]> git.ipfire.org Git - thirdparty/systemd.git/blob - TODO
Merge pull request #7991 from poettering/n-on-console
[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 * emulate properties of the root cgroup on controllers that don't support such
47 properties natively on cpu/io/memory, the way we already do it for
48 "pids". Also, add the same logic to cgtop.
49
50 * set TasksAccounting=1 on the root slice if we are running on the root cgroup,
51 and similar for the others, as soon as we emulate them properly. After all,
52 Linux keeps these system-wide stats anyway, and it costs nothing to expose
53 them.
54
55 * sd-bus: add vtable flag, that may be used to request client creds implicitly
56 and asynchronously before dispatching the operation
57
58 * implement transient socket unit.
59
60 * make systemd-run create transient path and socket unit.
61
62 * make use of ethtool veth peer info in machined, for automatically finding out
63 host-side interface pointing to the container.
64
65 * add some special mode to LogsDirectory=/StateDirectory=… that allows
66 declaring these directories without necessarily pulling in deps for them, or
67 creating them when starting up. That way, we could declare that
68 systemd-journald writes to /var/log/journal, which could be useful when we
69 doing disk usage calculations and so on.
70
71 * taint systemd if there are fewer than 65536 users assigned to the system.
72
73 * deprecate PermissionsStartOnly= and RootDirectoryStartOnly= in favour of the ExecStart= prefix chars
74
75 * add a new RuntimeDirectoryPreserve= mode that defines a similar lifecycle for
76 the runtime dir as we maintain for the fdstore: i.e. keep it around as long
77 as the unit is running or has a job queued.
78
79 * support projid-based quota in machinectl for containers, and then drop
80 implicit btrfs loopback magic in machined
81
82 * Add NetworkNamespacePath= to specify a path to a network namespace
83
84 * maybe use SOURCE_DATE_EPOCH (i.e. the env var the reproducible builds folks
85 introduced) as the RTC epoch, instead of the mtime of NEWS.
86
87 * add a way to lock down cgroup migration: a boolean, which when set for a unit
88 makes sure the processes in it can never migrate out of it
89
90 * blog about fd store and restartable services
91
92 * document Environment=SYSTEMD_LOG_LEVEL=debug drop-in in debugging document
93
94 * rework ExecOutput and ExecInput enums so that EXEC_OUTPUT_NULL loses its
95 magic meaning and is no longer upgraded to something else if set explicitly.
96
97 * in the long run: permit a system with /etc/machine-id linked to /dev/null, to
98 make it lose its identity, i.e. be anonymous. For this we'd have to patch
99 through the whole tree to make all code deal with the case where no machine
100 ID is available.
101
102 * optionally, collect cgroup resource data, and store it in per-unit RRD files,
103 suitable for processing with rrdtool. Add bus API to access this data, and
104 possibly implement a CPULoad property based on it.
105
106 * beef up pam_systemd to take unit file settings such as cgroups properties as
107 parameters
108
109 * a new "systemd-analyze security" tool outputting a checklist of security
110 features a service does and does not implement
111
112 * maybe hook of xfs/ext4 quotactl() with services? i.e. automatically manage
113 the quota of a the user indicated in User= via unit file settings, like the
114 other resource management concepts. Would mix nicely with DynamicUser=1. Or
115 alternatively, do this with projids, so that we can also cover services
116 running as root. Quota should probably cover all the special dirs such as
117 StateDirectory=, LogsDirectory=, CacheDirectory=, as well as RootDirectory= if it
118 is set, plus the whole disk space any image configured with RootImage=.
119
120 * Introduce "exit" as an EmergencyAction value, and allow to configure a
121 per-unit success/failure exit code to configure. This would be useful for
122 running commands inside of services inside of containers, which could then
123 propagate their failure state all the way up.
124
125 * In DynamicUser= mode: before selecting a UID, use disk quota APIs on relevant
126 disks to see if the UID is already in use.
127
128 * add dissect_image_warn() as a wrapper around dissect_image() that prints
129 friendly log messages for the returned errors, so that we don't have to
130 duplicate that in nspawn, systemd-dissect and PID 1.
131
132 * add "systemctl wait" or so, which does what "systemd-run --wait" does, but
133 for all units. It should be both a way to pin units into memory as well as a
134 wait to retrieve their exit data.
135
136 * maybe set a new set of env vars for services, based on RuntimeDirectory=,
137 StateDirectory=, LogsDirectory=, CacheDirectory= and ConfigurationDirectory=
138 automatically. For example, there could be $RUNTIME_DIRECTORY,
139 $STATE_DIRECTORY, $LOGS_DIRECTORY=, $CACHE_DIRECTORY and
140 $CONFIGURATION_DIRECTORY or so. This could be useful to write services that
141 can adapt to varying directories for these purposes. Special care has to be
142 taken if multiple dirs are configured. Maybe avoid setting the env vars in
143 that case?
144
145 * expose IO accounting data on the bus, show it in systemd-run --wait and log
146 about it in the resource log message
147
148 * add "systemctl purge" for flushing out configuration, state, logs, ... of a
149 unit when it is stopped
150
151 * show whether a service has out-of-date configuration in "systemctl status" by
152 using mtime data of ConfigurationDirectory=.
153
154 * replace all uses of fgets() + LINE_MAX by read_line()
155
156 * Add AddUser= setting to unit files, similar to DynamicUser=1 which however
157 creates a static, persistent user rather than a dynamic, transient user. We
158 can leverage code from sysusers.d for this.
159
160 * add some optional flag to ReadWritePaths= and friends, that has the effect
161 that we create the dir in question when the service is started. Example:
162
163 ReadWritePaths=:/var/lib/foobar
164
165 * maybe add call sd_journal_set_block_timeout() or so to set SO_SNDTIMEO for
166 the sd-journal logging socket, and, if the timeout is set to 0, sets
167 O_NONBLOCK on it. That way people can control if and when to block for
168 logging.
169
170 * tighten sd_notify() MAINPID= checks a bit: don't accept foreign PIDs (i.e.
171 PIDs not managed by the service manager)
172
173 * hostnamed: populate form factor data from a new hwdb database, so that old
174 yogas can be recognized as "convertible" too, even if they predate the DMI
175 "convertible" form factor
176
177 * Maybe add a small tool invoked early at boot, that adds in or resizes
178 partitions automatically, to be used when the media used is actually larger
179 than the image written onto it is.
180
181 * Maybe add PrivatePIDs= as new unit setting, and do minimal PID namespacing
182 after all. Be strict however, only support the equivalent of nspawn's
183 --as-pid2 switch, and sanely proxy sd_notify() messages dropping stuff such
184 as MAINPID.
185
186 * Add ExecMonitor= setting. May be used multiple times. Forks off a process in
187 the service cgroup, which is supposed to monitor the service, and when it
188 exits the service is considered failed by its monitor.
189
190 * track the per-service PAM process properly (i.e. as an additional control
191 process), so that it may be queried on the bus and everything.
192
193 * add a new "debug" job mode, that is propagated to unit_start() and for
194 services results in two things: we raise SIGSTOP right before invoking
195 execve() and turn off watchdog support. Then, use that to implement
196 "systemd-gdb" for attaching to the start-up of any system service in its
197 natural habitat.
198
199 * maybe introduce gpt auto discovery for /var/tmp?
200
201 * maybe add gpt-partition-based user management: each user gets his own
202 LUKS-encrypted GPT partition with a new GPT type. A small nss module
203 enumerates users via udev partition enumeration. UIDs are assigned in a fixed
204 way: the partition index is added as offset to some fixed base uid. User name
205 is stored in GPT partition name. A PAM module authenticates the user via the
206 LUKS partition password. Benefits: strong per-user security, compatibility
207 with stateless/read-only/verity-enabled root. (other idea: do this based on
208 loopback files in /home, without GPT involvement)
209
210 * gpt-auto logic: introduce support for discovering /var matching an image. For
211 that, use a partition type UUID that is hashed from the OS name (as encoded
212 in /etc/os-release), the architecture, and 4 new bits from the gpt flags
213 field of the root partition. This way can easily support multiple OS
214 installations on the same GPT partition table, without problems with
215 unmatched /var partitions.
216
217 * gpt-auto logic: related to the above, maybe support a "secondary" root
218 partition, that is mounted to / and is writable, and where the actual root's
219 /usr is mounted into.
220
221 * 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
222
223 * drop nss-myhostname in favour of nss-resolve?
224
225 * add a percentage syntax for TimeoutStopSec=, e.g. TimeoutStopSec=150%, and
226 then use that for the setting used in user@.service. It should be understood
227 relative to the configured default value.
228
229 * in networkd, when matching device types, fix up DEVTYPE rubbish the kernel passes to us
230
231 * enable LockMLOCK to take a percentage value relative to physical memory
232
233 * Permit masking specific netlink APIs with RestrictAddressFamily=
234
235 * nspawn: support that /proc, /sys/, /dev are pre-mounted
236
237 * define gpt header bits to select volatility mode
238
239 * ProtectKernelLogs= (drops CAP_SYSLOG, add seccomp for syslog() syscall, and DeviceAllow to /dev/kmsg) in service files
240
241 * ProtectClock= (drops CAP_SYS_TIMES, adds seecomp filters for settimeofday, adjtimex), sets DeviceAllow o /dev/rtc
242
243 * ProtectTracing= (drops CAP_SYS_PTRACE, blocks ptrace syscall, makes /sys/kernel/tracing go away)
244
245 * ProtectMount= (drop mount/umount/pivot_root from seccomp, disallow fuse via DeviceAllow, imply Mountflags=slave)
246
247 * ProtectKeyRing= to take keyring calls away
248
249 * RemoveKeyRing= to remove all keyring entries of the specified user
250
251 * ProtectReboot= that masks reboot() and kexec_load() syscalls, prohibits kill
252 on PID 1 with the relevant signals, and makes relevant files in /sys and
253 /proc (such as the sysrq stuff) unavailable
254
255 * DeviceAllow= should also generate seccomp filters for mknod()
256
257 * make sure the ratelimit object can deal with USEC_INFINITY as way to turn off things
258
259 * journalctl: make sure -f ends when the container indicated by -M terminates
260
261 * mount: automatically search for "main" partition of an image has multiple
262 partitions
263
264 * expose the "privileged" flag of ExecCommand on the bus, and open it up to
265 transient units
266
267 * in nss-systemd, if we run inside of RootDirectory= with PrivateUsers= set,
268 find a way to map the User=/Group= of the service to the right name. This way
269 a user/group for a service only has to exist on the host for the right
270 mapping to work.
271
272 * add bus API for creating unit files in /etc, reusing the code for transient units
273
274 * add bus API to remove unit files from /etc
275
276 * add bus API to retrieve current unit file contents (i.e. implement "systemctl cat" on the bus only)
277
278 * rework fopen_temporary() to make use of open_tmpfile_linkable() (problem: the
279 kernel doesn't support linkat() that replaces existing files, currently)
280
281 * check if DeviceAllow= should split first, resolve specifiers later
282
283 * transient units: don't bother with actually setting unit properties, we
284 reload the unit file anyway
285
286 * journald: sigbus API via a signal-handler safe function that people may call
287 from the SIGBUS handler
288
289 * optionally, also require WATCHDOG=1 notifications during service start-up and shutdown
290
291 * resolved: when routing queries, make sure only look for the *longest* suffix...
292
293 * delay activation of logind until somebody logs in, or when /dev/tty0 pulls it
294 in or lingering is on (so that containers don't bother with it until PAM is used). also exit-on-idle
295
296 * cache sd_event_now() result from before the first iteration...
297
298 * add systemctl stop --job-mode=triggering that follows TRIGGERED_BY deps and adds them to the same transaction
299
300 * PID1: find a way how we can reload unit file configuration for
301 specific units only, without reloading the whole of systemd
302
303 * add an explicit parser for LimitRTPRIO= that verifies
304 the specified range and generates sane error messages for incorrect
305 specifications.
306
307 * when we detect that there are waiting jobs but no running jobs, do something
308
309 * push CPUAffinity= also into the "cpuset" cgroup controller (only after the cpuset controller got ported to the unified hierarchy)
310
311 * PID 1 should send out sd_notify("WATCHDOG=1") messages (for usage in the --user mode, and when run via nspawn)
312
313 * there's probably something wrong with having user mounts below /sys,
314 as we have for debugfs. for exmaple, src/core/mount.c handles mounts
315 prefixed with /sys generally special.
316 http://lists.freedesktop.org/archives/systemd-devel/2015-June/032962.html
317
318 * fstab-generator: default to tmpfs-as-root if only usr= is specified on the kernel cmdline
319
320 * docs: bring http://www.freedesktop.org/wiki/Software/systemd/MyServiceCantGetRealtime up to date
321
322 * mounting and unmounting mount points manually with different source
323 devices will result in collected on all devices used.
324 http://lists.freedesktop.org/archives/systemd-devel/2015-April/030225.html
325
326 * add a job mode that will fail if a transaction would mean stopping
327 running units. Use this in timedated to manage the NTP service
328 state.
329 http://lists.freedesktop.org/archives/systemd-devel/2015-April/030229.html
330
331 * Maybe add support for the equivalent of "ethtool advertise" to .link files?
332 http://lists.freedesktop.org/archives/systemd-devel/2015-April/030112.html
333
334 * The udev blkid built-in should expose a property that reflects
335 whether media was sensed in USB CF/SD card readers. This should then
336 be used to control SYSTEMD_READY=1/0 so that USB card readers aren't
337 picked up by systemd unless they contain a medium. This would mirror
338 the behaviour we already have for CD drives.
339
340 * networkd/udev: implement SR_IOV configuration in .link files:
341 http://lists.freedesktop.org/archives/systemd-devel/2015-January/027451.html
342
343 * Rework systemctl's GetAll property parsing to use the generic bus_map_all_properties() API
344
345 * Port various tools to make use of verbs.[ch], where applicable: busctl,
346 coredumpctl, hostnamectl, localectl, timedatectl
347
348 * hostnamectl: show root image uuid
349
350 * sysfs set api in libudev is not const
351
352 * Find a solution for SMACK capabilities stuff:
353 http://lists.freedesktop.org/archives/systemd-devel/2014-December/026188.html
354
355 * "systemctl preset-all" should probably order the unit files it
356 operates on lexicographically before starting to work, in order to
357 ensure deterministic behaviour if two unit files conflict (like DMs
358 do, for example)
359
360 * synchronize console access with BSD locks:
361 http://lists.freedesktop.org/archives/systemd-devel/2014-October/024582.html
362
363 * as soon as we have sender timestamps, revisit coalescing multiple parallel daemon reloads:
364 http://lists.freedesktop.org/archives/systemd-devel/2014-December/025862.html
365
366 * in systemctl list-unit-files: show the install value the presets would suggest for a service in a third column
367
368 * figure out when we can use the coarse timers
369
370 * add "systemctl start -v foobar.service" that shows logs of a service
371 while the start command runs. This is non-trivial to do without
372 races though, since we should flush out all journal messages before
373 returning from the "systemctl stop".
374
375 * 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
376
377 * maybe add support for specifier expansion in user.conf, specifically DefaultEnvironment=
378
379 * introduce systemd-timesync-wait.service or so to sync on an NTP fix?
380
381 * consider showing the unit names during boot up in the status output, not just the unit descriptions
382
383 * maybe allow timer units with an empty Units= setting, so that they
384 can be used for resuming the system but nothing else.
385
386 * what to do about udev db binary stability for apps? (raw access is not an option)
387
388 * maybe provide an API to allow migration of foreign PIDs into existing scopes.
389
390 * man: maybe use the word "inspect" rather than "introspect"?
391
392 * systemctl: if some operation fails, show log output?
393
394 * systemctl edit: use equvalent of cat() to insert existing config as a comment, prepended with #.
395 Upon editor exit, lines with one # are removed, lines with two # are left with one #, etc.
396
397 * exponential backoff in timesyncd when we cannot reach a server
398
399 * timesyncd: add ugly bus calls to set NTP servers per-interface, for usage by NM
400
401 * merge ~/.local/share and ~/.local/lib into one similar /usr/lib and /usr/share....
402
403 * systemd.show_status= should probably have a mode where only failed
404 units are shown.
405
406 * add systemd.abort_on_kill or some other such flag to send SIGABRT instead of SIGKILL
407 (throughout the codebase, not only PID1)
408
409 * resolved:
410 - mDNS/DNS-SD
411 - service registration
412 - service/domain/types browsing
413 - avahi compat
414 - DNS-SD service registration from socket units
415 - resolved should optionally register additional per-interface LLMNR
416 names, so that for the container case we can establish the same name
417 (maybe "host") for referencing the server, everywhere.
418 - allow clients to request DNSSEC for a single lookup even if DNSSEC is off (?)
419 - hook up resolved with machined-based address resolution
420
421 * refcounting in sd-resolve is borked
422
423 * Add a new verb "systemctl top"
424
425 * add new gpt type for btrfs volumes
426
427 * support empty /etc boots nicely:
428 - nspawn/gpt-generator: introduce new gpt partition type for /usr
429
430 * generator that automatically discovers btrfs subvolumes, identifies their purpose based on some xattr on them.
431
432 * a way for container managers to turn off getty starting via $container_headless= or so...
433
434 * figure out a nice way how we can let the admin know what child/sibling unit causes cgroup membership for a specific unit
435
436 * For timer units: add some mechanisms so that timer units that trigger immediately on boot do not have the services
437 they run added to the initial transaction and thus confuse Type=idle.
438
439 * add bus api to query unit file's X fields.
440
441 * gpt-auto-generator:
442 - Define new partition type for encrypted swap? Support probed LUKS for encrypted swap?
443 - Make /home automount rather than mount?
444
445 * add generator that pulls in systemd-network from containers when
446 CAP_NET_ADMIN is set, more than the loopback device is defined, even
447 when it is otherwise off
448
449 * MessageQueueMessageSize= (and suchlike) should use parse_iec_size().
450
451 * implement Distribute= in socket units to allow running multiple
452 service instances processing the listening socket, and open this up
453 for ReusePort=
454
455 * socket units: support creating sockets in different namespace,
456 opening it up for JoinsNamespaceOf=. This would require to fork off
457 a tiny process that joins the namespace and creates/binds the socket
458 and passes this back to PID1 via SCM_RIGHTS. This also could be used
459 to allow Chown/chgrp on sockets without requiring NSS in PID 1.
460
461 * introduce bus call FreezeUnit(s, b), as well as "systemctl freeze
462 $UNIT" and "systemctl thaw $UNIT" as wrappers around this. The calls
463 should SIGSTOP all unit processes in a loop until all processes of
464 it are fully stopped. This can later be used for app management by
465 desktop UIs such as gnome-shell to freeze apps that are not visible
466 on screen, not unlike how job control works on the shell
467
468 * cgroups:
469 - implement per-slice CPUFairScheduling=1 switch
470 - handle jointly mounted controllers correctly
471 - introduce high-level settings for RT budget, swappiness
472 - how to reset dynamically changed unit cgroup attributes sanely?
473 - when reloading configuration, apply new cgroup configuration
474 - when recursively showing the cgroup hierarchy, optionally also show
475 the hierarchies of child processes
476
477 * transient units:
478 - add field to transient units that indicate whether systemd or somebody else saves/restores its settings, for integration with libvirt
479 - ensure scope units may be started only a single time
480
481 * Automatically configure swap partition to use for hibernation by looking for largest swap partition on the root disk?
482
483 * when we detect low battery and no AC on boot, show pretty splash and refuse boot
484
485 * libsystemd-journal, libsystemd-login, libudev: add calls to easily attach these objects to sd-event event loops
486
487 * be more careful what we export on the bus as (usec_t) 0 and (usec_t) -1
488
489 * 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
490
491 * After coming back from hibernation reset hibernation swap partition using the /dev/snapshot ioctl APIs
492
493 * If we try to find a unit via a dangling symlink, generate a clean
494 error. Currently, we just ignore it and read the unit from the search
495 path anyway.
496
497 * refuse boot if /usr/lib/os-release is missing or /etc/machine-id cannot be set up
498
499 * man: the documentation of Restart= currently is very misleading and suggests the tools from ExecStartPre= might get restarted.
500
501 * load .d/*.conf dropins for device units
502
503 * allow implementation of InaccessibleDirectories=/ plus
504 ReadOnlyDirectories=... for whitelisting files for a service.
505
506 * sd-bus:
507 - EBADSLT handling
508 - GetAllProperties() on a non-existing object does not result in a failure currently
509 - port to sd-resolve for connecting to TCP dbus servers
510 - 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
511 - see if we can drop more message validation on the sending side
512 - add API to clone sd_bus_message objects
513 - longer term: priority inheritance
514 - dbus spec updates:
515 - NameLost/NameAcquired obsolete
516 - GVariant
517 - path escaping
518 - update systemd.special(7) to mention that dbus.socket is only about the compatibility socket now
519
520 * sd-event
521 - allow multiple signal handlers per signal?
522 - document chaining of signal handler for SIGCHLD and child handlers
523 - define more intervals where we will shift wakeup intervals around in, 1h, 6h, 24h, ...
524 - generate a failure of a default event loop is executed out-of-thread
525 - maybe add support for inotify events
526
527 * investigate endianness issues of UUID vs. GUID
528
529 * dbus: when a unit failed to load (i.e. is in UNIT_ERROR state), we
530 should be able to safely try another attempt when the bus call LoadUnit() is invoked.
531
532 * add a pam module that passes the hdd passphrase into the PAM stack and then expires it, for usage by gdm auto-login.
533
534 * add a pam module that on password changes updates any LUKS slot where the password matches
535
536 * maybe add a generator that looks for "systemd.run=" on the kernel cmdline for container usercases...
537
538 * test/:
539 - add 'set -e' to scripts in test/
540 - make stuff in test/ work with separate output dir
541
542 * seems that when we follow symlinks to units we prefer the symlink
543 destination path over /etc and /usr. We should not do that. Instead
544 /etc should always override /run+/usr and also any symlink
545 destination.
546
547 * when isolating, try to figure out a way how we implicitly can order
548 all units we stop before the isolating unit...
549
550 * teach ConditionKernelCommandLine= globs or regexes (in order to match foobar={no,0,off})
551
552 * BootLoaderSpec: Clarify that the kernel has to be in $BOOT. Clarify
553 that the boot loader should be installed to the ESP. Define a way
554 how an installer can figure out whether a BLS compliant boot loader
555 is installed.
556
557 * think about requeuing jobs when daemon-reload is issued? usecase:
558 the initrd issues a reload after fstab from the host is accessible
559 and we might want to requeue the mounts local-fs acquired through
560 that automatically.
561
562 * systemd-inhibit: make taking delay locks useful: support sending SIGINT or SIGTERM on PrepareForSleep()
563
564 * remove any syslog support from log.c — we probably cannot do this before split-off udev is gone for good
565
566 * shutdown logging: store to EFI var, and store to USB stick?
567
568 * merge unit_kill_common() and unit_kill_context()
569
570 * introduce ExecCondition= in services
571
572 * EFI:
573 - honor language efi variables for default language selection (if there are any?)
574 - honor timezone efi variables for default timezone selection (if there are any?)
575 - change bootctl to be backed by systemd-bootd to control temporary and persistent default boot goal plus efi variables
576
577 * maybe do not install getty@tty1.service symlink in /etc but in /usr?
578
579 * print a nicer explanation if people use variable/specifier expansion in ExecStart= for the first word
580
581 * mount: turn dependency information from /proc/self/mountinfo into dependency information between systemd units.
582
583 * logind:
584 - logind: optionally, ignore idle-hint logic for autosuspend, block suspend as long as a session is around
585 - When we update the kernel all kind of hibernation should be prohibited until shutdown/reboot
586 - logind: wakelock/opportunistic suspend support
587 - Add pretty name for seats in logind
588 - logind: allow showing logout dialog from system?
589 - we should probably handle SIGTERM/SIGINT to not leave dot files around, just in case
590 - session scopes/user unit: add RequiresMountsFor for the home directory of the user
591 - add Suspend() bus calls which take timestamps to fix double suspend issues when somebody hits suspend and closes laptop quickly.
592 - if pam_systemd is invoked by su from a process that is outside of a
593 any session we should probably just become a NOP, since that's
594 usually not a real user session but just some system code that just
595 needs setuid().
596 - logind: make the Suspend()/Hibernate() bus calls wait for the for
597 the job to be completed. before returning, so that clients can wait
598 for "systemctl suspend" to finish to know when the suspending is
599 complete.
600 - logind: when the power button is pressed short, just popup a
601 logout dialog. If it is pressed for 1s, do the usual
602 shutdown. Inspiration are Macs here.
603 - expose "Locked" property on logind sesison objects
604 - given that logind now lets PID 1 do all nasty work, we can
605 probably reduce the capability set it retains substantially.
606 (we need CAP_SYS_ADMIN for drmSetMaster(), so maybe not worth it)
607 - expose orientation sensors and tablet mode through logind
608 - maybe allow configuration of the StopTimeout for session scopes
609 - rename session scope so that it includes the UID. THat way
610 the session scope can be arranged freely in slices and we don't have
611 make assumptions about their slice anymore.
612 - follow PropertiesChanged state more closely, to deal with quick logouts and
613 relogins
614
615 * exec: when deinitializating a tty device fix the perms and group, too, not only when initializing. Set access mode/gid to 0620/tty.
616
617 * service: watchdog logic: for testing purposes allow ping, but do not require pong
618
619 * journal:
620 - consider introducing implicit _TTY= + _PPID= + _EUID= + _EGID= + _FSUID= + _FSGID= fields
621 - import and delete pstore filesystem content at startup
622 - journald: also get thread ID from client, plus thread name
623 - journal: when waiting for journal additions in the client always sleep at least 1s or so, in order to minimize wakeups
624 - add API to close/reopen/get fd for journal client fd in libsystemd-journal.
625 - fallback to /dev/log based logging in libsystemd-journal, if we cannot log natively?
626 - declare the local journal protocol stable in the wiki interface chart
627 - sd-journal: speed up sd_journal_get_data() with transparent hash table in bg
628 - journald: when dropping msgs due to ratelimit make sure to write
629 "dropped %u messages" not only when we are about to print the next
630 message that works, but alraedy after a short tiemout
631 - check if we can make journalctl by default use --follow mode inside of less if called without args?
632 - maybe add API to send pairs of iovecs via sd_journal_send
633 - journal: add a setgid "systemd-journal" utility to invoke from libsystemd-journal, which passes fds via STDOUT and does PK access
634 - journactl: support negative filtering, i.e. FOOBAR!="waldo",
635 and !FOOBAR for events without FOOBAR.
636 - journal: store timestamp of journal_file_set_offline() int he header,
637 so it is possible to display when the file was last synced.
638 - 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.
639 - journal: find a way to allow dropping history early, based on priority, other rules
640 - journal: When used on NFS, check payload hashes
641 - journald: add kernel cmdline option to disable ratelimiting for debug purposes
642 - refuse taking lower-case variable names in sd_journal_send() and friends.
643 - journald: we currently rotate only after MaxUse+MaxFilesize has been reached.
644 - journal: deal nicely with byte-by-byte copied files, especially regards header
645 - journal: sanely deal with entries which are larger than the individual file size, but where the components would fit
646 - Replace utmp, wtmp, btmp, and lastlog completely with journal
647 - journalctl: instead --after-cursor= maybe have a --cursor=XYZ+1 syntax?
648 - when a kernel driver logs in a tight loop, we should ratelimit that too.
649 - journald: optionally, log debug messages to /run but everything else to /var
650 - journald: when we drop syslog messages because the syslog socket is
651 full, make sure to write how many messages are lost as first thing
652 to syslog when it works again.
653 - change systemd-journal-flush into a service that stays around during
654 boot, and causes the journal to be moved back to /run on shutdown,
655 so that we do not keep /var busy. This needs to happen synchronously,
656 hence doing this via signals is not going to work.
657 - optionally support running journald from the command line for testing purposes in external projects
658 - journald: allow per-priority and per-service retention times when rotating/vacuuming
659 - journald: make use of uid-range.h to managed uid ranges to split
660 journals in.
661 - journalctl: add the ability to look for the most recent process of a binary. journalctl /usr/bin/X11 --pid=-1 or so...
662 - improve journalctl performance by loading journal files
663 lazily. Encode just enough information in the file name, so that we
664 do not have to open it to know that it is not interesting for us, for
665 the most common operations.
666 - journal-or-kmsg is currently broken? See reverted
667 commit 4a01181e460686d8b4a543b1dfa7f77c9e3c5ab8.
668 - man: document that corrupted journal files is nothing to act on
669 - rework journald sigbus stuff to use mutex
670 - Set RLIMIT_NPROC for systemd-journal-xyz, and all other of our
671 services that run under their own user ids, and use User= (but only
672 in a world where userns is ubiquitous since otherwise we cannot
673 invoke those daemons on the host AND in a container anymore). Also,
674 if LimitNPROC= is used without User= we should warn and refuse
675 operation.
676 - journalctl --verify: don't show files that are currently being
677 written to as FAIL, but instead show that their are being written to.
678 - add journalctl -H that talks via ssh to a remote peer and passes through
679 binary logs data
680 - add a version of --merge which also merges /var/log/journal/remote
681 - journalctl: -m should access container journals directly by enumerating
682 them via machined, and also watch containers coming and going.
683 Benefit: nspawn --ephemeral would start working nicely with the journal.
684 - assign MESSAGE_ID to log messages about failed services
685
686 * add a test if all entries in the catalog are properly formatted.
687 (Adding dashes in a catalog entry currently results in the catalog entry
688 being silently skipped. journalctl --update-catalog must warn about this,
689 and we should also have a unit test to check that all our message are OK.)
690
691 * document:
692 - document that deps in [Unit] sections ignore Alias= fields in
693 [Install] units of other units, unless those units are disabled
694 - man: clarify that time-sync.target is not only sysv compat but also useful otherwise. Same for similar targets
695 - document that service reload may be implemented as service reexec
696 - add a man page containing packaging guidelines and recommending usage of things like Documentation=, PrivateTmp=, PrivateNetwork= and ReadOnlyDirectories=/etc /usr.
697 - document systemd-journal-flush.service properly
698 - documentation: recommend to connect the timer units of a service to the service via Also= in [Install]
699 - man: document the very specific env the shutdown drop-in tools live in
700 - man: add more examples to man pages
701 - man: maybe sort directives in man pages, and take sections from --help and apply them to man too
702
703 * systemctl:
704 - add systemctl switch to dump transaction without executing it
705 - Add a verbose mode to "systemctl start" and friends that explains what is being done or not done
706 - "systemctl disable" on a static unit prints no message and does
707 nothing. "systemctl enable" does nothing, and gives a bad message
708 about it. Should fix both to print nice actionable messages.
709 - print nice message from systemctl --failed if there are no entries shown, and hook that into ExecStartPre of rescue.service/emergency.service
710 - add new command to systemctl: "systemctl system-reexec" which reexecs as many daemons as virtually possible
711 - systemctl enable: fail if target to alias into does not exist? maybe show how many units are enabled afterwards?
712 - systemctl: "Journal has been rotated since unit was started." message is misleading
713 - systemctl status output should include list of triggering units and their status
714
715 * unit install:
716 - "systemctl mask" should find all names by which a unit is accessible
717 (i.e. by scanning for symlinks to it) and link them all to /dev/null
718
719 * timer units:
720 - timer units should get the ability to trigger when:
721 o CLOCK_REALTIME makes jumps (TFD_TIMER_CANCEL_ON_SET)
722 o DST changes
723 - Modulate timer frequency based on battery state
724
725 * add libsystemd-password or so to query passwords during boot using the password agent logic
726
727 * clean up date formatting and parsing so that all absolute/relative timestamps we format can also be parsed
728
729 * on shutdown: move utmp, wall, audit logic all into PID 1 (or logind?), get rid of systemd-update-utmp-runlevel
730
731 * make repeated alt-ctrl-del presses printing a dump
732
733 * hostnamed: before returning information from /etc/machine-info.conf check the modification data and reread. Similar for localed, ...
734
735 * currently x-systemd.timeout is lost in the initrd, since crypttab is copied into dracut, but fstab is not
736
737 * nspawn:
738 - emulate /dev/kmsg using CUSE and turn off the syslog syscall
739 with seccomp. That should provide us with a useful log buffer that
740 systemd can log to during early boot, and disconnect container logs
741 from the kernel's logs.
742 - as soon as networkd has a bus interface, hook up --network-interface=,
743 --network-bridge= with networkd, to trigger netdev creation should an
744 interface be missing
745 - a nice way to boot up without machine id set, so that it is set at boot
746 automatically for supporting --ephemeral. Maybe hash the host machine id
747 together with the machine name to generate the machine id for the container
748 - fix logic always print a final newline on output.
749 https://github.com/systemd/systemd/pull/272#issuecomment-113153176
750 - should optionally support receiving WATCHDOG=1 messages from its payload
751 PID 1...
752 - optionally automatically add FORWARD rules to iptables whenever nspawn is
753 running, remove them when shut down.
754 - maybe make copying of /etc/resolv.conf optional, and skip it if --read-only
755 is used
756
757 * dissect
758 - refuse mounting over a mount point
759 - automatically discover .roothash files in dissect, similarly to nspawn
760
761 * machined:
762 - add an API so that libvirt-lxc can inform us about network interfaces being
763 removed or added to an existing machine
764 - "machinectl migrate" or similar to copy a container from or to a
765 difference host, via ssh
766 - introduce systemd-nspawn-ephemeral@.service, and hook it into
767 "machinectl start" with a new --ephemeral switch
768 - "machinectl status" should also show internal logs of the container in
769 question
770 - "machinectl list-images" should show os-release data, as well as
771 machine-info data (including deployment level)
772 - "machinectl history"
773 - "machinectl diff"
774 - "machinectl commit" that takes a writable snapshot of a tree, invokes a
775 shell in it, and marks it read-only after use
776
777 * importd:
778 - generate a nice warning if mkfs.btrfs is missing
779
780 * cryptsetup:
781 - cryptsetup-generator: allow specification of passwords in crypttab itself
782 - support rd.luks.allow-discards= kernel cmdline params in cryptsetup generator
783
784 * hw watchdog: optionally try to use the preset watchdog timeout instead of always overriding it
785 https://bugs.freedesktop.org/show_bug.cgi?id=54712
786
787 * create /sbin/init symlinks from the build system
788
789 * add a dependency on standard-conf.xml and other included files to man pages
790
791 * MountFlags=shared acts as MountFlags=slave right now.
792
793 * properly handle loop back mounts via fstab, especially regards to fsck/passno
794
795 * initialize the hostname from the fs label of /, if /etc/hostname does not exist?
796
797 * udev:
798 - move to LGPL
799 - kill scsi_id
800 - add trigger --subsystem-match=usb/usb_device device
801 - reimport udev db after MOVE events for devices without dev_t
802
803 * There's currently no way to cancel fsck (used to be possible via C-c or c on the console)
804
805 * add option to sockets to avoid activation. Instead just drop packets/connections, see http://cyberelk.net/tim/2012/02/15/portreserve-systemd-solution/
806
807 * coredump:
808 - save coredump in Windows/Mozilla minidump format
809 - when truncating coredumps, also log the full size that the process had, and make a metadata field so we can report truncated coredumps
810
811 * support crash reporting operation modes (https://live.gnome.org/GnomeOS/Design/Whiteboards/ProblemReporting)
812
813 * default to actual 32-bit PIDs, via /proc/sys/kernel/pid_max
814
815 * be able to specify a forced restart of service A where service B depends on, in case B
816 needs to be auto-respawned?
817
818 * tmpfiles:
819 - apply "x" on "D" too (see patch from William Douglas)
820 - replace F with f+.
821 - instead of ignoring unknown fields, reject them.
822 - creating new directories/subvolumes/fifos/device nodes
823 should not follow symlinks. None of the other adjustment or creation
824 calls follow symlinks.
825
826 * make sure systemd-ask-password-wall does not shutdown systemd-ask-password-console too early
827
828 * verify that the AF_UNIX sockets of a service in the fs still exist
829 when we start a service in order to avoid confusion when a user
830 assumes starting a service is enough to make it accessible
831
832 * Make it possible to set the keymap independently from the font on
833 the kernel cmdline. Right now setting one resets also the other.
834
835 * and a dbus call to generate target from current state
836
837 * write blog stories about:
838 - hwdb: what belongs into it, lsusb
839 - enabling dbus services
840 - how to make changes to sysctl and sysfs attributes
841 - remote access
842 - how to pass throw-away units to systemd, or dynamically change properties of existing units
843 - testing with Harald's awesome test kit
844 - auto-restart
845 - how to develop against journal browsing APIs
846 - the journal HTTP iface
847 - non-cgroup resource management
848 - dynamic resource management with cgroups
849 - refreshed, longer missions statement
850 - calendar time events
851 - init=/bin/sh vs. "emergency" mode, vs. "rescue" mode, vs. "multi-user" mode, vs. "graphical" mode, and the debug shell
852 - how to create your own target
853 - instantiated apache, dovecot and so on
854 - hooking a script into various stages of shutdown/rearly booot
855
856 * investigate whether the gnome pty helper should be moved into systemd, to provide cgroup support.
857
858 * dot output for --test showing the 'initial transaction'
859
860 * pid1:
861 - When logging about multiple units (stopping BoundTo units, conflicts, etc.),
862 log both units as UNIT=, so that journalctl -u triggers on both.
863 - generate better errors when people try to set transient properties
864 that are not supported...
865 http://lists.freedesktop.org/archives/systemd-devel/2015-February/028076.html
866 - maybe introduce WantsMountsFor=? Usecase:
867 http://lists.freedesktop.org/archives/systemd-devel/2015-January/027729.html
868 - recreate systemd's D-Bus private socket file on SIGUSR2
869 - move PAM code into its own binary
870 - when we automatically restart a service, ensure we restart its rdeps, too.
871 - hide PAM options in fragment parser when compile time disabled
872 - Support --test based on current system state
873 - 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().
874 - after deserializing sockets in socket.c we should reapply sockopts and things
875 - drop PID 1 reloading, only do reexecing (difficult: Reload()
876 currently is properly synchronous, Reexec() is weird, because we
877 cannot delay the response properly until we are back, so instead of
878 being properly synchronous we just keep open the fd and close it
879 when done. That means clients do not get a successful method reply,
880 but much rather a disconnect on success.
881 - when breaking cycles drop sysv services first, then services from /run, then from /etc, then from /usr
882 - when a bus name of a service disappears from the bus make sure to queue further activation requests
883
884 * unit files:
885 - allow port=0 in .socket units
886 - maybe introduce ExecRestartPre=
887 - add ReloadSignal= for configuring a reload signal to use
888 - implement Register= switch in .socket units to enable registration
889 in Avahi, RPC and other socket registration services.
890 - allow Type=simple with PIDFile=
891 https://bugzilla.redhat.com/show_bug.cgi?id=723942
892 - allow writing multiple conditions in unit files on one line
893 - load-fragment: when loading a unit file via a chain of symlinks
894 verify that it is not masked via any of the names traversed.
895 - introduce Type=pid-file
896 - introduce mix of BindTo and Requisite
897 - add a concept of RemainAfterExit= to scope units
898 - Allow multiple ExecStart= for all Type= settings, so that we can cover rescue.service nicely
899
900 * udev-link-config:
901 - Make sure ID_PATH is always exported and complete for
902 network devices where possible, so we can safely rely
903 on Path= matching
904 - check MTUBytes parsing (expecting size_t but we are using unsigned)
905
906 * sd-rtnl:
907 - add support for more attribute types
908 - inbuilt piping support (essentially degenerate async)? see loopback-setup.c and other places
909
910 * networkd:
911 - add more keys to [Route] and [Address] sections
912 - add support for more DHCPv4 options (and, longer term, other kinds of dynamic config)
913 - add proper initrd support (in particular generate .network/.link files based on /proc/cmdline)
914 - add reduced [Link] support to .network files
915 - add Scope= parsing option for [Network]
916 - properly handle routerless dhcp leases
917 - work with non-Ethernet devices
918 - add support for more bond options
919 - dhcp: do we allow configuring dhcp routes on interfaces that are not the one we got the dhcp info from?
920 - the DHCP lease data (such as NTP/DNS) is still made available when
921 a carrier is lost on a link. It should be removed instantly.
922 - expose in the API the following bits:
923 - option 15, domain name and/or option 119, search list
924 - option 12, host name and/or option 81, fqdn
925 - option 123, 144, geolocation
926 - option 252, configure http proxy (PAC/wpad)
927 - provide a way to define a per-network interface default metric value
928 for all routes to it. possibly a second default for DHCP routes.
929 - allow Name= to be specified repeatedly in the [Match] section. Maybe also
930 support Name=foo*|bar*|baz ?
931 - duplicate address check for static IPs (like ARPCHECK in network-scripts)
932 - allow DUID/IAID to be customized, see issue #394.
933 - whenever uplink info changes, make DHCP server send out FORCERENEW
934
935 * networkd-wait-online:
936 - make operstates to wait for configurable?
937
938 * dhcp:
939 - figure out how much we can increase Maximum Message Size
940
941 * dhcp6:
942 - add functions to set previously stored IPv6 addresses on startup and get
943 them at shutdown; store them in client->ia_na
944 - write more test cases
945 - implement reconfigure support, see 5.3., 15.11. and 22.20.
946 - implement support for temporary adressess (IA_TA)
947 - implement dhcpv6 authentication
948 - investigate the usefulness of Confirm messages; i.e. are there any
949 situations where the link changes without any loss in carrier detection
950 or interface down
951 - some servers don't do rapid commit without a filled in IA_NA, verify
952 this behavior
953 - RouteTable= ?
954
955 External:
956
957 * dbus:
958 - natively watch for dbus-*.service symlinks (PENDING)
959 - teach dbus to activate all services it finds in /etc/systemd/services/org-*.service
960
961 * fix alsa mixer restore to not print error when no config is stored
962
963 * make cryptsetup lower --iter-time
964
965 * patch kernel for xattr support in /dev, /proc/, /sys?
966
967 * kernel: add device_type = "fb", "fbcon" to class "graphics"
968
969 * drop accountsservice's StandardOutput=syslog and Type=dbus fields
970
971 * /usr/bin/service should actually show the new command line
972
973 * 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=
974
975 * fedora: F20: go timer units all the way, leave cron.daily for cron
976
977 * neither pkexec nor sudo initialize environ[] from the PAM environment?
978
979 * fedora: update policy to declare access mode and ownership of unit files to root:root 0644, and add an rpmlint check for it
980
981 * register catalog database signature as file magic
982
983 * zsh shell completion:
984 - <command> <verb> -<TAB> should complete options, but currently does not
985 - systemctl add-wants,add-requires
986
987
988 Regularly:
989
990 * look for close() vs. close_nointr() vs. close_nointr_nofail()
991
992 * check for strerror(r) instead of strerror(-r)
993
994 * pahole
995
996 * set_put(), hashmap_put() return values check. i.e. == 0 does not free()!
997
998 * use secure_getenv() instead of getenv() where appropriate
999
1000 * link up selected blog stories from man pages and unit files Documentation= fields