]> git.ipfire.org Git - thirdparty/systemd.git/blob - TODO
man: document the new logic
[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 * expose IO accounting data on the bus, show it in systemd-run --wait and log
28 about it in the resource log message
29
30 * add "systemctl purge" for flushing out configuration, state, logs, ... of a
31 unit when it is stopped
32
33 * show whether a service has out-of-date configuration in "systemctl status" by
34 using mtime data of ConfigurationDirectory=.
35
36 * Properly chmod() RuntimeDirectory=, StateDirectory=, LogsDirectory= and
37 CacheDirectory= when we start up and the directory isn't properly owned. In
38 particular to make DynamicUser= work
39
40 * replace all uses of fgets() + LINE_MAX by read_line()
41
42 * set IPAddressDeny=any on all services that shouldn't do networking (possibly
43 combined with IPAddressAllow=localhost).
44
45 * dissect: when we discover squashfs, don't claim we had a "writable" partition
46 in systemd-dissect
47
48 * Add AddUser= setting to unit files, similar to DynamicUser=1 which however
49 creates a static, persistent user rather than a dynamic, transient user. We
50 can leverage code from sysusers.d for this.
51
52 * add some optional flag to ReadWritePaths= and friends, that has the effect
53 that we create the dir in question when the service is started. Example:
54
55 ReadWritePaths=:/var/lib/foobar
56
57 * sort generated hwdb files alphabetically when we import them, so that git
58 diffs remain minimal (in particular: the OUI databases we import are not
59 sorted, and not stable)
60
61 * maybe add call sd_journal_set_block_timeout() or so to set SO_SNDTIMEO for
62 the sd-journal logging socket, and, if the timeout is set to 0, sets
63 O_NONBLOCK on it. That way people can control if and when to block for
64 logging.
65
66 * tighten sd_notify() MAINPID= checks a bit: don't accept foreign PIDs (i.e.
67 PIDs not managed by the service manager)
68
69 * hostnamed: populate form factor data from a new hwdb database, so that old
70 yogas can be recognized as "convertible" too, even if they predate the DMI
71 "convertible" form factor
72
73 * Maybe add a small tool invoked early at boot, that adds in or resizes
74 partitions automatically, to be used when the media used is actually larger
75 than the image written onto it is.
76
77 * Maybe add PrivatePIDs= as new unit setting, and do minimal PID namespacing
78 after all. Be strict however, only support the equivalent of nspawn's
79 --as-pid2 switch, and sanely proxy sd_notify() messages dropping stuff such
80 as MAINPID.
81
82 * change the dependency Set* objects in Unit structures to become Hashmap*, and
83 then store a bit mask who created a specific dependency: the source unit via
84 fragment configuration, the destination unit via fragment configuration, or
85 the source unit via udev rules (in case of .device units), or any combination
86 thereof. This information can then be used to flush out old udev-created
87 dependencies when the udev properties change, and eventually to implement a
88 "systemctl refresh" operation for reloading the configuration of individual
89 units without reloading the whole set.
90
91 * Add ExecMonitor= setting. May be used multiple times. Forks off a process in
92 the service cgroup, which is supposed to monitor the service, and when it
93 exits the service is considered failed by its monitor.
94
95 * track the per-service PAM process properly (i.e. as an additional control
96 process), so that it may be queried on the bus and everything.
97
98 * add a new "debug" job mode, that is propagated to unit_start() and for
99 services results in two things: we raise SIGSTOP right before invoking
100 execve() and turn off watchdog support. Then, use that to implement
101 "systemd-gdb" for attaching to the start-up of any system service in its
102 natural habitat.
103
104 * replace all canonicalize_file_name() invocations by chase_symlinks(), in
105 particulr those where a rootdir is relevant.
106
107 * maybe introduce gpt auto discovery for /var/tmp?
108
109 * fix PrivateNetwork= so that we fall back gracefully on kernels lacking
110 namespacing support (similar for the other namespacing options)
111
112 * maybe add gpt-partition-based user management: each user gets his own
113 LUKS-encrypted GPT partition with a new GPT type. A small nss module
114 enumerates users via udev partition enumeration. UIDs are assigned in a fixed
115 way: the partition index is added as offset to some fixed base uid. User name
116 is stored in GPT partition name. A PAM module authenticates the user via the
117 LUKS partition password. Benefits: strong per-user security, compatibility
118 with stateless/read-only/verity-enabled root. (other idea: do this based on
119 loopback files in /home, without GPT involvement)
120
121 * gpt-auto logic: introduce support for discovering /var matching an image. For
122 that, use a partition type UUID that is hashed from the OS name (as encoded
123 in /etc/os-release), the architecture, and 4 new bits from the gpt flags
124 field of the root partition. This way can easily support multiple OS
125 installations on the same GPT partition table, without problems with
126 unmatched /var partitions.
127
128 * gpt-auto logic: related to the above, maybe support a "secondary" root
129 partition, that is mounted to / and is writable, and where the actual root's
130 /usr is mounted into.
131
132 * machined: add apis to query /etc/machine-info data of a container
133
134 * .mount and .swap units: add Format=yes|no option that formats the partition before mounting/enabling it, implicitly
135
136 * 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
137
138 * drop nss-myhostname in favour of nss-resolve?
139
140 * drop internal dlopen() based nss-dns fallback in nss-resolve, and rely on the
141 external nsswitch.conf based one
142
143 * add a percentage syntax for TimeoutStopSec=, e.g. TimeoutStopSec=150%, and
144 then use that for the setting used in user@.service. It should be understood
145 relative to the configured default value.
146
147 * on cgroupsv2 add DelegateControllers=, to pick the precise cgroup controllers to delegate
148
149 * in networkd, when matching device types, fix up DEVTYPE rubbish the kernel passes to us
150
151 * enable LockMLOCK to take a percentage value relative to physical memory
152
153 * Permit masking specific netlink APIs with RestrictAddressFamily=
154
155 * nspawn: start UID allocation loop from hash of container name
156
157 * nspawn: support that /proc, /sys/, /dev are pre-mounted
158
159 * define gpt header bits to select volatility mode
160
161 * ProtectKernelLogs= (drops CAP_SYSLOG, add seccomp for syslog() syscall, and DeviceAllow to /dev/kmsg) in service files
162
163 * ProtectClock= (drops CAP_SYS_TIMES, adds seecomp filters for settimeofday, adjtimex), sets DeviceAllow o /dev/rtc
164
165 * ProtectTracing= (drops CAP_SYS_PTRACE, blocks ptrace syscall, makes /sys/kernel/tracing go away)
166
167 * ProtectMount= (drop mount/umount/pivot_root from seccomp, disallow fuse via DeviceAllow, imply Mountflags=slave)
168
169 * ProtectKeyRing= to take keyring calls away
170
171 * RemoveKeyRing= to remove all keyring entries of the specified user
172
173 * ProtectReboot= that masks reboot() and kexec_load() syscalls, prohibits kill
174 on PID 1 with the relevant signals, and makes relevant files in /sys and
175 /proc (such as the sysrq stuff) unavailable
176
177 * DeviceAllow= should also generate seccomp filters for mknod()
178
179 * Add DataDirectory=, CacheDirectory= and LogDirectory= to match
180 RuntimeDirectory=, and create it as necessary when starting a service, owned by the right user.
181
182 * make sure the ratelimit object can deal with USEC_INFINITY as way to turn off things
183
184 * journalctl: make sure -f ends when the container indicated by -M terminates
185
186 * mount: automatically search for "main" partition of an image has multiple
187 partitions
188
189 * expose the "privileged" flag of ExecCommand on the bus, and open it up to
190 transient units
191
192 * in nss-systemd, if we run inside of RootDirectory= with PrivateUsers= set,
193 find a way to map the User=/Group= of the service to the right name. This way
194 a user/group for a service only has to exist on the host for the right
195 mapping to work.
196
197 * allow attaching additional journald log fields to cgroups
198
199 * add bus API for creating unit files in /etc, reusing the code for transient units
200
201 * add bus API to remove unit files from /etc
202
203 * add bus API to retrieve current unit file contents (i.e. implement "systemctl cat" on the bus only)
204
205 * rework fopen_temporary() to make use of open_tmpfile_linkable() (problem: the
206 kernel doesn't support linkat() that replaces existing files, currently)
207
208 * check if DeviceAllow= should split first, resolve specifiers later
209
210 * transient units: don't bother with actually setting unit properties, we
211 reload the unit file anyway
212
213 * journald: sigbus API via a signal-handler safe function that people may call
214 from the SIGBUS handler
215
216 * optionally, also require WATCHDOG=1 notifications during service start-up and shutdown
217
218 * resolved: when routing queries, make sure only look for the *longest* suffix...
219
220 * delay activation of logind until somebody logs in, or when /dev/tty0 pulls it
221 in or lingering is on (so that containers don't bother with it until PAM is used). also exit-on-idle
222
223 * cache sd_event_now() result from before the first iteration...
224
225 * add systemctl stop --job-mode=triggering that follows TRIGGERED_BY deps and adds them to the same transaction
226
227 * PID1: find a way how we can reload unit file configuration for
228 specific units only, without reloading the whole of systemd
229
230 * add an explicit parser for LimitRTPRIO= that verifies
231 the specified range and generates sane error messages for incorrect
232 specifications.
233
234 * do something about "/control" subcgroups in the unified cgroup hierarchy
235
236 * when we detect that there are waiting jobs but no running jobs, do something
237
238 * push CPUAffinity= also into the "cpuset" cgroup controller (only after the cpuset controller got ported to the unified hierarchy)
239
240 * PID 1 should send out sd_notify("WATCHDOG=1") messages (for usage in the --user mode, and when run via nspawn)
241
242 * there's probably something wrong with having user mounts below /sys,
243 as we have for debugfs. for exmaple, src/core/mount.c handles mounts
244 prefixed with /sys generally special.
245 http://lists.freedesktop.org/archives/systemd-devel/2015-June/032962.html
246
247 * man: document that unless you use StandardError=null the shell >/dev/stderr won't work in shell scripts in services
248
249 * fstab-generator: default to tmpfs-as-root if only usr= is specified on the kernel cmdline
250
251 * docs: bring http://www.freedesktop.org/wiki/Software/systemd/MyServiceCantGetRealtime up to date
252
253 * mounting and unmounting mount points manually with different source
254 devices will result in collected on all devices used.
255 http://lists.freedesktop.org/archives/systemd-devel/2015-April/030225.html
256
257 * add a job mode that will fail if a transaction would mean stopping
258 running units. Use this in timedated to manage the NTP service
259 state.
260 http://lists.freedesktop.org/archives/systemd-devel/2015-April/030229.html
261
262 * Maybe add support for the equivalent of "ethtool advertise" to .link files?
263 http://lists.freedesktop.org/archives/systemd-devel/2015-April/030112.html
264
265 * The udev blkid built-in should expose a property that reflects
266 whether media was sensed in USB CF/SD card readers. This should then
267 be used to control SYSTEMD_READY=1/0 so that USB card readers aren't
268 picked up by systemd unless they contain a medium. This would mirror
269 the behaviour we already have for CD drives.
270
271 * networkd/udev: implement SR_IOV configuration in .link files:
272 http://lists.freedesktop.org/archives/systemd-devel/2015-January/027451.html
273
274 * Rework systemctl's GetAll property parsing to use the generic bus_map_all_properties() API
275
276 * implement a per-service firewall based on net_cls
277
278 * Port various tools to make use of verbs.[ch], where applicable: busctl,
279 coredumpctl, hostnamectl, localectl, systemd-analyze, timedatectl
280
281 * hostnamectl: show root image uuid
282
283 * sysfs set api in libudev is not const
284
285 * Find a solution for SMACK capabilities stuff:
286 http://lists.freedesktop.org/archives/systemd-devel/2014-December/026188.html
287
288 * "systemctl preset-all" should probably order the unit files it
289 operates on lexicographically before starting to work, in order to
290 ensure deterministic behaviour if two unit files conflict (like DMs
291 do, for example)
292
293 * synchronize console access with BSD locks:
294 http://lists.freedesktop.org/archives/systemd-devel/2014-October/024582.html
295
296 * as soon as we have sender timestamps, revisit coalescing multiple parallel daemon reloads:
297 http://lists.freedesktop.org/archives/systemd-devel/2014-December/025862.html
298
299 * in systemctl list-unit-files: show the install value the presets would suggest for a service in a third column
300
301 * figure out when we can use the coarse timers
302
303 * add "systemctl start -v foobar.service" that shows logs of a service
304 while the start command runs. This is non-trivial to do without
305 races though, since we should flush out all journal messages before
306 returning from the "systemctl stop".
307
308 * 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
309
310 * maybe add support for specifier expansion in user.conf, specifically DefaultEnvironment=
311
312 * introduce systemd-timesync-wait.service or so to sync on an NTP fix?
313
314 * systemd --user should issue sd_notify() upon reaching basic.target, not on becoming idle
315
316 * consider showing the unit names during boot up in the status output, not just the unit descriptions
317
318 * maybe allow timer units with an empty Units= setting, so that they
319 can be used for resuming the system but nothing else.
320
321 * what to do about udev db binary stability for apps? (raw access is not an option)
322
323 * maybe provide an API to allow migration of foreign PIDs into existing scopes.
324
325 * man: maybe use the word "inspect" rather than "introspect"?
326
327 * systemctl: if some operation fails, show log output?
328
329 * systemctl edit: use equvalent of cat() to insert existing config as a comment, prepended with #.
330 Upon editor exit, lines with one # are removed, lines with two # are left with one #, etc.
331
332 * exponential backoff in timesyncd when we cannot reach a server
333
334 * timesyncd: add ugly bus calls to set NTP servers per-interface, for usage by NM
335
336 * merge ~/.local/share and ~/.local/lib into one similar /usr/lib and /usr/share....
337
338 * systemd.show_status= should probably have a mode where only failed
339 units are shown.
340
341 * add systemd.abort_on_kill or some other such flag to send SIGABRT instead of SIGKILL
342 (throughout the codebase, not only PID1)
343
344 * resolved:
345 - mDNS/DNS-SD
346 - service registration
347 - service/domain/types browsing
348 - avahi compat
349 - DNS-SD service registration from socket units
350 - resolved should optionally register additional per-interface LLMNR
351 names, so that for the container case we can establish the same name
352 (maybe "host") for referencing the server, everywhere.
353 - allow clients to request DNSSEC for a single lookup even if DNSSEC is off (?)
354 - hook up resolved with machined-based address resolution
355
356 * refcounting in sd-resolve is borked
357
358 * Add a new verb "systemctl top"
359
360 * add new gpt type for btrfs volumes
361
362 * support empty /etc boots nicely:
363 - nspawn/gpt-generator: introduce new gpt partition type for /usr
364
365 * generator that automatically discovers btrfs subvolumes, identifies their purpose based on some xattr on them.
366
367 * a way for container managers to turn off getty starting via $container_headless= or so...
368
369 * figure out a nice way how we can let the admin know what child/sibling unit causes cgroup membership for a specific unit
370
371 * mount_cgroup_controllers(): symlinks need to get the label applied
372
373 * For timer units: add some mechanisms so that timer units that trigger immediately on boot do not have the services
374 they run added to the initial transaction and thus confuse Type=idle.
375
376 * add bus api to query unit file's X fields.
377
378 * gpt-auto-generator:
379 - Define new partition type for encrypted swap? Support probed LUKS for encrypted swap?
380 - Make /home automount rather than mount?
381
382 * add generator that pulls in systemd-network from containers when
383 CAP_NET_ADMIN is set, more than the loopback device is defined, even
384 when it is otherwise off
385
386 * MessageQueueMessageSize= (and suchlike) should use parse_iec_size().
387
388 * implement Distribute= in socket units to allow running multiple
389 service instances processing the listening socket, and open this up
390 for ReusePort=
391
392 * socket units: support creating sockets in different namespace,
393 opening it up for JoinsNamespaceOf=. This would require to fork off
394 a tiny process that joins the namespace and creates/binds the socket
395 and passes this back to PID1 via SCM_RIGHTS. This also could be used
396 to allow Chown/chgrp on sockets without requiring NSS in PID 1.
397
398 * introduce bus call FreezeUnit(s, b), as well as "systemctl freeze
399 $UNIT" and "systemctl thaw $UNIT" as wrappers around this. The calls
400 should SIGSTOP all unit processes in a loop until all processes of
401 it are fully stopped. This can later be used for app management by
402 desktop UIs such as gnome-shell to freeze apps that are not visible
403 on screen, not unlike how job control works on the shell
404
405 * cgroups:
406 - implement per-slice CPUFairScheduling=1 switch
407 - handle jointly mounted controllers correctly
408 - introduce high-level settings for RT budget, swappiness
409 - how to reset dynamically changed unit cgroup attributes sanely?
410 - when reloading configuration, apply new cgroup configuration
411 - when recursively showing the cgroup hierarchy, optionally also show
412 the hierarchies of child processes
413
414 * transient units:
415 - add field to transient units that indicate whether systemd or somebody else saves/restores its settings, for integration with libvirt
416 - ensure scope units may be started only a single time
417
418 * Automatically configure swap partition to use for hibernation by looking for largest swap partition on the root disk?
419
420 * when we detect low battery and no AC on boot, show pretty splash and refuse boot
421
422 * libsystemd-journal, libsystemd-login, libudev: add calls to easily attach these objects to sd-event event loops
423
424 * be more careful what we export on the bus as (usec_t) 0 and (usec_t) -1
425
426 * 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
427
428 * After coming back from hibernation reset hibernation swap partition using the /dev/snapshot ioctl APIs
429
430 * If we try to find a unit via a dangling symlink, generate a clean
431 error. Currently, we just ignore it and read the unit from the search
432 path anyway.
433
434 * refuse boot if /usr/lib/os-release is missing or /etc/machine-id cannot be set up
435
436 * man: the documentation of Restart= currently is very misleading and suggests the tools from ExecStartPre= might get restarted.
437
438 * load .d/*.conf dropins for device units
439
440 * allow implementation of InaccessibleDirectories=/ plus
441 ReadOnlyDirectories=... for whitelisting files for a service.
442
443 * sd-bus:
444 - EBADSLT handling
445 - GetAllProperties() on a non-existing object does not result in a failure currently
446 - port to sd-resolve for connecting to TCP dbus servers
447 - 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
448 - see if we can drop more message validation on the sending side
449 - add API to clone sd_bus_message objects
450 - make AddMatch calls on dbus1 transports async?
451 - longer term: priority inheritance
452 - dbus spec updates:
453 - NameLost/NameAcquired obsolete
454 - GVariant
455 - path escaping
456 - update systemd.special(7) to mention that dbus.socket is only about the compatibility socket now
457 - test bloom filter generation indexes
458
459 * sd-event
460 - allow multiple signal handlers per signal?
461 - document chaining of signal handler for SIGCHLD and child handlers
462 - define more intervals where we will shift wakeup intervals around in, 1h, 6h, 24h, ...
463 - generate a failure of a default event loop is executed out-of-thread
464 - maybe add support for inotify events
465
466 * investigate endianness issues of UUID vs. GUID
467
468 * dbus: when a unit failed to load (i.e. is in UNIT_ERROR state), we
469 should be able to safely try another attempt when the bus call LoadUnit() is invoked.
470
471 * add a pam module that passes the hdd passphrase into the PAM stack and then expires it, for usage by gdm auto-login.
472
473 * add a pam module that on password changes updates any LUKS slot where the password matches
474
475 * maybe add a generator that looks for "systemd.run=" on the kernel cmdline for container usercases...
476
477 * test/:
478 - add 'set -e' to scripts in test/
479 - make stuff in test/ work with separate output dir
480
481 * seems that when we follow symlinks to units we prefer the symlink
482 destination path over /etc and /usr. We should not do that. Instead
483 /etc should always override /run+/usr and also any symlink
484 destination.
485
486 * when isolating, try to figure out a way how we implicitly can order
487 all units we stop before the isolating unit...
488
489 * teach ConditionKernelCommandLine= globs or regexes (in order to match foobar={no,0,off})
490
491 * BootLoaderSpec: Clarify that the kernel has to be in $BOOT. Clarify
492 that the boot loader should be installed to the ESP. Define a way
493 how an installer can figure out whether a BLS compliant boot loader
494 is installed.
495
496 * think about requeuing jobs when daemon-reload is issued? usecase:
497 the initrd issues a reload after fstab from the host is accessible
498 and we might want to requeue the mounts local-fs acquired through
499 that automatically.
500
501 * systemd-inhibit: make taking delay locks useful: support sending SIGINT or SIGTERM on PrepareForSleep()
502
503 * remove any syslog support from log.c — we probably cannot do this before split-off udev is gone for good
504
505 * shutdown logging: store to EFI var, and store to USB stick?
506
507 * think about window-manager-run-as-user-service problem: exit 0 → activate shutdown.target; exit != 0 → restart service
508
509 * merge unit_kill_common() and unit_kill_context()
510
511 * introduce ExecCondition= in services
512
513 * EFI:
514 - honor language efi variables for default language selection (if there are any?)
515 - honor timezone efi variables for default timezone selection (if there are any?)
516 - change bootctl to be backed by systemd-bootd to control temporary and persistent default boot goal plus efi variables
517
518 * maybe do not install getty@tty1.service symlink in /etc but in /usr?
519
520 * print a nicer explanation if people use variable/specifier expansion in ExecStart= for the first word
521
522 * mount: turn dependency information from /proc/self/mountinfo into dependency information between systemd units.
523
524 * logind:
525 - logind: optionally, ignore idle-hint logic for autosuspend, block suspend as long as a session is around
526 - When we update the kernel all kind of hibernation should be prohibited until shutdown/reboot
527 - logind: wakelock/opportunistic suspend support
528 - Add pretty name for seats in logind
529 - logind: allow showing logout dialog from system?
530 - we should probably handle SIGTERM/SIGINT to not leave dot files around, just in case
531 - session scopes/user unit: add RequiresMountsFor for the home directory of the user
532 - add Suspend() bus calls which take timestamps to fix double suspend issues when somebody hits suspend and closes laptop quickly.
533 - if pam_systemd is invoked by su from a process that is outside of a
534 any session we should probably just become a NOP, since that's
535 usually not a real user session but just some system code that just
536 needs setuid().
537 - logind: make the Suspend()/Hibernate() bus calls wait for the for
538 the job to be completed. before returning, so that clients can wait
539 for "systemctl suspend" to finish to know when the suspending is
540 complete.
541 - logind: when the power button is pressed short, just popup a
542 logout dialog. If it is pressed for 1s, do the usual
543 shutdown. Inspiration are Macs here.
544 - expose "Locked" property on logind sesison objects
545 - given that logind now lets PID 1 do all nasty work, we can
546 probably reduce the capability set it retains substantially.
547 (we need CAP_SYS_ADMIN for drmSetMaster(), so maybe not worth it)
548 - expose orientation sensors and tablet mode through logind
549 - maybe allow configuration of the StopTimeout for session scopes
550 - rename session scope so that it includes the UID. THat way
551 the session scope can be arranged freely in slices and we don't have
552 make assumptions about their slice anymore.
553 - follow PropertiesChanged state more closely, to deal with quick logouts and
554 relogins
555
556 * exec: when deinitializating a tty device fix the perms and group, too, not only when initializing. Set access mode/gid to 0620/tty.
557
558 * service: watchdog logic: for testing purposes allow ping, but do not require pong
559
560 * journal:
561 - consider introducing implicit _TTY= + _PPID= + _EUID= + _EGID= + _FSUID= + _FSGID= fields
562 - import and delete pstore filesystem content at startup
563 - journald: also get thread ID from client, plus thread name
564 - journal: when waiting for journal additions in the client always sleep at least 1s or so, in order to minimize wakeups
565 - add API to close/reopen/get fd for journal client fd in libsystemd-journal.
566 - fallback to /dev/log based logging in libsystemd-journal, if we cannot log natively?
567 - declare the local journal protocol stable in the wiki interface chart
568 - sd-journal: speed up sd_journal_get_data() with transparent hash table in bg
569 - journald: when dropping msgs due to ratelimit make sure to write
570 "dropped %u messages" not only when we are about to print the next
571 message that works, but alraedy after a short tiemout
572 - check if we can make journalctl by default use --follow mode inside of less if called without args?
573 - maybe add API to send pairs of iovecs via sd_journal_send
574 - journal: add a setgid "systemd-journal" utility to invoke from libsystemd-journal, which passes fds via STDOUT and does PK access
575 - journactl: support negative filtering, i.e. FOOBAR!="waldo",
576 and !FOOBAR for events without FOOBAR.
577 - journal: store timestamp of journal_file_set_offline() int he header,
578 so it is possible to display when the file was last synced.
579 - 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.
580 - journal: find a way to allow dropping history early, based on priority, other rules
581 - journal: When used on NFS, check payload hashes
582 - journald: add kernel cmdline option to disable ratelimiting for debug purposes
583 - refuse taking lower-case variable names in sd_journal_send() and friends.
584 - journald: we currently rotate only after MaxUse+MaxFilesize has been reached.
585 - journal: deal nicely with byte-by-byte copied files, especially regards header
586 - journal: sanely deal with entries which are larger than the individual file size, but where the components would fit
587 - Replace utmp, wtmp, btmp, and lastlog completely with journal
588 - journalctl: instead --after-cursor= maybe have a --cursor=XYZ+1 syntax?
589 - when a kernel driver logs in a tight loop, we should ratelimit that too.
590 - journald: optionally, log debug messages to /run but everything else to /var
591 - journald: when we drop syslog messages because the syslog socket is
592 full, make sure to write how many messages are lost as first thing
593 to syslog when it works again.
594 - journald: make sure ratelimit is actually really per-service with the new cgroup changes
595 - change systemd-journal-flush into a service that stays around during
596 boot, and causes the journal to be moved back to /run on shutdown,
597 so that we do not keep /var busy. This needs to happen synchronously,
598 hence doing this via signals is not going to work.
599 - optionally support running journald from the command line for testing purposes in external projects
600 - journald: allow per-priority and per-service retention times when rotating/vacuuming
601 - journald: make use of uid-range.h to managed uid ranges to split
602 journals in.
603 - journalctl: add the ability to look for the most recent process of a binary. journalctl /usr/bin/X11 --pid=-1 or so...
604 - improve journalctl performance by loading journal files
605 lazily. Encode just enough information in the file name, so that we
606 do not have to open it to know that it is not interesting for us, for
607 the most common operations.
608 - journal-or-kmsg is currently broken? See reverted
609 commit 4a01181e460686d8b4a543b1dfa7f77c9e3c5ab8.
610 - man: document that corrupted journal files is nothing to act on
611 - rework journald sigbus stuff to use mutex
612 - Set RLIMIT_NPROC for systemd-journal-xyz, and all other of our
613 services that run under their own user ids, and use User= (but only
614 in a world where userns is ubiquitous since otherwise we cannot
615 invoke those daemons on the host AND in a container anymore). Also,
616 if LimitNPROC= is used without User= we should warn and refuse
617 operation.
618 - journalctl --verify: don't show files that are currently being
619 written to as FAIL, but instead show that their are being written to.
620 - add journalctl -H that talks via ssh to a remote peer and passes through
621 binary logs data
622 - add a version of --merge which also merges /var/log/journal/remote
623 - log accumulated resource usage after each service invocation
624 - journalctl: -m should access container journals directly by enumerating
625 them via machined, and also watch containers coming and going.
626 Benefit: nspawn --ephemeral would start working nicely with the journal.
627 - assign MESSAGE_ID to log messages about failed services
628
629 * document:
630 - document that deps in [Unit] sections ignore Alias= fields in
631 [Install] units of other units, unless those units are disabled
632 - man: clarify that time-sync.target is not only sysv compat but also useful otherwise. Same for similar targets
633 - document the exit codes when services fail before they are exec()ed
634 - document that service reload may be implemented as service reexec
635 - document in wiki how to map ical recurrence events to systemd timer unit calendar specifications
636 - add a man page containing packaging guidelines and recommending usage of things like Documentation=, PrivateTmp=, PrivateNetwork= and ReadOnlyDirectories=/etc /usr.
637 - document systemd-journal-flush.service properly
638 - documentation: recommend to connect the timer units of a service to the service via Also= in [Install]
639 - man: document the very specific env the shutdown drop-in tools live in
640 - man: add more examples to man pages
641 - man: maybe sort directives in man pages, and take sections from --help and apply them to man too
642
643 * systemctl:
644 - add systemctl switch to dump transaction without executing it
645 - Add a verbose mode to "systemctl start" and friends that explains what is being done or not done
646 - "systemctl disable" on a static unit prints no message and does
647 nothing. "systemctl enable" does nothing, and gives a bad message
648 about it. Should fix both to print nice actionable messages.
649 - print nice message from systemctl --failed if there are no entries shown, and hook that into ExecStartPre of rescue.service/emergency.service
650 - add new command to systemctl: "systemctl system-reexec" which reexecs as many daemons as virtually possible
651 - systemctl enable: fail if target to alias into does not exist? maybe show how many units are enabled afterwards?
652 - systemctl: "Journal has been rotated since unit was started." message is misleading
653 - better error message if you run systemctl without systemd running
654 - systemctl status output should include list of triggering units and their status
655
656 * unit install:
657 - "systemctl mask" should find all names by which a unit is accessible
658 (i.e. by scanning for symlinks to it) and link them all to /dev/null
659
660 * timer units:
661 - timer units should get the ability to trigger when:
662 o CLOCK_REALTIME makes jumps (TFD_TIMER_CANCEL_ON_SET)
663 o DST changes
664 - Modulate timer frequency based on battery state
665
666 * add libsystemd-password or so to query passwords during boot using the password agent logic
667
668 * clean up date formatting and parsing so that all absolute/relative timestamps we format can also be parsed
669
670 * on shutdown: move utmp, wall, audit logic all into PID 1 (or logind?), get rid of systemd-update-utmp-runlevel
671
672 * make repeated alt-ctrl-del presses printing a dump
673
674 * hostnamed: before returning information from /etc/machine-info.conf check the modification data and reread. Similar for localed, ...
675
676 * currently x-systemd.timeout is lost in the initrd, since crypttab is copied into dracut, but fstab is not
677
678 * nspawn:
679 - emulate /dev/kmsg using CUSE and turn off the syslog syscall
680 with seccomp. That should provide us with a useful log buffer that
681 systemd can log to during early boot, and disconnect container logs
682 from the kernel's logs.
683 - as soon as networkd has a bus interface, hook up --network-interface=,
684 --network-bridge= with networkd, to trigger netdev creation should an
685 interface be missing
686 - a nice way to boot up without machine id set, so that it is set at boot
687 automatically for supporting --ephemeral. Maybe hash the host machine id
688 together with the machine name to generate the machine id for the container
689 - fix logic always print a final newline on output.
690 https://github.com/systemd/systemd/pull/272#issuecomment-113153176
691 - should optionally support receiving WATCHDOG=1 messages from its payload
692 PID 1...
693 - should send out sd_notify("WATCHDOG=1") messages
694 - optionally automatically add FORWARD rules to iptables whenever nspawn is
695 running, remove them when shut down.
696 - Improve error message when --bind= is used on a non-existing source
697 directory
698 - maybe make copying of /etc/resolv.conf optional, and skip it if --read-only
699 is used
700
701 * dissect
702 - refuse mounting over a mount point
703 - automatically discover .roothash files in dissect, similarly to nspawn
704
705 * machined:
706 - add an API so that libvirt-lxc can inform us about network interfaces being
707 removed or added to an existing machine
708 - "machinectl migrate" or similar to copy a container from or to a
709 difference host, via ssh
710 - introduce systemd-nspawn-ephemeral@.service, and hook it into
711 "machinectl start" with a new --ephemeral switch
712 - "machinectl status" should also show internal logs of the container in
713 question
714 - "machinectl list-images" should show os-release data, as well as
715 machine-info data (including deployment level)
716 - "machinectl history"
717 - "machinectl diff"
718 - "machinectl commit" that takes a writable snapshot of a tree, invokes a
719 shell in it, and marks it read-only after use
720
721 * importd:
722 - generate a nice warning if mkfs.btrfs is missing
723
724 * cryptsetup:
725 - cryptsetup-generator: allow specification of passwords in crypttab itself
726 - support rd.luks.allow-discards= kernel cmdline params in cryptsetup generator
727
728 * hw watchdog: optionally try to use the preset watchdog timeout instead of always overriding it
729 https://bugs.freedesktop.org/show_bug.cgi?id=54712
730
731 * create /sbin/init symlinks from the build system
732
733 * MountFlags=shared acts as MountFlags=slave right now.
734
735 * properly handle loop back mounts via fstab, especially regards to fsck/passno
736
737 * initialize the hostname from the fs label of /, if /etc/hostname does not exist?
738
739 * udev:
740 - move to LGPL
741 - kill scsi_id
742 - add trigger --subsystem-match=usb/usb_device device
743 - reimport udev db after MOVE events for devices without dev_t
744
745 * There's currently no way to cancel fsck (used to be possible via C-c or c on the console)
746
747 * add option to sockets to avoid activation. Instead just drop packets/connections, see http://cyberelk.net/tim/2012/02/15/portreserve-systemd-solution/
748
749 * coredump:
750 - save coredump in Windows/Mozilla minidump format
751 - when truncating coredumps, also log the full size that the process had, and make a metadata field so we can report truncated coredumps
752
753 * support crash reporting operation modes (https://live.gnome.org/GnomeOS/Design/Whiteboards/ProblemReporting)
754
755 * default to actual 32-bit PIDs, via /proc/sys/kernel/pid_max
756
757 * be able to specify a forced restart of service A where service B depends on, in case B
758 needs to be auto-respawned?
759
760 * tmpfiles:
761 - apply "x" on "D" too (see patch from William Douglas)
762 - replace F with f+.
763 - instead of ignoring unknown fields, reject them.
764 - creating new directories/subvolumes/fifos/device nodes
765 should not follow symlinks. None of the other adjustment or creation
766 calls follow symlinks.
767
768 * make sure systemd-ask-password-wall does not shutdown systemd-ask-password-console too early
769
770 * verify that the AF_UNIX sockets of a service in the fs still exist
771 when we start a service in order to avoid confusion when a user
772 assumes starting a service is enough to make it accessible
773
774 * Make it possible to set the keymap independently from the font on
775 the kernel cmdline. Right now setting one resets also the other.
776
777 * and a dbus call to generate target from current state
778
779 * write blog stories about:
780 - hwdb: what belongs into it, lsusb
781 - enabling dbus services
782 - status update
783 - how to make changes to sysctl and sysfs attributes
784 - remote access
785 - how to pass throw-away units to systemd, or dynamically change properties of existing units
786 - testing with Harald's awesome test kit
787 - auto-restart
788 - how to develop against journal browsing APIs
789 - the journal HTTP iface
790 - non-cgroup resource management
791 - dynamic resource management with cgroups
792 - refreshed, longer missions statement
793 - calendar time events
794 - init=/bin/sh vs. "emergency" mode, vs. "rescue" mode, vs. "multi-user" mode, vs. "graphical" mode, and the debug shell
795 - how to create your own target
796 - instantiated apache, dovecot and so on
797 - hooking a script into various stages of shutdown/rearly booot
798
799 * investigate whether the gnome pty helper should be moved into systemd, to provide cgroup support.
800
801 * dot output for --test showing the 'initial transaction'
802
803 * pid1:
804 - When logging about multiple units (stopping BoundTo units, conflicts, etc.),
805 log both units as UNIT=, so that journalctl -u triggers on both.
806 - generate better errors when people try to set transient properties
807 that are not supported...
808 http://lists.freedesktop.org/archives/systemd-devel/2015-February/028076.html
809 - maybe introduce WantsMountsFor=? Usecase:
810 http://lists.freedesktop.org/archives/systemd-devel/2015-January/027729.html
811 - recreate systemd's D-Bus private socket file on SIGUSR2
812 - move PAM code into its own binary
813 - when we automatically restart a service, ensure we restart its rdeps, too.
814 - hide PAM options in fragment parser when compile time disabled
815 - Support --test based on current system state
816 - 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().
817 - after deserializing sockets in socket.c we should reapply sockopts and things
818 - drop PID 1 reloading, only do reexecing (difficult: Reload()
819 currently is properly synchronous, Reexec() is weird, because we
820 cannot delay the response properly until we are back, so instead of
821 being properly synchronous we just keep open the fd and close it
822 when done. That means clients do not get a successful method reply,
823 but much rather a disconnect on success.
824 - when breaking cycles drop sysv services first, then services from /run, then from /etc, then from /usr
825 - when a bus name of a service disappears from the bus make sure to queue further activation requests
826
827 * unit files:
828 - allow port=0 in .socket units
829 - maybe introduce ExecRestartPre=
830 - add ReloadSignal= for configuring a reload signal to use
831 - implement Register= switch in .socket units to enable registration
832 in Avahi, RPC and other socket registration services.
833 - allow Type=simple with PIDFile=
834 https://bugzilla.redhat.com/show_bug.cgi?id=723942
835 - allow writing multiple conditions in unit files on one line
836 - load-fragment: when loading a unit file via a chain of symlinks
837 verify that it is not masked via any of the names traversed.
838 - introduce Type=pid-file
839 - introduce mix of BindTo and Requisite
840 - add a concept of RemainAfterExit= to scope units
841 - Allow multiple ExecStart= for all Type= settings, so that we can cover rescue.service nicely
842
843 * udev-link-config:
844 - Make sure ID_PATH is always exported and complete for
845 network devices where possible, so we can safely rely
846 on Path= matching
847 - check MTUBytes parsing (expecting size_t but we are using unsigned)
848
849 * sd-rtnl:
850 - add support for more attribute types
851 - inbuilt piping support (essentially degenerate async)? see loopback-setup.c and other places
852
853 * networkd:
854 - add more keys to [Route] and [Address] sections
855 - add support for more DHCPv4 options (and, longer term, other kinds of dynamic config)
856 - add proper initrd support (in particular generate .network/.link files based on /proc/cmdline)
857 - add reduced [Link] support to .network files
858 - add Scope= parsing option for [Network]
859 - properly handle routerless dhcp leases
860 - work with non-Ethernet devices
861 - add support for more bond options
862 - dhcp: do we allow configuring dhcp routes on interfaces that are not the one we got the dhcp info from?
863 - the DHCP lease data (such as NTP/DNS) is still made available when
864 a carrier is lost on a link. It should be removed instantly.
865 - expose in the API the following bits:
866 - option 15, domain name and/or option 119, search list
867 - option 12, host name and/or option 81, fqdn
868 - option 123, 144, geolocation
869 - option 252, configure http proxy (PAC/wpad)
870 - provide a way to define a per-network interface default metric value
871 for all routes to it. possibly a second default for DHCP routes.
872 - allow Name= to be specified repeatedly in the [Match] section. Maybe also
873 support Name=foo*|bar*|baz ?
874 - duplicate address check for static IPs (like ARPCHECK in network-scripts)
875 - allow DUID/IAID to be customized, see issue #394.
876 - whenever uplink info changes, make DHCP server send out FORCERENEW
877
878 * networkd-wait-online:
879 - make operstates to wait for configurable?
880
881 * dhcp:
882 - figure out how much we can increase Maximum Message Size
883
884 * dhcp6:
885 - add functions to set previously stored IPv6 addresses on startup and get
886 them at shutdown; store them in client->ia_na
887 - write more test cases
888 - implement reconfigure support, see 5.3., 15.11. and 22.20.
889 - implement support for temporary adressess (IA_TA)
890 - implement dhcpv6 authentication
891 - investigate the usefulness of Confirm messages; i.e. are there any
892 situations where the link changes without any loss in carrier detection
893 or interface down
894 - some servers don't do rapid commit without a filled in IA_NA, verify
895 this behavior
896 - RouteTable= ?
897
898 External:
899
900 * dbus:
901 - natively watch for dbus-*.service symlinks (PENDING)
902 - teach dbus to activate all services it finds in /etc/systemd/services/org-*.service
903
904 * fix alsa mixer restore to not print error when no config is stored
905
906 * make cryptsetup lower --iter-time
907
908 * patch kernel for xattr support in /dev, /proc/, /sys?
909
910 * kernel: add device_type = "fb", "fbcon" to class "graphics"
911
912 * drop accountsservice's StandardOutput=syslog and Type=dbus fields
913
914 * /usr/bin/service should actually show the new command line
915
916 * 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=
917
918 * fedora: F20: go timer units all the way, leave cron.daily for cron
919
920 * neither pkexec nor sudo initialize environ[] from the PAM environment?
921
922 * fedora: update policy to declare access mode and ownership of unit files to root:root 0644, and add an rpmlint check for it
923
924 * register catalog database signature as file magic
925
926 * zsh shell completion:
927 - <command> <verb> -<TAB> should complete options, but currently does not
928 - systemctl add-wants,add-requires
929
930
931 Regularly:
932
933 * look for close() vs. close_nointr() vs. close_nointr_nofail()
934
935 * check for strerror(r) instead of strerror(-r)
936
937 * Use PR_SET_PROCTITLE_AREA if it becomes available in the kernel
938
939 * pahole
940
941 * set_put(), hashmap_put() return values check. i.e. == 0 does not free()!
942
943 * use secure_getenv() instead of getenv() where appropriate
944
945 * link up selected blog stories from man pages and unit files Documentation= fields