this in timedated to manage the NTP service
319 state.
320 http://lists.freedesktop.org/archives/systemd-devel/2015-April/030229.html
321
322 * Maybe add support for the equivalent of "ethtool advertise" to .link files?
323 http://lists.freedesktop.org/archives/systemd-devel/2015-April/030112.html
324
325 * The udev blkid built-in should expose a property that reflects
326 whether media was sensed in USB CF/SD card readers. This should then
327 be used to control SYSTEMD_READY=1/0 so that USB card readers aren't
328 picked up by systemd unless they contain a medium. This would mirror
329 the behaviour we already have for CD drives.
330
331 * networkd/udev: implement SR_IOV configuration in .link files:
332 http://lists.freedesktop.org/archives/systemd-devel/2015-January/027451.html
333
334 * Rework systemctl's GetAll property parsing to use the generic bus_map_all_properties() API
335
336 * Port various tools to make use of verbs.[ch], where applicable: busctl,
337 coredumpctl, hostnamectl, localectl, timedatectl
338
339 * hostnamectl: show root image uuid
340
341 * sysfs set api in libudev is not const
342
343 * Find a solution for SMACK capabilities stuff:
344 http://lists.freedesktop.org/archives/systemd-devel/2014-December/026188.html
345
346 * "systemctl preset-all" should probably order the unit files it
347 operates on lexicographically before starting to work, in order to
348 ensure deterministic behaviour if two unit files conflict (like DMs
349 do, for example)
350
351 * synchronize console access with BSD locks:
352 http://lists.freedesktop.org/archives/systemd-devel/2014-October/024582.html
353
354 * as soon as we have sender timestamps, revisit coalescing multiple parallel daemon reloads:
355 http://lists.freedesktop.org/archives/systemd-devel/2014-December/025862.html
356
357 * in systemctl list-unit-files: show the install value the presets would suggest for a service in a third column
358
359 * figure out when we can use the coarse timers
360
361 * add "systemctl start -v foobar.service" that shows logs of a service
362 while the start command runs. This is non-trivial to do without
363 races though, since we should flush out all journal messages before
364 returning from the "systemctl stop".
365
366 * 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
367
368 * maybe add support for specifier expansion in user.conf, specifically DefaultEnvironment=
369
370 * introduce systemd-timesync-wait.service or so to sync on an NTP fix?
371
372 * consider showing the unit names during boot up in the status output, not just the unit descriptions
373
374 * maybe allow timer units with an empty Units= setting, so that they
375 can be used for resuming the system but nothing else.
376
377 * what to do about udev db binary stability for apps? (raw access is not an option)
378
379 * maybe provide an API to allow migration of foreign PIDs into existing scopes.
380
381 * man: maybe use the word "inspect" rather than "introspect"?
382
383 * systemctl: if some operation fails, show log output?
384
385 * systemctl edit: use equvalent of cat() to insert existing config as a comment, prepended with #.
386 Upon editor exit, lines with one # are removed, lines with two # are left with one #, etc.
387
388 * exponential backoff in timesyncd when we cannot reach a server
389
390 * timesyncd: add ugly bus calls to set NTP servers per-interface, for usage by NM
391
392 * merge ~/.local/share and ~/.local/lib into one similar /usr/lib and /usr/share....
393
394 * systemd.show_status= should probably have a mode where only failed
395 units are shown.
396
397 * add systemd.abort_on_kill or some other such flag to send SIGABRT instead of SIGKILL
398 (throughout the codebase, not only PID1)
399
400 * resolved:
401 - mDNS/DNS-SD
402 - service registration
403 - service/domain/types browsing
404 - avahi compat
405 - DNS-SD service registration from socket units
406 - resolved should optionally register additional per-interface LLMNR
407 names, so that for the container case we can establish the same name
408 (maybe "host") for referencing the server, everywhere.
409 - allow clients to request DNSSEC for a single lookup even if DNSSEC is off (?)
410 - hook up resolved with machined-based address resolution
411
412 * refcounting in sd-resolve is borked
413
414 * Add a new verb "systemctl top"
415
416 * add new gpt type for btrfs volumes
417
418 * support empty /etc boots nicely:
419 - nspawn/gpt-generator: introduce new gpt partition type for /usr
420
421 * generator that automatically discovers btrfs subvolumes, identifies their purpose based on some xattr on them.
422
423 * a way for container managers to turn off getty starting via $container_headless= or so...
424
425 * figure out a nice way how we can let the admin know what child/sibling unit causes cgroup membership for a specific unit
426
427 * For timer units: add some mechanisms so that timer units that trigger immediately on boot do not have the services
428 they run added to the initial transaction and thus confuse Type=idle.
429
430 * add bus api to query unit file's X fields.
431
432 * gpt-auto-generator:
433 - Define new partition type for encrypted swap? Support probed LUKS for encrypted swap?
434 - Make /home automount rather than mount?
435
436 * add generator that pulls in systemd-network from containers when
437 CAP_NET_ADMIN is set, more than the loopback device is defined, even
438 when it is otherwise off
439
440 * MessageQueueMessageSize= (and suchlike) should use parse_iec_size().
441
442 * implement Distribute= in socket units to allow running multiple
443 service instances processing the listening socket, and open this up
444 for ReusePort=
445
446 * socket units: support creating sockets in different namespace,
447 opening it up for JoinsNamespaceOf=. This would require to fork off
448 a tiny process that joins the namespace and creates/binds the socket
449 and passes this back to PID1 via SCM_RIGHTS. This also could be used
450 to allow Chown/chgrp on sockets without requiring NSS in PID 1.
451
452 * introduce bus call FreezeUnit(s, b), as well as "systemctl freeze
453 $UNIT" and "systemctl thaw $UNIT" as wrappers around this. The calls
454 should SIGSTOP all unit processes in a loop until all processes of
455 it are fully stopped. This can later be used for app management by
456 desktop UIs such as gnome-shell to freeze apps that are not visible
457 on screen, not unlike how job control works on the shell
458
459 * cgroups:
460 - implement per-slice CPUFairScheduling=1 switch
461 - handle jointly mounted controllers correctly
462 - introduce high-level settings for RT budget, swappiness
463 - how to reset dynamically changed unit cgroup attributes sanely?
464 - when reloading configuration, apply new cgroup configuration
465 - when recursively showing the cgroup hierarchy, optionally also show
466 the hierarchies of child processes
467
468 * transient units:
469 - add field to transient units that indicate whether systemd or somebody else saves/restores its settings, for integration with libvirt
470 - ensure scope units may be started only a single time
471
472 * Automatically configure swap partition to use for hibernation by looking for largest swap partition on the root disk?
473
474 * when we detect low battery and no AC on boot, show pretty splash and refuse boot
475
476 * libsystemd-journal, libsystemd-login, libudev: add calls to easily attach these objects to sd-event event loops
477
478 * be more careful what we export on the bus as (usec_t) 0 and (usec_t) -1
479
480 * 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
481
482 * After coming back from hibernation reset hibernation swap partition using the /dev/snapshot ioctl APIs
483
484 * If we try to find a unit via a dangling symlink, generate a clean
485 error. Currently, we just ignore it and read the unit from the search
486 path anyway.
487
488 * refuse boot if /usr/lib/os-release is missing or /etc/machine-id cannot be set up
489
490 * man: the documentation of Restart= currently is very misleading and suggests the tools from ExecStartPre= might get restarted.
491
492 * load .d/*.conf dropins for device units
493
494 * allow implementation of InaccessibleDirectories=/ plus
495 ReadOnlyDirectories=... for whitelisting files for a service.
496
497 * sd-bus:
498 - EBADSLT handling
499 - GetAllProperties() on a non-existing object does not result in a failure currently
500 - port to sd-resolve for connecting to TCP dbus servers
501 - 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
502 - see if we can drop more message validation on the sending side
503 - add API to clone sd_bus_message objects
504 - longer term: priority inheritance
505 - dbus spec updates:
506 - NameLost/NameAcquired obsolete
507 - GVariant
508 - path escaping
509 - update systemd.special(7) to mention that dbus.socket is only about the compatibility socket now
510
511 * sd-event
512 - allow multiple signal handlers per signal?
513 - document chaining of signal handler for SIGCHLD and child handlers
514 - define more intervals where we will shift wakeup intervals around in, 1h, 6h, 24h, ...
515 - generate a failure of a default event loop is executed out-of-thread
516 - maybe add support for inotify events
517
518 * investigate endianness issues of UUID vs. GUID
519
520 * dbus: when a unit failed to load (i.e. is in UNIT_ERROR state), we
521 should be able to safely try another attempt when the bus call LoadUnit() is invoked.
522
523 * add a pam module that passes the hdd passphrase into the PAM stack and then expires it, for usage by gdm auto-login.
524
525 * add a pam module that on password changes updates any LUKS slot where the password matches
526
527 * maybe add a generator that looks for "systemd.run=" on the kernel cmdline for container usercases...
528
529 * test/:
530 - add 'set -e' to scripts in test/
531 - make stuff in test/ work with separate output dir
532
533 * seems that when we follow symlinks to units we prefer the symlink
534 destination path over /etc and /usr. We should not do that. Instead
535 /etc should always override /run+/usr and also any symlink
536 destination.
537
538 * when isolating, try to figure out a way how we implicitly can order
539 all units we stop before the isolating unit...
540
541 * teach ConditionKernelCommandLine= globs or regexes (in order to match foobar={no,0,off})
542
543 * BootLoaderSpec: Clarify that the kernel has to be in $BOOT. Clarify
544 that the boot loader should be installed to the ESP. Define a way
545 how an installer can figure out whether a BLS compliant boot loader
546 is installed.
547
548 * think about requeuing jobs when daemon-reload is issued? usecase:
549 the initrd issues a reload after fstab from the host is accessible
550 and we might want to requeue the mounts local-fs acquired through
551 that automatically.
552
553 * systemd-inhibit: make taking delay locks useful: support sending SIGINT or SIGTERM on PrepareForSleep()
554
555 * remove any syslog support from log.c — we probably cannot do this before split-off udev is gone for good
556
557 * shutdown logging: store to EFI var, and store to USB stick?
558
559 * merge unit_kill_common() and unit_kill_context()
560
561 * introduce ExecCondition= in services
562
563 * EFI:
564 - honor language efi variables for default language selection (if there are any?)
565 - honor timezone efi variables for default timezone selection (if there are any?)
566 - change bootctl to be backed by systemd-bootd to control temporary and persistent default boot goal plus efi variables
567
568 * maybe do not install getty@tty1.service symlink in /etc but in /usr?
569
570 * print a nicer explanation if people use variable/specifier expansion in ExecStart= for the first word
571
572 * mount: turn dependency information from /proc/self/mountinfo into dependency information between systemd units.
573
574 * logind:
575 - logind: optionally, ignore idle-hint logic for autosuspend, block suspend as long as a session is around
576 - When we update the kernel all kind of hibernation should be prohibited until shutdown/reboot
577 - logind: wakelock/opportunistic suspend support
578 - Add pretty name for seats in logind
579 - logind: allow showing logout dialog from system?
580 - we should probably handle SIGTERM/SIGINT to not leave dot files around, just in case
581 - session scopes/user unit: add RequiresMountsFor for the home directory of the user
582 - add Suspend() bus calls which take timestamps to fix double suspend issues when somebody hits suspend and closes laptop quickly.
583 - if pam_systemd is invoked by su from a process that is outside of a
584 any session we should probably just become a NOP, since that's
585 usually not a real user session but just some system code that just
586 needs setuid().
587 - logind: make the Suspend()/Hibernate() bus calls wait for the for
588 the job to be completed. before returning, so that clients can wait
589 for "systemctl suspend" to finish to know when the suspending is
590 complete.
591 - logind: when the power button is pressed short, just popup a
592 logout dialog. If it is pressed for 1s, do the usual
593 shutdown. Inspiration are Macs here.
594 - expose "Locked" property on logind sesison objects
595 - given that logind now lets PID 1 do all nasty work, we can
596 probably reduce the capability set it retains substantially.
597 (we need CAP_SYS_ADMIN for drmSetMaster(), so maybe not worth it)
598 - expose orientation sensors and tablet mode through logind
599 - maybe allow configuration of the StopTimeout for session scopes
600 - rename session scope so that it includes the UID. THat way
601 the session scope can be arranged freely in slices and we don't have
602 make assumptions about their slice anymore.
603 - follow PropertiesChanged state more closely, to deal with quick logouts and
604 relogins
605
606 * exec: when deinitializating a tty device fix the perms and group, too, not only when initializing. Set access mode/gid to 0620/tty.
607
608 * service: watchdog logic: for testing purposes allow ping, but do not require pong
609
610 * journal:
611 - consider introducing implicit _TTY= + _PPID= + _EUID= + _EGID= + _FSUID= + _FSGID= fields
612 - import and delete pstore filesystem content at startup
613 - journald: also get thread ID from client, plus thread name
614 - journal: when waiting for journal additions in the client always sleep at least 1s or so, in order to minimize wakeups
615 - add API to close/reopen/get fd for journal client fd in libsystemd-journal.
616 - fallback to /dev/log based logging in libsystemd-journal, if we cannot log natively?
617 - declare the local journal protocol stable in the wiki interface chart
618 - sd-journal: speed up sd_journal_get_data() with transparent hash table in bg
619 - journald: when dropping msgs due to ratelimit make sure to write
620 "dropped %u messages" not only when we are about to print the next
621 message that works, but alraedy after a short tiemout
622 - check if we can make journalctl by default use --follow mode inside of less if called without args?
623 - maybe add API to send pairs of iovecs via sd_journal_send
624 - journal: add a setgid "systemd-journal" utility to invoke from libsystemd-journal, which passes fds via STDOUT and does PK access
625 - journactl: support negative filtering, i.e. FOOBAR!="waldo",
626 and !FOOBAR for events without FOOBAR.
627 - journal: store timestamp of journal_file_set_offline() int he header,
628 so it is possible to display when the file was last synced.
629 - 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.
630 - journal: find a way to allow dropping history early, based on priority, other rules
631 - journal: When used on NFS, check payload hashes
632 - journald: add kernel cmdline option to disable ratelimiting for debug purposes
633 - refuse taking lower-case variable names in sd_journal_send() and friends.
634 - journald: we currently rotate only after MaxUse+MaxFilesize has been reached.
635 - journal: deal nicely with byte-by-byte copied files, especially regards header
636 - journal: sanely deal with entries which are larger than the individual file size, but where the components would fit
637 - Replace utmp, wtmp, btmp, and lastlog completely with journal
638 - journalctl: instead --after-cursor= maybe have a --cursor=XYZ+1 syntax?
639 - when a kernel driver logs in a tight loop, we should ratelimit that too.
640 - journald: optionally, log debug messages to /run but everything else to /var
641 - journald: when we drop syslog messages because the syslog socket is
642 full, make sure to write how many messages are lost as first thing
643 to syslog when it works again.
644 - change systemd-journal-flush into a service that stays around during
645 boot, and causes the journal to be moved back to /run on shutdown,
646 so that we do not keep /var busy. This needs to happen synchronously,
647 hence doing this via signals is not going to work.
648 - optionally support running journald from the command line for testing purposes in external projects
649 - journald: allow per-priority and per-service retention times when rotating/vacuuming
650 - journald: make use of uid-range.h to managed uid ranges to split
651 journals in.
652 - journalctl: add the ability to look for the most recent process of a binary. journalctl /usr/bin/X11 --pid=-1 or so...
653 - improve journalctl performance by loading journal files
654 lazily. Encode just enough information in the file name, so that we
655 do not have to open it to know that it is not interesting for us, for
656 the most common operations.
657 - journal-or-kmsg is currently broken? See reverted
658 commit 4a01181e460686d8b4a543b1dfa7f77c9e3c5ab8.
659 - man: document that corrupted journal files is nothing to act on
660 - rework journald sigbus stuff to use mutex
661 - Set RLIMIT_NPROC for systemd-journal-xyz, and all other of our
662 services that run under their own user ids, and use User= (but only
663 in a world where userns is ubiquitous since otherwise we cannot
664 invoke those daemons on the host AND in a container anymore). Also,
665 if LimitNPROC= is used without User= we should warn and refuse
666 operation.
667 - journalctl --verify: don't show files that are currently being
668 written to as FAIL, but instead show that their are being written to.
669 - add journalctl -H that talks via ssh to a remote peer and passes through
670 binary logs data
671 - add a version of --merge which also merges /var/log/journal/remote
672 - journalctl: -m should access container journals directly by enumerating
673 them via machined, and also watch containers coming and going.
674 Benefit: nspawn --ephemeral would start working nicely with the journal.
675 - assign MESSAGE_ID to log messages about failed services
676
677 * add a test if all entries in the catalog are properly formatted.
678 (Adding dashes in a catalog entry currently results in the catalog entry
679 being silently skipped. journalctl --update-catalog must warn about this,
680 and we should also have a unit test to check that all our message are OK.)
681
682 * document:
683 - document that deps in [Unit] sections ignore Alias= fields in
684 [Install] units of other units, unless those units are disabled
685 - man: clarify that time-sync.target is not only sysv compat but also useful otherwise. Same for similar targets
686 - document that service reload may be implemented as service reexec
687 - add a man page containing packaging guidelines and recommending usage of things like Documentation=, PrivateTmp=, PrivateNetwork= and ReadOnlyDirectories=/etc /usr.
688 - document systemd-journal-flush.service properly
689 - documentation: recommend to connect the timer units of a service to the service via Also= in [Install]
690 - man: document the very specific env the shutdown drop-in tools live in
691 - man: add more examples to man pages
692 - man: maybe sort directives in man pages, and take sections from --help and apply them to man too
693
694 * systemctl:
695 - add systemctl switch to dump transaction without executing it
696 - Add a verbose mode to "systemctl start" and friends that explains what is being done or not done
697 - "systemctl disable" on a static unit prints no message and does
698 nothing. "systemctl enable" does nothing, and gives a bad message
699 about it. Should fix both to print nice actionable messages.
700 - print nice message from systemctl --failed if there are no entries shown, and hook that into ExecStartPre of rescue.service/emergency.service
701 - add new command to systemctl: "systemctl system-reexec" which reexecs as many daemons as virtually possible
702 - systemctl enable: fail if target to alias into does not exist? maybe show how many units are enabled afterwards?
703 - systemctl: "Journal has been rotated since unit was started." message is misleading
704 - systemctl status output should include list of triggering units and their status
705
706 * unit install:
707 - "systemctl mask" should find all names by which a unit is accessible
708 (i.e. by scanning for symlinks to it) and link them all to /dev/null
709
710 * timer units:
711 - timer units should get the ability to trigger when:
712 o CLOCK_REALTIME makes jumps (TFD_TIMER_CANCEL_ON_SET)
713 o DST changes
714 - Modulate timer frequency based on battery state
715
716 * add libsystemd-password or so to query passwords during boot using the password agent logic
717
718 * clean up date formatting and parsing so that all absolute/relative timestamps we format can also be parsed
719
720 * on shutdown: move utmp, wall, audit logic all into PID 1 (or logind?), get rid of systemd-update-utmp-runlevel
721
722 * make repeated alt-ctrl-del presses printing a dump
723
724 * hostnamed: before returning information from /etc/machine-info.conf check the modification data and reread. Similar for localed, ...
725
726 * currently x-systemd.timeout is lost in the initrd, since crypttab is copied into dracut, but fstab is not
727
728 * nspawn:
729 - emulate /dev/kmsg using CUSE and turn off the syslog syscall
730 with seccomp. That should provide us with a useful log buffer that
731 systemd can log to during early boot, and disconnect container logs
732 from the kernel's logs.
733 - as soon as networkd has a bus interface, hook up --network-interface=,
734 --network-bridge= with networkd, to trigger netdev creation should an
735 interface be missing
736 - a nice way to boot up without machine id set, so that it is set at boot
737 automatically for supporting --ephemeral. Maybe hash the host machine id
738 together with the machine name to generate the machine id for the container
739 - fix logic always print a final newline on output.
740 https://github.com/systemd/systemd/pull/272#issuecomment-113153176
741 - should optionally support receiving WATCHDOG=1 messages from its payload
742 PID 1...
743 - optionally automatically add FORWARD rules to iptables whenever nspawn is
744 running, remove them when shut down.
745 - maybe make copying of /etc/resolv.conf optional, and skip it if --read-only
746 is used
747
748 * dissect
749 - refuse mounting over a mount point
750 - automatically discover .roothash files in dissect, similarly to nspawn
751
752 * machined:
753 - add an API so that libvirt-lxc can inform us about network interfaces being
754 removed or added to an existing machine
755 - "machinectl migrate" or similar to copy a container from or to a
756 difference host, via ssh
757 - introduce systemd-nspawn-ephemeral@.service, and hook it into
758 "machinectl start" with a new --ephemeral switch
759 - "machinectl status" should also show internal logs of the container in
760 question
761 - "machinectl list-images" should show os-release data, as well as
762 machine-info data (including deployment level)
763 - "machinectl history"
764 - "machinectl diff"
765 - "machinectl commit" that takes a writable snapshot of a tree, invokes a
766 shell in it, and marks it read-only after use
767
768 * importd:
769 - generate a nice warning if mkfs.btrfs is missing
770
771 * cryptsetup:
772 - cryptsetup-generator: allow specification of passwords in crypttab itself
773 - support rd.luks.allow-discards= kernel cmdline params in cryptsetup generator
774
775 * hw watchdog: optionally try to use the preset watchdog timeout instead of always overriding it
776 https://bugs.freedesktop.org/show_bug.cgi?id=54712
777
778 * create /sbin/init symlinks from the build system
779
780 * add a dependency on standard-conf.xml and other included files to man pages
781
782 * MountFlags=shared acts as MountFlags=slave right now.
783
784 * properly handle loop back mounts via fstab, especially regards to fsck/passno
785
786 * initialize the hostname from the fs label of /, if /etc/hostname does not exist?
787
788 * udev:
789 - move to LGPL
790 - kill scsi_id
791 - add trigger --subsystem-match=usb/usb_device device
792 - reimport udev db after MOVE events for devices without dev_t
793
794 * There's currently no way to cancel fsck (used to be possible via C-c or c on the console)
795
796 * add option to sockets to avoid activation. Instead just drop packets/connections, see http://cyberelk.net/tim/2012/02/15/portreserve-systemd-solution/
797
798 * coredump:
799 - save coredump in Windows/Mozilla minidump format
800 - when truncating coredumps, also log the full size that the process had, and make a metadata field so we can report truncated coredumps
801
802 * support crash reporting operation modes (https://live.gnome.org/GnomeOS/Design/Whiteboards/ProblemReporting)
803
804 * default to actual 32-bit PIDs, via /proc/sys/kernel/pid_max
805
806 * be able to specify a forced restart of service A where service B depends on, in case B
807 needs to be auto-respawned?
808
809 * tmpfiles:
810 - apply "x" on "D" too (see patch from William Douglas)
811 - replace F with f+.
812 - instead of ignoring unknown fields, reject them.
813 - creating new directories/subvolumes/fifos/device nodes
814 should not follow symlinks. None of the other adjustment or creation
815 calls follow symlinks.
816
817 * make sure systemd-ask-password-wall does not shutdown systemd-ask-password-console too early
818
819 * verify that the AF_UNIX sockets of a service in the fs still exist
820 when we start a service in order to avoid confusion when a user
821 assumes starting a service is enough to make it accessible
822
823 * Make it possible to set the keymap independently from the font on
824 the kernel cmdline. Right now setting one resets also the other.
825
826 * and a dbus call to generate target from current state
827
828 * write blog stories about:
829 - hwdb: what belongs into it, lsusb
830 - enabling dbus services
831 - how to make changes to sysctl and sysfs attributes
832 - remote access
833 - how to pass throw-away units to systemd, or dynamically change properties of existing units
834 - testing with Harald's awesome test kit
835 - auto-restart
836 - how to develop against journal browsing APIs
837 - the journal HTTP iface
838 - non-cgroup resource management
839 - dynamic resource management with cgroups
840 - refreshed, longer missions statement
841 - calendar time events
842 - init=/bin/sh vs. "emergency" mode, vs. "rescue" mode, vs. "multi-user" mode, vs. "graphical" mode, and the debug shell
843 - how to create your own target
844 - instantiated apache, dovecot and so on
845 - hooking a script into various stages of shutdown/rearly booot
846
847 * investigate whether the gnome pty helper should be moved into systemd, to provide cgroup support.
848
849 * dot output for --test showing the 'initial transaction'
850
851 * pid1:
852 - When logging about multiple units (stopping BoundTo units, conflicts, etc.),
853 log both units as UNIT=, so that journalctl -u triggers on both.
854 - generate better errors when people try to set transient properties
855 that are not supported...
856 http://lists.freedesktop.org/archives/systemd-devel/2015-February/028076.html
857 - maybe introduce WantsMountsFor=? Usecase:
858 http://lists.freedesktop.org/archives/systemd-devel/2015-January/027729.html
859 - recreate systemd's D-Bus private socket file on SIGUSR2
860 - move PAM code into its own binary
861 - when we automatically restart a service, ensure we restart its rdeps, too.
862 - hide PAM options in fragment parser when compile time disabled
863 - Support --test based on current system state
864 - 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().
865 - after deserializing sockets in socket.c we should reapply sockopts and things
866 - drop PID 1 reloading, only do reexecing (difficult: Reload()
867 currently is properly synchronous, Reexec() is weird, because we
868 cannot delay the response properly until we are back, so instead of
869 being properly synchronous we just keep open the fd and close it
870 when done. That means clients do not get a successful method reply,
871 but much rather a disconnect on success.
872 - when breaking cycles drop sysv services first, then services from /run, then from /etc, then from /usr
873 - when a bus name of a service disappears from the bus make sure to queue further activation requests
874
875 * unit files:
876 - allow port=0 in .socket units
877 - maybe introduce ExecRestartPre=
878 - add ReloadSignal= for configuring a reload signal to use
879 - implement Register= switch in .socket units to enable registration
880 in Avahi, RPC and other socket registration services.
881 - allow Type=simple with PIDFile=
882 https://bugzilla.redhat.com/show_bug.cgi?id=723942
883 - allow writing multiple conditions in unit files on one line
884 - load-fragment: when loading a unit file via a chain of symlinks
885 verify that it is not masked via any of the names traversed.
886 - introduce Type=pid-file
887 - introduce mix of BindTo and Requisite
888 - add a concept of RemainAfterExit= to scope units
889 - Allow multiple ExecStart= for all Type= settings, so that we can cover rescue.service nicely
890
891 * udev-link-config:
892 - Make sure ID_PATH is always exported and complete for
893 network devices where possible, so we can safely rely
894 on Path= matching
895 - check MTUBytes parsing (expecting size_t but we are using unsigned)
896
897 * sd-rtnl:
898 - add support for more attribute types
899 - inbuilt piping support (essentially degenerate async)? see loopback-setup.c and other places
900
901 * networkd:
902 - add more keys to [Route] and [Address] sections
903 - add support for more DHCPv4 options (and, longer term, other kinds of dynamic config)
904 - add proper initrd support (in particular generate .network/.link files based on /proc/cmdline)
905 - add reduced [Link] support to .network files
906 - add Scope= parsing option for [Network]
907 - properly handle routerless dhcp leases
908 - work with non-Ethernet devices
909 - add support for more bond options
910 - dhcp: do we allow configuring dhcp routes on interfaces that are not the one we got the dhcp info from?
911 - the DHCP lease data (such as NTP/DNS) is still made available when
912 a carrier is lost on a link. It should be removed instantly.
913 - expose in the API the following bits:
914 - option 15, domain name and/or option 119, search list
915 - option 12, host name and/or option 81, fqdn
916 - option 123, 144, geolocation
917 - option 252, configure http proxy (PAC/wpad)
918 - provide a way to define a per-network interface default metric value
919 for all routes to it. possibly a second default for DHCP routes.
920 - allow Name= to be specified repeatedly in the [Match] section. Maybe also
921 support Name=foo*|bar*|baz ?
922 - duplicate address check for static IPs (like ARPCHECK in network-scripts)
923 - allow DUID/IAID to be customized, see issue #394.
924 - whenever uplink info changes, make DHCP server send out FORCERENEW
925
926 * networkd-wait-online:
927 - make operstates to wait for configurable?
928
929 * dhcp:
930 - figure out how much we can increase Maximum Message Size
931
932 * dhcp6:
933 - add functions to set previously stored IPv6 addresses on startup and get
934 them at shutdown; store them in client->ia_na
935 - write more test cases
936 - implement reconfigure support, see 5.3., 15.11. and 22.20.
937 - implement support for temporary adressess (IA_TA)
938 - implement dhcpv6 authentication
939 - investigate the usefulness of Confirm messages; i.e. are there any
940 situations where the link changes without any loss in carrier detection
941 or interface down
942 - some servers don't do rapid commit without a filled in IA_NA, verify
943 this behavior
944 - RouteTable= ?
945
946 External:
947
948 * dbus:
949 - natively watch for dbus-*.service symlinks (PENDING)
950 - teach dbus to activate all services it finds in /etc/systemd/services/org-*.service
951
952 * fix alsa mixer restore to not print error when no config is stored
953
954 * make cryptsetup lower --iter-time
955
956 * patch kernel for xattr support in /dev, /proc/, /sys?
957
958 * kernel: add device_type = "fb", "fbcon" to class "graphics"
959
960 * drop accountsservice's StandardOutput=syslog and Type=dbus fields
961
962 * /usr/bin/service should actually show the new command line
963
964 * 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=
965
966 * fedora: F20: go timer units all the way, leave cron.daily for cron
967
968 * neither pkexec nor sudo initialize environ[] from the PAM environment?
969
970 * fedora: update policy to declare access mode and ownership of unit files to root:root 0644, and add an rpmlint check for it
971
972 * register catalog database signature as file magic
973
974 * zsh shell completion:
975 - <command> <verb> -<TAB> should complete options, but currently does not
976 - systemctl add-wants,add-requires
977
978
979 Regularly:
980
981 * look for close() vs. close_nointr() vs. close_nointr_nofail()
982
983 * check for strerror(r) instead of strerror(-r)
984
985 * pahole
986
987 * set_put(), hashmap_put() return values check. i.e. == 0 does not free()!
988
989 * use secure_getenv() instead of getenv() where appropriate
990
991 * link up selected blog stories from man pages and unit files Documentation= fields