]> git.ipfire.org Git - thirdparty/systemd.git/blob - TODO
Merge pull request #15934 from keszybz/docs-and-test-fix
[thirdparty/systemd.git] / TODO
1 Bugfixes:
2
3 * Many manager configuration settings that are only applicable to user
4 manager or system manager can be always set. It would be better to reject
5 them when parsing config.
6
7 External:
8
9 * Fedora: add an rpmlint check that verifies that all unit files in the RPM are listed in %systemd_post macros.
10
11 * wiki: update journal format documentation for lz4 additions
12
13 Janitorial Clean-ups:
14
15 * Rearrange tests so that the various test-xyz.c match a specific src/basic/xyz.c again
16
17 * rework mount.c and swap.c to follow proper state enumeration/deserialization
18 semantics, like we do for device.c now
19
20 Features:
21
22 * if /usr/bin/swapoff fails due to OOM, log a friendly explanatory message about it
23
24 * build short web pages out of each catalog entry, build them along with man
25 pages, and include hyperlinks to them in the journal output
26
27 * machined: add API to acquire UID range. add API to mount/dissect loopback
28 file. Both protected by PK. Then make nspawn use these APIs to run
29 unprivileged containers. i.e. push the truly privileged bits into machined,
30 so that the client side can remain entirely unprivileged, with SUID or
31 anything like that.
32
33 * add "throttling" to sd-event event sources: optionally, when we wake up too
34 often for one, let's turn it off entirely for a while. Use that for the
35 /proc/self/mountinfo logic.
36
37 * move our systemd-user PAM snippet to /usr/, which PAM appears to support
38 these days
39
40 * nspawn: support time namespaces
41
42 * add ConditionSecurity=tpm2
43
44 * pid1: Move to tracking of main pid/control pid of units per pidfd
45
46 * pid1: support new clone3() fork-into-cgroup feature
47
48 * pid1: also remove PID files of a service when the service starts, not just
49 when it exits
50
51 * make "systemd-dissect" an official supported tool, i.e. move to /usr/bin/ and
52 provide man page. Given that we now have a tool that can generate images like
53 this, it's useful to have one that can dump contents of them, too.
54
55 * All tools that support --root= should also learn --image= so that they can
56 operate on disk images directly. Specifically: bootctl, firstboot, tmpfiles,
57 sysusers, systemctl, repart, journalctl, coredumpctl.
58
59 * per-service credential system. Specifically: add LoadCredential= (for loading
60 cred from file), AcquireCredential= (for asking user for cred, via
61 ask-password), PassCredential= (for passing on credential systemd itself
62 got). Then, place credentials in a per-service, immutable ramfs instance (so
63 that it cannot be swapped out), destroy after use. Also pass via keyring
64 (with graceful fallback to cover for containers). Define CredentialPath= for
65 defining subdir of /run/credentials/ where to place it. Set $CREDENTIAL_PATH
66 env var for services to the result. Also pass via fd passing (optionally).
67
68 * homed: add native recovery key support. use 48 lowercase modhex characters
69 (192bit), show qr code of it, include pattern expression in user record.
70
71 * homed: introduce "degraded" state for home directories that weren't cleanly
72 unmounted (use xattr we add and remove on the loop back file)
73
74 * homed: during login resize fs automatically towards size goal. Specifically,
75 resize to diskSize if possible, but leave a certain amount (configured by a
76 new value diskLeaveFreeSize) of space free on the backing fs.
77
78 * homed: permit multiple user record signing keys to be used locally, and pick
79 the right one for signing records automatically depending on a pre-existing
80 signature
81
82 * homed: add a way to "adopt" a home directory, i.e. strip foreign signatures
83 and insert a local signature instead.
84
85 * homed: as an extension to the directory+subvolume backend: if located on
86 especially marked fs, then sync down password into LUKS header of that fs,
87 and always verify passwords against it too. Bootstrapping is a problem
88 though: if no one is logged in (or no other user even exists yet), how do you
89 unlock the volume in order to create the first user and add the first pw.
90
91 * homed: support new FS_IOC_ADD_ENCRYPTION_KEY ioctl for setting up fscrypt
92
93 * busctl: maybe expose a verb "ping" for pinging a dbus service to see if it
94 exists and responds.
95
96 * homed: maybe pre-create ~/.cache as subvol so that it can have separate quota
97 easily?
98
99 * when systemd-nspawn and suchlike dissect an OS image, and there are multiple
100 root partitions, do an strverscmp() on the partition label and boot
101 first. That is inspired how sd-boot figures out which kernel to boot, and
102 thus allows defining OS images which can be A/B updated and we default to the
103 newest version automatically, both in nspawn and in sd-boot
104
105 * cryptsetup/homed: also support FIDO2 HMAC password logic for unlocking
106 devices. (see: https://github.com/mjec/fido2-hmac-secret)
107
108 * systemd-gpt-auto should probably set x-systemd.growfs on the mounts it
109 creates
110
111 * homed/userdb: distinguish passwords and recovery keys in the records, since
112 we probably want to use different PBKDF algorithms/settings for them:
113 passwords have low entropy but recovery keys should have good entropy key
114 hence we can make them quicker to work.
115
116 * bootctl:
117 - teach it to prepare an ESP wholesale, i.e. with mkfs.vfat invocation
118 - teach it to copy in unified kernel images and maybe type #1 boot loader spec entries from host
119 - make it operate on loopback files, dissecting enough to find ESP to operate on
120
121 * by default, in systemd --user service bump the OOMAdjust to 100, as privs
122 allow so that systemd survives
123
124 * cryptsetup: allow encoding key directly in /etc/crypttab, maybe with a
125 "base64:" prefix. Useful in particular for pkcs11 mode.
126
127 * cryptsetup: reimplement the mkswap/mke2fs in cryptsetup-generator to use
128 systemd-makefs.service instead.
129
130 * socket units: allow creating a udev monitor socket with ListenDevices= or so,
131 with matches, then activate app through that passing socket over
132
133 * unify on openssl:
134 - port sd_id128_get_machine_app_specific() over from khash
135 - port resolved over from libgcrypt (DNSSEC code)
136 - port journald + fsprg over from libgcrypt
137 - port importd over from libgcrypt
138 - when that's done: kill khash.c
139 - when that's done: kill gnutls support in resolved
140
141 * kill zenata, all hail weblate?
142
143 * when we resize disks (homed?) always round up to 4K sectors, not 512K
144
145 * add growvol and makevol options for /etc/crypttab, similar to
146 x-systemd.growfs and x-systemd-makefs.
147
148 * hook up the TPM to /etc/crypttab, with a new option that is similar to the
149 new PKCS#11 option in crypttab, and allows unlocking a LUKS volume via a key
150 unsealed from the TPM. Optionally, if TPM is not available fall back to
151 TPM-less mode, and set up linear DM mapping instead (inspired by kpartx), so
152 that the device paths stay the same, regardless if crypto is used or not.
153
154 * systemd-repart: by default generate minimized partition tables (i.e. tables
155 that only cover the space actually used, excluding any free space at the
156 end), in order to maximize dd'ability. Requires libfdisk work, see
157 https://github.com/karelzak/util-linux/issues/907
158
159 * systemd-repart: optionally, allow specifying a path to initialize new
160 partitions from, i.e. an fs image file or a source device node. This would
161 then turn systemd-repart into a simple installer: with a few .repart files
162 you could replicate the host system on another device. a full installer would
163 then be: "systemd-repart /dev/sda && bootctl install /dev/sda &&
164 systemd-firstboot --image= …"
165
166 * systemd-repart: MBR partition table support. Care needs to be taken regarding
167 Type=, so that partition definitions can sanely apply to both the GPT and the
168 MBR case. Idea: accept syntax "Type=gpt:home mbr:0x83" for setting the types
169 for the two partition types explicitly. And provide an internal mapping so
170 that "Type=linux-generic" maps to the right types for both partition tables
171 automatically.
172
173 * systemd-repart: allow sizing partitions as factor of available RAM, so that
174 we can reasonably size swap partitions for hibernation.
175
176 * systemd-repart: allow running mkfs before making partitions pop up +
177 encryption via LUKS to allow booting into an empty root with only /usr mounted in
178
179 * systemd-repart: allow managing the gpt read-only partition flag + auto-mount flag
180
181 * systemd-repart: allow disabling growing of specific partitions, or making
182 them (think ESP: we don't ever want to grow it, since we cannot resize vfat)
183
184 * systemd-repart: add specifier expansion, add especifier that refers to root
185 device node of current system, /usr device node, and matching verity, so that
186 an installer can be made a "copy" installer of the booted OS
187
188 * systemd-repart: make it a static checker during early boot for existence and
189 absence of other partitions for trusted boot environments
190
191 * systemd-repart: allow config of partition uuid
192
193 * userdb: allow username prefix searches in varlink API, allow realname and
194 realname substr searches in varlink API
195
196 * userdb: allow uid/gid range checks
197
198 * userdb: allow existence checks
199
200 * pid1: activation by journal search expression
201
202 * when switching root from initrd to host, set the machine_id env var so that
203 if the host has no machine ID set yet we continue to use the random one the
204 initrd had set.
205
206 * sd-event: add native support for P_ALL waitid() watching, then move PID 1 to
207 it fo reaping assigned but unknown children. This needs to some special care
208 to operate somewhat sensibly in light of priorities: P_ALL will return
209 arbitrary processes, regardless of the priority we want to watch them with,
210 hence on each event loop iteration check all processes which we shall watch
211 with higher prio explicitly, and then watch the entire rest with P_ALL.
212
213 * tweak sd-event's child watching: keep a prioq of children to watch and use
214 waitid() only on the children with the highest priority until one is waitable
215 and ignore all lower-prio ones from that point on
216
217 * maybe introduce xattrs that can be set on the root dir of the root fs
218 partition that declare the volatility mode to use the image in. Previously I
219 thought marking this via GPT partition flags but that's not ideal since
220 that's outside of the LUKS encryption/verity verification, and we probably
221 shouldn't operate in a volatile mode unless we got told so from a trusted
222 source.
223
224 * figure out automatic partition discovery when combining writable root dir
225 with immutable /usr
226
227 * coredump: maybe when coredumping read a new xattr from /proc/$PID/exe that
228 may be used to mark a whole binary as non-coredumpable. Would fix:
229 https://bugs.freedesktop.org/show_bug.cgi?id=69447
230
231 * teach parse_timestamp() timezones like the calendar spec already knows it
232
233 * beef up hibernation to optionally do swapon/swapoff immediately before/after
234 the hibernation
235
236 * beef up s2h to implement a battery watch loop: instead of entering
237 hibernation unconditionally after coming back from resume make a decision
238 based on the battery load level: if battery level is above a specific
239 threshold, go to suspend again, only hibernate if below it. This means we'd
240 stick to suspend usually, but fall back to hibernation only when battery runs
241 empty (well, subject to our sampling interval). Related to this, check if we
242 can make ACPI _BTP (i.e. /sys/class/power_supply/*/alarm) work for us too,
243 i.e. see if it can wake up machines from suspend, so that we could resume
244 automatically when the system is low on power and move automatically to
245 hibernation mode. (see
246 https://uefi.org/sites/default/files/resources/ACPI%206_2_A_Sept29.pdf
247 section 10.2.2.8 and
248 https://docs.microsoft.com/en-us/windows-hardware/design/device-experiences/modern-standby-wake-sources
249 at the end).
250
251 * add an explicit "vertical" mode to format-table, so that "systemctl
252 status"-like outputs (i.e. with a series of field names left and values
253 right) become genuine first class citizens, and we gain automatic, sane JSON
254 output for them.
255
256 * systemd-firstboot: teach it dissector magic, so that you can point it to some
257 disk image and it will just set everything in it all behind the scenes.
258
259 * systemd-firstboot: add --force mode that replaces existing configuration.
260
261 * We should probably replace /var/log/README, /etc/rc.d/README with symlinks
262 that are linked to these places instead of copied. After all they are
263 constant vendor data.
264
265 * maybe add kernel cmdline params: to force random seed crediting
266
267 * nspawn: on cgroupsv1 issue cgroup empty handler process based on host events,
268 so that we make cgroup agent logic safe
269
270 * nspawn/machined: add API to invoke binary in container, then use that as
271 fallback in "machinectl shell"
272
273 * logind: rework pam_logind to also do a bus call in case of invocation from
274 user@.service, which returns the XDG_RUNTIME_DIR value, and make this
275 behaviour selectable via pam module option.
276
277 * homed:
278 - when user tries to log into record signed by unrecognized key, automatically add key to our chain after polkit auth
279 - hook up machined/nspawn users with a varlink user query interface
280 - rollback when resize fails mid-operation
281 - GNOME's side for forget key on suspend (requires rework so that lock screen runs outside of uid)
282 - resize on login?
283 - shrink fs on logout?
284 - update LUKS password on login if we find there's a password that unlocks the JSON record but not the LUKS device.
285 - create on activate?
286 - properties: icon url?, preferred session type?, administrator bool (which translates to 'wheel' membership)?, address?, telephone?, vcard?, samba stuff?, parental controls?
287 - communicate clearly when usb stick is safe to remove. probably involves
288 beefing up logind to make pam session close hook synchronous and wait until
289 systemd --user is shut down.
290 - logind: maybe keep a "busy fd" as long as there's a non-released session around or the user@.service
291 - maybe make automatic, read-only, time-based reflink-copies of LUKS disk images (think: time machine)
292 - distinguish destroy / remove (i.e. currently we can unregister a user, unregister+remove their home directory, but not just remove their home directory)
293 - in systemd's PAMName= logic: query passwords with ssh-askpassword, so that we can make "loginctl set-linger" mode work
294 - fingerprint authentication, pattern authentication, …
295 - make sure "classic" user records can also be managed by homed
296 - description field for groups
297 - make size of $XDG_RUNTIME_DIR configurable in user record
298 - reuse pwquality magic in firstboot
299 - query password from kernel keyring first
300 - update even if record is "absent"
301 - add a "access mode" + "fstype" field to the "status" section of json identity records reflecting the actually used access mode and fstype, even on non-luks backends
302 - move acct mgmt stuff from pam_systemd_home to pam_systemd?
303 - when "homectl --pkcs11-token-uri=" is used, synthesize ssh-authorized-keys records for all keys we have private keys on the stick for
304 - make slice for users configurable (requires logind rework)
305 - logind: populate auto-login list bus property from PKCS#11 token
306 - when determining state of a LUKS home directory, check DM suspended sysfs file
307
308 * introduce a new per-process uuid, similar to the boot id, the machine id, the
309 invocation id, that is derived from process creds, specifically a hashed
310 combination of AT_RANDOM + getpid() + the starttime from
311 /proc/self/status. Then add these ids implicitly when logging. Deriving this
312 uuid from these three things has the benefit that it can be derived easily
313 from /proc/$PID/ in a stable, and unique way that changes on both fork() and
314 exec().
315
316 * let's not GC a unit while its ratelimits are still pending
317
318 * when killing due to service watchdog timeout maybe detect whether target
319 process is under ptracing and then log loudly and continue instead.
320
321 * introduce a new group to own TPM devices
322
323 * make rfkill uaccess controllable by default, i.e. steal rule from
324 gnome-bluetooth and friends
325
326 * tweak journald context caching. In addition to caching per-process attributes
327 keyed by PID, cache per-cgroup attributes (i.e. the various xattrs we read)
328 keyed by cgroup path, and guarded by ctime changes. This should provide us
329 with a nice speed-up on services that have many processes running in the same
330 cgroup.
331
332 * make MAINPID= message reception checks even stricter: if service uses User=,
333 then check sending UID and ignore message if it doesn't match the user or
334 root.
335
336 * maybe trigger a uevent "change" on a device if "systemctl reload xyz.device"
337 is issued.
338
339 * when importing an fs tree with machined, optionally apply userns-rec-chown
340
341 * when importing an fs tree with machined, complain if image is not an OS
342
343 * Maybe introduce a helper safe_exec() or so, which is to execve() which
344 safe_fork() is to fork(). And then make revert the RLIMIT_NOFILE soft limit
345 to 1K implicitly, unless explicitly opted-out.
346
347 * rework seccomp/nnp logic that even if User= is used in combination with
348 a seccomp option we don't have to set NNP. For that, change uid first whil
349 keeping CAP_SYS_ADMIN, then apply seccomp, the drop cap.
350
351 * add a concept for automatically loading per-unit secrets off disk and
352 inserting them into the kernel keyring. Maybe SecretsDirectory= similar to
353 ConfigurationDirectory=.
354
355 * when no locale is configured, default to UEFI's PlatformLang variable
356
357 * bootctl,sd-boot: actually honour the "architecture" key
358
359 * add a new syscall group "@esoteric" for more esoteric stuff such as bpf() and
360 usefaultd() and make systemd-analyze check for it.
361
362 * paranoia: whenever we process passwords, call mlock() on the memory
363 first. i.e. look for all places we use free_and_erasep() and
364 augment them with mlock(). Also use MADV_DONTDUMP.
365
366 * Move RestrictAddressFamily= to the new cgroup create socket
367
368 * support the bind/connect/sendmsg cgroup stuff for sandboxing, and possibly
369 patching around
370
371 * maybe implicitly attach monotonic+realtime timestamps to outgoing messages in
372 log.c and sd-journal-send
373
374 * optionally: turn on cgroup delegation for per-session scope units
375
376 * introduce per-unit (i.e. per-slice, per-service) journal log size limits.
377
378 * sd-boot: automatically load EFI modules from some drop-in dir, so that people
379 can add in file system drivers and such
380
381 * sd-boot: optionally, show boot menu when previous default boot item has
382 non-zero "tries done" count
383
384 * introduce an option (or replacement) for "systemctl show" that outputs all
385 properties as JSON, similar to busctl's new JSON output. In contrast to that
386 it should skip the variant type string though.
387
388 * augment CODE_FILE=, CODE_LINE= with something like CODE_BASE= or so which
389 contains some identifier for the project, which allows us to include
390 clickable links to source files generating these log messages. The identifier
391 could be some abberviated URL prefix or so (taking inspiration from Go
392 imports). For example, for systemd we could use
393 CODE_BASE=github.com/systemd/systemd/blob/98b0b1123cc or so which is
394 sufficient to build a link by prefixing "http://" and suffixing the
395 CODE_FILE.
396
397 * Augment MESSAGE_ID with MESSAGE_BASE, in a similar fashion so that we can
398 make clickable links from log messages carrying a MESSAGE_ID, that lead to
399 some explanatory text online.
400
401 * maybe extend .path units to expose fanotify() per-mount change events
402
403 * Add a "systemctl list-units --by-slice" mode or so, which rearranges the
404 output of "systemctl list-units" slightly by showing the tree structure of
405 the slices, and the units attached to them.
406
407 * the a-posteriori stopping of units bound to units that disappeared logic
408 should be reworked: there should be a queue of units, and we should only
409 enqueue stop jobs from a defer event that processes queue instead of
410 right-away when we find a unit that is bound to one that doesn't exist
411 anymore. (similar to how the stop-unneeded queue has been reworked the same
412 way)
413
414 * nspawn: make nspawn suitable for shell pipelines: instead of triggering a
415 hangup when input is finished, send ^D, which synthesizes an EOF. Then wait
416 for hangup or ^D before passing on the EOF.
417
418 * When reloading configuration PID 1 should reset all its properties to the
419 original defaults before calling parse_config()
420
421 * nspawn: greater control over selinux label?
422
423 * hibernate/s2h: make this robust and safe to enable in Fedora by default.
424 Specifically:
425
426 1. add resume_offset support to the resume code (i.e. support swap files
427 properly)
428 2. check if swap is on weird storage and refuse if so
429 3. add auto-detection of hibernation images
430
431 * cgroups: use inotify to get notified when somebody else modifies cgroups
432 owned by us, then log a friendly warning.
433
434 * beef up log.c with support for stripping ANSI sequences from strings, so that
435 it is OK to include them in log strings. This would be particularly useful so
436 that our log messages could contain clickable links for example for unit
437 files and suchlike we operate on.
438
439 * add support for "portablectl attach http://foobar.com/waaa.raw (i.e. importd integration)
440
441 * add attach --enable and attach --now (for attach+enable+start)
442
443 * sync dynamic uids/gids between host+portable srvice (i.e. if DynamicUser=1 is set for a service, make sure that the
444 selected user is resolvable in the service even if it ships its own /etc/passwd)
445
446 * Fix DECIMAL_STR_MAX or DECIMAL_STR_WIDTH. One includes a trailing NUL, the
447 other doesn't. What a disaster. Probably to exclude it. Also
448 DECIMAL_STR_WIDTH should probably add an extra "-" into account for negative
449 numbers.
450
451 * Check that users of inotify's IN_DELETE_SELF flag are using it properly, as
452 usually IN_ATTRIB is the right way to watch deleted files, as the former only
453 fires when a file is actually removed from disk, i.e. the link count drops to
454 zero and is not open anymore, while the latter happens when a file is
455 unlinked from any dir.
456
457 * port systemctl, busctl, … over to format-table.[ch]'s table formatters
458
459 * pid1: lock image configured with RootDirectory=/RootImage= using the usual nspawn semantics while the unit is up
460
461 * add --vacuum-xyz options to coredumpctl, matching those journalctl already has.
462
463 * introduce Ephemeral= unit file switch, that creates an ephemeral copy of all
464 files and directories that are left writable for a unit, and which are
465 removed after the unit goes down again. A bit like --ephemeral for
466 systemd-nspawn but for system services. If used together with RootImage= this
467 should reflink the image file itself.
468
469 Related: add Ephemeral=<path1> <path2> … which would allow marking
470 specific paths only like this.
471
472 * add CopyFile= or so as unit file setting that may be used to copy files or
473 directory trees from the host to the services RootImage= and RootDirectory=
474 environment. Which we can use for /etc/machine-id and in particular
475 /etc/resolv.conf. Should be smart and do something useful on read-only
476 images, for example fallback to read-only bind mounting the file instead.
477
478 * show invocation ID in systemd-run output
479
480 * bypass SIGTERM state in unit files if KillSignal is SIGKILL
481
482 * add proper dbus APIs for the various sd_notify() commands, such as MAINPID=1
483 and so on, which would mean we could report errors and such.
484
485 * teach tmpfiles.d q/Q logic something sensible in the context of XFS/ext4
486 project quota
487
488 * introduce DefaultSlice= or so in system.conf that allows changing where we
489 place our units by default, i.e. change system.slice to something
490 else. Similar, ManagerSlice= should exist so that PID1's own scope unit could
491 be moved somewhere else too. Finally machined and logind should get similar
492 options so that it is possible to move user session scopes and machines to a
493 different slice too by default. Usecase: people who want to put resources on
494 the entire system, with the exception of one specific service. See:
495 https://lists.freedesktop.org/archives/systemd-devel/2018-February/040369.html
496
497 * maybe rework get_user_creds() to query the user database if $SHELL is used
498 for root, but only then.
499
500 * be stricter with fds we receive for the fdstore: close them asynchronously
501
502 * calenderspec: add support for week numbers and day numbers within a
503 year. This would allow us to define "bi-weekly" triggers safely.
504
505 * sd-bus: add vtable flag, that may be used to request client creds implicitly
506 and asynchronously before dispatching the operation
507
508 * make use of ethtool veth peer info in machined, for automatically finding out
509 host-side interface pointing to the container.
510
511 * add some special mode to LogsDirectory=/StateDirectory=… that allows
512 declaring these directories without necessarily pulling in deps for them, or
513 creating them when starting up. That way, we could declare that
514 systemd-journald writes to /var/log/journal, which could be useful when we
515 doing disk usage calculations and so on.
516
517 * taint systemd if there are fewer than 65536 users assigned (userns) to the system.
518
519 * deprecate RootDirectoryStartOnly= in favour of a new ExecStart= prefix char
520
521 * add a new RuntimeDirectoryPreserve= mode that defines a similar lifecycle for
522 the runtime dir as we maintain for the fdstore: i.e. keep it around as long
523 as the unit is running or has a job queued.
524
525 * support projid-based quota in machinectl for containers
526
527 * add a way to lock down cgroup migration: a boolean, which when set for a unit
528 makes sure the processes in it can never migrate out of it
529
530 * blog about fd store and restartable services
531
532 * document Environment=SYSTEMD_LOG_LEVEL=debug drop-in in debugging document
533
534 * rework ExecOutput and ExecInput enums so that EXEC_OUTPUT_NULL loses its
535 magic meaning and is no longer upgraded to something else if set explicitly.
536
537 * in the long run: permit a system with /etc/machine-id linked to /dev/null, to
538 make it lose its identity, i.e. be anonymous. For this we'd have to patch
539 through the whole tree to make all code deal with the case where no machine
540 ID is available.
541
542 * optionally, collect cgroup resource data, and store it in per-unit RRD files,
543 suitable for processing with rrdtool. Add bus API to access this data, and
544 possibly implement a CPULoad property based on it.
545
546 * beef up pam_systemd to take unit file settings such as cgroups properties as
547 parameters
548
549 * maybe hook of xfs/ext4 quotactl() with services? i.e. automatically manage
550 the quota of a the user indicated in User= via unit file settings, like the
551 other resource management concepts. Would mix nicely with DynamicUser=1. Or
552 alternatively, do this with projids, so that we can also cover services
553 running as root. Quota should probably cover all the special dirs such as
554 StateDirectory=, LogsDirectory=, CacheDirectory=, as well as RootDirectory= if it
555 is set, plus the whole disk space any image configured with RootImage=.
556
557 * In DynamicUser= mode: before selecting a UID, use disk quota APIs on relevant
558 disks to see if the UID is already in use.
559
560 * add "systemctl wait" or so, which does what "systemd-run --wait" does, but
561 for all units. It should be both a way to pin units into memory as well as a
562 wait to retrieve their exit data.
563
564 * expose IO accounting data on the bus, show it in systemd-run --wait and log
565 about it in the resource log message
566
567 * show whether a service has out-of-date configuration in "systemctl status" by
568 using mtime data of ConfigurationDirectory=.
569
570 * Add AddUser= setting to unit files, similar to DynamicUser=1 which however
571 creates a static, persistent user rather than a dynamic, transient user. We
572 can leverage code from sysusers.d for this.
573
574 * add some optional flag to ReadWritePaths= and friends, that has the effect
575 that we create the dir in question when the service is started. Example:
576
577 ReadWritePaths=:/var/lib/foobar
578
579 * maybe add call sd_journal_set_block_timeout() or so to set SO_SNDTIMEO for
580 the sd-journal logging socket, and, if the timeout is set to 0, sets
581 O_NONBLOCK on it. That way people can control if and when to block for
582 logging.
583
584 * hostnamed: populate form factor data from a new hwdb database, so that old
585 yogas can be recognized as "convertible" too, even if they predate the DMI
586 "convertible" form factor
587
588 * Maybe add PrivatePIDs= as new unit setting, and do minimal PID namespacing
589 after all. Be strict however, only support the equivalent of nspawn's
590 --as-pid2 switch, and sanely proxy sd_notify() messages dropping stuff such
591 as MAINPID.
592
593 * Add ExecMonitor= setting. May be used multiple times. Forks off a process in
594 the service cgroup, which is supposed to monitor the service, and when it
595 exits the service is considered failed by its monitor.
596
597 * track the per-service PAM process properly (i.e. as an additional control
598 process), so that it may be queried on the bus and everything.
599
600 * add a new "debug" job mode, that is propagated to unit_start() and for
601 services results in two things: we raise SIGSTOP right before invoking
602 execve() and turn off watchdog support. Then, use that to implement
603 "systemd-gdb" for attaching to the start-up of any system service in its
604 natural habitat.
605
606 * gpt-auto logic: related to the above, maybe support a "secondary" root
607 partition, that is mounted to / and is writable, and where the actual root's
608 /usr is mounted into.
609
610 * 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
611
612 * drop nss-myhostname in favour of nss-resolve?
613
614 * add a percentage syntax for TimeoutStopSec=, e.g. TimeoutStopSec=150%, and
615 then use that for the setting used in user@.service. It should be understood
616 relative to the configured default value.
617
618 * in networkd, when matching device types, fix up DEVTYPE rubbish the kernel passes to us
619
620 * enable LockMLOCK to take a percentage value relative to physical memory
621
622 * Permit masking specific netlink APIs with RestrictAddressFamily=
623
624 * nspawn: support that /proc, /sys/, /dev are pre-mounted
625
626 * define gpt header bits to select volatility mode
627
628 * ProtectClock= (drops CAP_SYS_TIMES, adds seecomp filters for settimeofday, adjtimex), sets DeviceAllow o /dev/rtc
629
630 * ProtectTracing= (drops CAP_SYS_PTRACE, blocks ptrace syscall, makes /sys/kernel/tracing go away)
631
632 * ProtectMount= (drop mount/umount/pivot_root from seccomp, disallow fuse via DeviceAllow, imply Mountflags=slave)
633
634 * ProtectKeyRing= to take keyring calls away
635
636 * RemoveKeyRing= to remove all keyring entries of the specified user
637
638 * ProtectReboot= that masks reboot() and kexec_load() syscalls, prohibits kill
639 on PID 1 with the relevant signals, and makes relevant files in /sys and
640 /proc (such as the sysrq stuff) unavailable
641
642 * make sure the ratelimit object can deal with USEC_INFINITY as way to turn off things
643
644 * journalctl: make sure -f ends when the container indicated by -M terminates
645
646 * mount: automatically search for "main" partition of an image has multiple
647 partitions
648
649 * in nss-systemd, if we run inside of RootDirectory= with PrivateUsers= set,
650 find a way to map the User=/Group= of the service to the right name. This way
651 a user/group for a service only has to exist on the host for the right
652 mapping to work.
653
654 * add bus API for creating unit files in /etc, reusing the code for transient units
655
656 * add bus API to remove unit files from /etc
657
658 * add bus API to retrieve current unit file contents (i.e. implement "systemctl cat" on the bus only)
659
660 * rework fopen_temporary() to make use of open_tmpfile_linkable() (problem: the
661 kernel doesn't support linkat() that replaces existing files, currently)
662
663 * transient units: don't bother with actually setting unit properties, we
664 reload the unit file anyway
665
666 * journald: sigbus API via a signal-handler safe function that people may call
667 from the SIGBUS handler
668
669 * optionally, also require WATCHDOG=1 notifications during service start-up and shutdown
670
671 * delay activation of logind until somebody logs in, or when /dev/tty0 pulls it
672 in or lingering is on (so that containers don't bother with it until PAM is used). also exit-on-idle
673
674 * cache sd_event_now() result from before the first iteration...
675
676 * PID1: find a way how we can reload unit file configuration for
677 specific units only, without reloading the whole of systemd
678
679 * add an explicit parser for LimitRTPRIO= that verifies
680 the specified range and generates sane error messages for incorrect
681 specifications.
682
683 * when we detect that there are waiting jobs but no running jobs, do something
684
685 * push CPUAffinity= also into the "cpuset" cgroup controller
686
687 * PID 1 should send out sd_notify("WATCHDOG=1") messages (for usage in the --user mode, and when run via nspawn)
688
689 * there's probably something wrong with having user mounts below /sys,
690 as we have for debugfs. for example, src/core/mount.c handles mounts
691 prefixed with /sys generally special.
692 http://lists.freedesktop.org/archives/systemd-devel/2015-June/032962.html
693
694 * fstab-generator: default to tmpfs-as-root if only usr= is specified on the kernel cmdline
695
696 * docs: bring http://www.freedesktop.org/wiki/Software/systemd/MyServiceCantGetRealtime up to date
697
698 * add a job mode that will fail if a transaction would mean stopping
699 running units. Use this in timedated to manage the NTP service
700 state.
701 http://lists.freedesktop.org/archives/systemd-devel/2015-April/030229.html
702
703 * The udev blkid built-in should expose a property that reflects
704 whether media was sensed in USB CF/SD card readers. This should then
705 be used to control SYSTEMD_READY=1/0 so that USB card readers aren't
706 picked up by systemd unless they contain a medium. This would mirror
707 the behaviour we already have for CD drives.
708
709 * networkd/udev: implement SR_IOV configuration in .link files:
710 http://lists.freedesktop.org/archives/systemd-devel/2015-January/027451.html
711
712 * hostnamectl: show root image uuid
713
714 * Find a solution for SMACK capabilities stuff:
715 http://lists.freedesktop.org/archives/systemd-devel/2014-December/026188.html
716
717 * "systemctl preset-all" should probably order the unit files it
718 operates on lexicographically before starting to work, in order to
719 ensure deterministic behaviour if two unit files conflict (like DMs
720 do, for example)
721
722 * synchronize console access with BSD locks:
723 http://lists.freedesktop.org/archives/systemd-devel/2014-October/024582.html
724
725 * as soon as we have sender timestamps, revisit coalescing multiple parallel daemon reloads:
726 http://lists.freedesktop.org/archives/systemd-devel/2014-December/025862.html
727
728 * figure out when we can use the coarse timers
729
730 * add "systemctl start -v foobar.service" that shows logs of a service
731 while the start command runs. This is non-trivial to do without
732 races though, since we should flush out all journal messages before
733 returning from the "systemctl stop".
734
735 * 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
736
737 * maybe add support for specifier expansion in user.conf, specifically DefaultEnvironment=
738
739 * maybe allow timer units with an empty Units= setting, so that they
740 can be used for resuming the system but nothing else.
741
742 * what to do about udev db binary stability for apps? (raw access is not an option)
743
744 * systemctl: if some operation fails, show log output?
745
746 * systemctl edit: use equivalent of cat() to insert existing config as a comment, prepended with #.
747 Upon editor exit, lines with one # are removed, lines with two # are left with one #, etc.
748
749 * exponential backoff in timesyncd when we cannot reach a server
750
751 * timesyncd: add ugly bus calls to set NTP servers per-interface, for usage by NM
752
753 * merge ~/.local/share and ~/.local/lib into one similar /usr/lib and /usr/share....
754
755 * add systemd.abort_on_kill or some other such flag to send SIGABRT instead of SIGKILL
756 (throughout the codebase, not only PID1)
757
758 * resolved:
759 - mDNS/DNS-SD
760 - service registration
761 - service/domain/types browsing
762 - avahi compat
763 - DNS-SD service registration from socket units
764 - resolved should optionally register additional per-interface LLMNR
765 names, so that for the container case we can establish the same name
766 (maybe "host") for referencing the server, everywhere.
767 - allow clients to request DNSSEC for a single lookup even if DNSSEC is off (?)
768 - hook up resolved with machined-based address resolution
769
770 * refcounting in sd-resolve is borked
771
772 * Add a new verb "systemctl top"
773
774 * add new gpt type for btrfs volumes
775
776 * support empty /etc boots nicely:
777 - nspawn/gpt-generator: introduce new gpt partition type for /usr
778
779 * generator that automatically discovers btrfs subvolumes, identifies their purpose based on some xattr on them.
780
781 * a way for container managers to turn off getty starting via $container_headless= or so...
782
783 * figure out a nice way how we can let the admin know what child/sibling unit causes cgroup membership for a specific unit
784
785 * For timer units: add some mechanisms so that timer units that trigger immediately on boot do not have the services
786 they run added to the initial transaction and thus confuse Type=idle.
787
788 * add bus api to query unit file's X fields.
789
790 * gpt-auto-generator:
791 - Define new partition type for encrypted swap? Support probed LUKS for encrypted swap?
792 - Make /home automount rather than mount?
793
794 * add generator that pulls in systemd-network from containers when
795 CAP_NET_ADMIN is set, more than the loopback device is defined, even
796 when it is otherwise off
797
798 * MessageQueueMessageSize= (and suchlike) should use parse_iec_size().
799
800 * implement Distribute= in socket units to allow running multiple
801 service instances processing the listening socket, and open this up
802 for ReusePort=
803
804 * introduce bus call FreezeUnit(s, b), as well as "systemctl freeze
805 $UNIT" and "systemctl thaw $UNIT" as wrappers around this. The calls
806 should SIGSTOP all unit processes in a loop until all processes of
807 it are fully stopped. This can later be used for app management by
808 desktop UIs such as gnome-shell to freeze apps that are not visible
809 on screen, not unlike how job control works on the shell
810
811 * cgroups:
812 - implement per-slice CPUFairScheduling=1 switch
813 - introduce high-level settings for RT budget, swappiness
814 - how to reset dynamically changed unit cgroup attributes sanely?
815 - when reloading configuration, apply new cgroup configuration
816 - when recursively showing the cgroup hierarchy, optionally also show
817 the hierarchies of child processes
818
819 * transient units:
820 - add field to transient units that indicate whether systemd or somebody else saves/restores its settings, for integration with libvirt
821
822 * when we detect low battery and no AC on boot, show pretty splash and refuse boot
823
824 * libsystemd-journal, libsystemd-login, libudev: add calls to easily attach these objects to sd-event event loops
825
826 * be more careful what we export on the bus as (usec_t) 0 and (usec_t) -1
827
828 * 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
829
830 * After coming back from hibernation reset hibernation swap partition using the /dev/snapshot ioctl APIs
831
832 * If we try to find a unit via a dangling symlink, generate a clean
833 error. Currently, we just ignore it and read the unit from the search
834 path anyway.
835
836 * refuse boot if /usr/lib/os-release is missing or /etc/machine-id cannot be set up
837
838 * man: the documentation of Restart= currently is very misleading and suggests the tools from ExecStartPre= might get restarted.
839
840 * load .d/*.conf dropins for device units
841
842 * sd-bus:
843 - EBADSLT handling
844 - GetAllProperties() on a non-existing object does not result in a failure currently
845 - port to sd-resolve for connecting to TCP dbus servers
846 - 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
847 - see if we can drop more message validation on the sending side
848 - add API to clone sd_bus_message objects
849 - longer term: priority inheritance
850 - dbus spec updates:
851 - NameLost/NameAcquired obsolete
852 - GVariant
853 - path escaping
854 - update systemd.special(7) to mention that dbus.socket is only about the compatibility socket now
855
856 * sd-event
857 - allow multiple signal handlers per signal?
858 - document chaining of signal handler for SIGCHLD and child handlers
859 - define more intervals where we will shift wakeup intervals around in, 1h, 6h, 24h, ...
860
861 * investigate endianness issues of UUID vs. GUID
862
863 * dbus: when a unit failed to load (i.e. is in UNIT_ERROR state), we
864 should be able to safely try another attempt when the bus call LoadUnit() is invoked.
865
866 * add a pam module that passes the hdd passphrase into the PAM stack and then expires it, for usage by gdm auto-login.
867
868 * add a pam module that on password changes updates any LUKS slot where the password matches
869
870 * test/:
871 - add unit tests for config_parse_device_allow()
872
873 * seems that when we follow symlinks to units we prefer the symlink
874 destination path over /etc and /usr. We should not do that. Instead
875 /etc should always override /run+/usr and also any symlink
876 destination.
877
878 * when isolating, try to figure out a way how we implicitly can order
879 all units we stop before the isolating unit...
880
881 * teach ConditionKernelCommandLine= globs or regexes (in order to match foobar={no,0,off})
882
883 * Add ConditionDirectoryNotEmpty= handle non-absoute paths as a search path or add
884 ConditionConfigSearchPathNotEmpty= or different syntax? See the discussion starting at
885 https://github.com/systemd/systemd/pull/15109#issuecomment-607740136.
886
887 * BootLoaderSpec: Clarify that the kernel has to be in $BOOT. Clarify
888 that the boot loader should be installed to the ESP. Define a way
889 how an installer can figure out whether a BLS compliant boot loader
890 is installed.
891
892 * think about requeuing jobs when daemon-reload is issued? usecase:
893 the initrd issues a reload after fstab from the host is accessible
894 and we might want to requeue the mounts local-fs acquired through
895 that automatically.
896
897 * systemd-inhibit: make taking delay locks useful: support sending SIGINT or SIGTERM on PrepareForSleep()
898
899 * remove any syslog support from log.c — we probably cannot do this before split-off udev is gone for good
900
901 * shutdown logging: store to EFI var, and store to USB stick?
902
903 * merge unit_kill_common() and unit_kill_context()
904
905 * EFI:
906 - honor language efi variables for default language selection (if there are any?)
907 - honor timezone efi variables for default timezone selection (if there are any?)
908 - change bootctl to be backed by systemd-bootd to control temporary and persistent default boot goal plus efi variables
909 * bootctl
910 - recognize the case when not booted on EFI
911
912 * maybe do not install getty@tty1.service symlink in /etc but in /usr?
913
914 * print a nicer explanation if people use variable/specifier expansion in ExecStart= for the first word
915
916 * mount: turn dependency information from /proc/self/mountinfo into dependency information between systemd units.
917
918 * logind:
919 - logind: optionally, ignore idle-hint logic for autosuspend, block suspend as long as a session is around
920 - logind: wakelock/opportunistic suspend support
921 - Add pretty name for seats in logind
922 - logind: allow showing logout dialog from system?
923 - add Suspend() bus calls which take timestamps to fix double suspend issues when somebody hits suspend and closes laptop quickly.
924 - if pam_systemd is invoked by su from a process that is outside of a
925 any session we should probably just become a NOP, since that's
926 usually not a real user session but just some system code that just
927 needs setuid().
928 - logind: make the Suspend()/Hibernate() bus calls wait for the for
929 the job to be completed. before returning, so that clients can wait
930 for "systemctl suspend" to finish to know when the suspending is
931 complete.
932 - logind: when the power button is pressed short, just popup a
933 logout dialog. If it is pressed for 1s, do the usual
934 shutdown. Inspiration are Macs here.
935 - expose "Locked" property on logind sesison objects
936 - maybe allow configuration of the StopTimeout for session scopes
937 - rename session scope so that it includes the UID. THat way
938 the session scope can be arranged freely in slices and we don't have
939 make assumptions about their slice anymore.
940 - follow PropertiesChanged state more closely, to deal with quick logouts and
941 relogins
942 - (optionally?) spawn seat-manager@$SEAT.service whenever a seat shows up that as CanGraphical set
943
944 * journal:
945 - consider introducing implicit _TTY= + _PPID= + _EUID= + _EGID= + _FSUID= + _FSGID= fields
946 - journald: also get thread ID from client, plus thread name
947 - journal: when waiting for journal additions in the client always sleep at least 1s or so, in order to minimize wakeups
948 - add API to close/reopen/get fd for journal client fd in libsystemd-journal.
949 - fallback to /dev/log based logging in libsystemd-journal, if we cannot log natively?
950 - declare the local journal protocol stable in the wiki interface chart
951 - sd-journal: speed up sd_journal_get_data() with transparent hash table in bg
952 - journald: when dropping msgs due to ratelimit make sure to write
953 "dropped %u messages" not only when we are about to print the next
954 message that works, but already after a short timeout
955 - check if we can make journalctl by default use --follow mode inside of less if called without args?
956 - maybe add API to send pairs of iovecs via sd_journal_send
957 - journal: add a setgid "systemd-journal" utility to invoke from libsystemd-journal, which passes fds via STDOUT and does PK access
958 - journactl: support negative filtering, i.e. FOOBAR!="waldo",
959 and !FOOBAR for events without FOOBAR.
960 - journal: store timestamp of journal_file_set_offline() int he header,
961 so it is possible to display when the file was last synced.
962 - 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.
963 - journal: find a way to allow dropping history early, based on priority, other rules
964 - journal: When used on NFS, check payload hashes
965 - journald: add kernel cmdline option to disable ratelimiting for debug purposes
966 - refuse taking lower-case variable names in sd_journal_send() and friends.
967 - journald: we currently rotate only after MaxUse+MaxFilesize has been reached.
968 - journal: deal nicely with byte-by-byte copied files, especially regards header
969 - journal: sanely deal with entries which are larger than the individual file size, but where the components would fit
970 - Replace utmp, wtmp, btmp, and lastlog completely with journal
971 - journalctl: instead --after-cursor= maybe have a --cursor=XYZ+1 syntax?
972 - when a kernel driver logs in a tight loop, we should ratelimit that too.
973 - journald: optionally, log debug messages to /run but everything else to /var
974 - journald: when we drop syslog messages because the syslog socket is
975 full, make sure to write how many messages are lost as first thing
976 to syslog when it works again.
977 - journald: allow per-priority and per-service retention times when rotating/vacuuming
978 - journald: make use of uid-range.h to managed uid ranges to split
979 journals in.
980 - journalctl: add the ability to look for the most recent process of a binary. journalctl /usr/bin/X11 --pid=-1 or so...
981 - improve journalctl performance by loading journal files
982 lazily. Encode just enough information in the file name, so that we
983 do not have to open it to know that it is not interesting for us, for
984 the most common operations.
985 - man: document that corrupted journal files is nothing to act on
986 - rework journald sigbus stuff to use mutex
987 - Set RLIMIT_NPROC for systemd-journal-xyz, and all other of our
988 services that run under their own user ids, and use User= (but only
989 in a world where userns is ubiquitous since otherwise we cannot
990 invoke those daemons on the host AND in a container anymore). Also,
991 if LimitNPROC= is used without User= we should warn and refuse
992 operation.
993 - journalctl --verify: don't show files that are currently being
994 written to as FAIL, but instead show that their are being written to.
995 - add journalctl -H that talks via ssh to a remote peer and passes through
996 binary logs data
997 - add a version of --merge which also merges /var/log/journal/remote
998 - journalctl: -m should access container journals directly by enumerating
999 them via machined, and also watch containers coming and going.
1000 Benefit: nspawn --ephemeral would start working nicely with the journal.
1001 - assign MESSAGE_ID to log messages about failed services
1002
1003 * add a test if all entries in the catalog are properly formatted.
1004 (Adding dashes in a catalog entry currently results in the catalog entry
1005 being silently skipped. journalctl --update-catalog must warn about this,
1006 and we should also have a unit test to check that all our message are OK.)
1007
1008 * document:
1009 - document that deps in [Unit] sections ignore Alias= fields in
1010 [Install] units of other units, unless those units are disabled
1011 - man: clarify that time-sync.target is not only sysv compat but also useful otherwise. Same for similar targets
1012 - document that service reload may be implemented as service reexec
1013 - add a man page containing packaging guidelines and recommending usage of things like Documentation=, PrivateTmp=, PrivateNetwork= and ReadOnlyDirectories=/etc /usr.
1014 - document systemd-journal-flush.service properly
1015 - documentation: recommend to connect the timer units of a service to the service via Also= in [Install]
1016 - man: document the very specific env the shutdown drop-in tools live in
1017 - man: add more examples to man pages
1018 - man: maybe sort directives in man pages, and take sections from --help and apply them to man too
1019 - document root=gpt-auto properly
1020
1021 * systemctl:
1022 - add systemctl switch to dump transaction without executing it
1023 - Add a verbose mode to "systemctl start" and friends that explains what is being done or not done
1024 - "systemctl disable" on a static unit prints no message and does
1025 nothing. "systemctl enable" does nothing, and gives a bad message
1026 about it. Should fix both to print nice actionable messages.
1027 - print nice message from systemctl --failed if there are no entries shown, and hook that into ExecStartPre of rescue.service/emergency.service
1028 - add new command to systemctl: "systemctl system-reexec" which reexecs as many daemons as virtually possible
1029 - systemctl enable: fail if target to alias into does not exist? maybe show how many units are enabled afterwards?
1030 - systemctl: "Journal has been rotated since unit was started." message is misleading
1031 - systemctl status output should include list of triggering units and their status
1032
1033 * unit install:
1034 - "systemctl mask" should find all names by which a unit is accessible
1035 (i.e. by scanning for symlinks to it) and link them all to /dev/null
1036
1037 * timer units:
1038 - timer units should get the ability to trigger when:
1039 o DST changes
1040 - Modulate timer frequency based on battery state
1041
1042 * add libsystemd-password or so to query passwords during boot using the password agent logic
1043
1044 * clean up date formatting and parsing so that all absolute/relative timestamps we format can also be parsed
1045
1046 * on shutdown: move utmp, wall, audit logic all into PID 1 (or logind?), get rid of systemd-update-utmp-runlevel
1047
1048 * make repeated alt-ctrl-del presses printing a dump
1049
1050 * hostnamed: before returning information from /etc/machine-info.conf check the modification data and reread. Similar for localed, ...
1051
1052 * currently x-systemd.timeout is lost in the initrd, since crypttab is copied into dracut, but fstab is not
1053
1054 * nspawn:
1055 - emulate /dev/kmsg using CUSE and turn off the syslog syscall
1056 with seccomp. That should provide us with a useful log buffer that
1057 systemd can log to during early boot, and disconnect container logs
1058 from the kernel's logs.
1059 - as soon as networkd has a bus interface, hook up --network-interface=,
1060 --network-bridge= with networkd, to trigger netdev creation should an
1061 interface be missing
1062 - a nice way to boot up without machine id set, so that it is set at boot
1063 automatically for supporting --ephemeral. Maybe hash the host machine id
1064 together with the machine name to generate the machine id for the container
1065 - fix logic always print a final newline on output.
1066 https://github.com/systemd/systemd/pull/272#issuecomment-113153176
1067 - should optionally support receiving WATCHDOG=1 messages from its payload
1068 PID 1...
1069 - optionally automatically add FORWARD rules to iptables whenever nspawn is
1070 running, remove them when shut down.
1071
1072 * dissect
1073 - refuse mounting over a mount point
1074 - automatically discover .roothash files in dissect, similarly to nspawn
1075
1076 * machined:
1077 - add an API so that libvirt-lxc can inform us about network interfaces being
1078 removed or added to an existing machine
1079 - "machinectl migrate" or similar to copy a container from or to a
1080 difference host, via ssh
1081 - introduce systemd-nspawn-ephemeral@.service, and hook it into
1082 "machinectl start" with a new --ephemeral switch
1083 - "machinectl status" should also show internal logs of the container in
1084 question
1085 - "machinectl history"
1086 - "machinectl diff"
1087 - "machinectl commit" that takes a writable snapshot of a tree, invokes a
1088 shell in it, and marks it read-only after use
1089
1090 * cryptsetup:
1091 - cryptsetup-generator: allow specification of passwords in crypttab itself
1092 - support rd.luks.allow-discards= kernel cmdline params in cryptsetup generator
1093
1094 * hw watchdog: optionally try to use the preset watchdog timeout instead of always overriding it
1095 https://bugs.freedesktop.org/show_bug.cgi?id=54712
1096
1097 * add a dependency on standard-conf.xml and other included files to man pages
1098
1099 * MountFlags=shared acts as MountFlags=slave right now.
1100
1101 * properly handle loop back mounts via fstab, especially regards to fsck/passno
1102
1103 * initialize the hostname from the fs label of /, if /etc/hostname does not exist?
1104
1105 * udev:
1106 - move to LGPL
1107 - kill scsi_id
1108 - add trigger --subsystem-match=usb/usb_device device
1109 - reimport udev db after MOVE events for devices without dev_t
1110
1111 * There's currently no way to cancel fsck (used to be possible via C-c or c on the console)
1112
1113 * add option to sockets to avoid activation. Instead just drop packets/connections, see http://cyberelk.net/tim/2012/02/15/portreserve-systemd-solution/
1114
1115 * coredump:
1116 - save coredump in Windows/Mozilla minidump format
1117 - when truncating coredumps, also log the full size that the process had, and make a metadata field so we can report truncated coredumps
1118
1119 * support crash reporting operation modes (https://live.gnome.org/GnomeOS/Design/Whiteboards/ProblemReporting)
1120
1121 * be able to specify a forced restart of service A where service B depends on, in case B
1122 needs to be auto-respawned?
1123
1124 * tmpfiles:
1125 - apply "x" on "D" too (see patch from William Douglas)
1126 - instead of ignoring unknown fields, reject them.
1127 - creating new directories/subvolumes/fifos/device nodes
1128 should not follow symlinks. None of the other adjustment or creation
1129 calls follow symlinks.
1130
1131 * make sure systemd-ask-password-wall does not shutdown systemd-ask-password-console too early
1132
1133 * verify that the AF_UNIX sockets of a service in the fs still exist
1134 when we start a service in order to avoid confusion when a user
1135 assumes starting a service is enough to make it accessible
1136
1137 * Make it possible to set the keymap independently from the font on
1138 the kernel cmdline. Right now setting one resets also the other.
1139
1140 * and a dbus call to generate target from current state
1141
1142 * write blog stories about:
1143 - hwdb: what belongs into it, lsusb
1144 - enabling dbus services
1145 - how to make changes to sysctl and sysfs attributes
1146 - remote access
1147 - how to pass throw-away units to systemd, or dynamically change properties of existing units
1148 - testing with Harald's awesome test kit
1149 - auto-restart
1150 - how to develop against journal browsing APIs
1151 - the journal HTTP iface
1152 - non-cgroup resource management
1153 - dynamic resource management with cgroups
1154 - refreshed, longer missions statement
1155 - calendar time events
1156 - init=/bin/sh vs. "emergency" mode, vs. "rescue" mode, vs. "multi-user" mode, vs. "graphical" mode, and the debug shell
1157 - how to create your own target
1158 - instantiated apache, dovecot and so on
1159 - hooking a script into various stages of shutdown/rearly booot
1160
1161 * investigate whether the gnome pty helper should be moved into systemd, to provide cgroup support.
1162
1163 * dot output for --test showing the 'initial transaction'
1164
1165 * pid1:
1166 - When logging about multiple units (stopping BoundTo units, conflicts, etc.),
1167 log both units as UNIT=, so that journalctl -u triggers on both.
1168 - generate better errors when people try to set transient properties
1169 that are not supported...
1170 http://lists.freedesktop.org/archives/systemd-devel/2015-February/028076.html
1171 - maybe introduce WantsMountsFor=? Usecase:
1172 http://lists.freedesktop.org/archives/systemd-devel/2015-January/027729.html
1173 - recreate systemd's D-Bus private socket file on SIGUSR2
1174 - move PAM code into its own binary
1175 - when we automatically restart a service, ensure we restart its rdeps, too.
1176 - hide PAM options in fragment parser when compile time disabled
1177 - Support --test based on current system state
1178 - 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().
1179 - after deserializing sockets in socket.c we should reapply sockopts and things
1180 - drop PID 1 reloading, only do reexecing (difficult: Reload()
1181 currently is properly synchronous, Reexec() is weird, because we
1182 cannot delay the response properly until we are back, so instead of
1183 being properly synchronous we just keep open the fd and close it
1184 when done. That means clients do not get a successful method reply,
1185 but much rather a disconnect on success.
1186 - when breaking cycles drop sysv services first, then services from /run, then from /etc, then from /usr
1187 - when a bus name of a service disappears from the bus make sure to queue further activation requests
1188
1189 * unit files:
1190 - allow port=0 in .socket units
1191 - maybe introduce ExecRestartPre=
1192 - add ReloadSignal= for configuring a reload signal to use
1193 - implement Register= switch in .socket units to enable registration
1194 in Avahi, RPC and other socket registration services.
1195 - allow Type=simple with PIDFile=
1196 https://bugzilla.redhat.com/show_bug.cgi?id=723942
1197 - allow writing multiple conditions in unit files on one line
1198 - introduce Type=pid-file
1199 - introduce mix of BindTo and Requisite
1200 - add a concept of RemainAfterExit= to scope units
1201 - Allow multiple ExecStart= for all Type= settings, so that we can cover rescue.service nicely
1202 - add verification of [Install] section to systemd-analyze verify
1203
1204 * udev-link-config:
1205 - Make sure ID_PATH is always exported and complete for
1206 network devices where possible, so we can safely rely
1207 on Path= matching
1208
1209 * sd-rtnl:
1210 - add support for more attribute types
1211 - inbuilt piping support (essentially degenerate async)? see loopback-setup.c and other places
1212
1213 * networkd:
1214 - add more keys to [Route] and [Address] sections
1215 - add support for more DHCPv4 options (and, longer term, other kinds of dynamic config)
1216 - add reduced [Link] support to .network files
1217 - properly handle routerless dhcp leases
1218 - work with non-Ethernet devices
1219 - dhcp: do we allow configuring dhcp routes on interfaces that are not the one we got the dhcp info from?
1220 - the DHCP lease data (such as NTP/DNS) is still made available when
1221 a carrier is lost on a link. It should be removed instantly.
1222 - expose in the API the following bits:
1223 - option 15, domain name
1224 - option 12, hostname and/or option 81, fqdn
1225 - option 123, 144, geolocation
1226 - option 252, configure http proxy (PAC/wpad)
1227 - provide a way to define a per-network interface default metric value
1228 for all routes to it. possibly a second default for DHCP routes.
1229 - allow Name= to be specified repeatedly in the [Match] section. Maybe also
1230 support Name=foo*|bar*|baz ?
1231 - duplicate address check for static IPs (like ARPCHECK in network-scripts)
1232 - whenever uplink info changes, make DHCP server send out FORCERENEW
1233
1234 * Figure out how to do unittests of networkd's state serialization
1235
1236 * dhcp:
1237 - figure out how much we can increase Maximum Message Size
1238
1239 * dhcp6:
1240 - add functions to set previously stored IPv6 addresses on startup and get
1241 them at shutdown; store them in client->ia_na
1242 - write more test cases
1243 - implement reconfigure support, see 5.3., 15.11. and 22.20.
1244 - implement support for temporary adressess (IA_TA)
1245 - implement dhcpv6 authentication
1246 - investigate the usefulness of Confirm messages; i.e. are there any
1247 situations where the link changes without any loss in carrier detection
1248 or interface down
1249 - some servers don't do rapid commit without a filled in IA_NA, verify
1250 this behavior
1251 - RouteTable= ?
1252
1253 External:
1254
1255 * dbus:
1256 - natively watch for dbus-*.service symlinks (PENDING)
1257 - teach dbus to activate all services it finds in /etc/systemd/services/org-*.service
1258
1259 * make cryptsetup lower --iter-time
1260
1261 * kernel: add device_type = "fb", "fbcon" to class "graphics"
1262
1263 * /usr/bin/service should actually show the new command line
1264
1265 * 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=
1266
1267 * neither pkexec nor sudo initialize environ[] from the PAM environment?
1268
1269 * fedora: update policy to declare access mode and ownership of unit files to root:root 0644, and add an rpmlint check for it
1270
1271 * register catalog database signature as file magic
1272
1273 * zsh shell completion:
1274 - <command> <verb> -<TAB> should complete options, but currently does not
1275 - systemctl add-wants,add-requires
1276
1277 * systemctl status should know about 'systemd-analyze calendar ... --iterations='
1278 * If timer has just OnInactiveSec=..., it should fire after a specified time
1279 after being started.
1280
1281 Regularly:
1282
1283 * look for close() vs. close_nointr() vs. close_nointr_nofail()
1284
1285 * check for strerror(r) instead of strerror(-r)
1286
1287 * pahole
1288
1289 * set_put(), hashmap_put() return values check. i.e. == 0 does not free()!
1290
1291 * use secure_getenv() instead of getenv() where appropriate
1292
1293 * link up selected blog stories from man pages and unit files Documentation= fields
1294 String is not UTF-8 clean, ignoring assignment