]> git.ipfire.org Git - thirdparty/systemd.git/blob - TODO
Merge pull request #15914 from poettering/ubsan-float-check
[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 * busctl prints errors to stdout:
8 busctl tree org.freedesktop.systemd1 /org/freedesktop/systemd1
9
10 External:
11
12 * Fedora: add an rpmlint check that verifies that all unit files in the RPM are listed in %systemd_post macros.
13
14 * wiki: update journal format documentation for lz4 additions
15
16 Janitorial Clean-ups:
17
18 * Rearrange tests so that the various test-xyz.c match a specific src/basic/xyz.c again
19
20 * rework mount.c and swap.c to follow proper state enumeration/deserialization
21 semantics, like we do for device.c now
22
23 Features:
24
25 * build short web pages out of each catalog entry, build them along with man
26 pages, and include hyperlinks to them in the journal output
27
28 * machined: add API to acquire UID range. add API to mount/dissect loopback
29 file. Both protected by PK. Then make nspawn use these APIs to run
30 unprivileged containers. i.e. push the truly privileged bits into machined,
31 so that the client side can remain entirely unprivileged, with SUID or
32 anything like that.
33
34 * add "throttling" to sd-event event sources: optionally, when we wake up too
35 often for one, let's turn it off entirely for a while. Use that for the
36 /proc/self/mountinfo logic.
37
38 * move our systemd-user PAM snippet to /usr/, which PAM appears to support
39 these days
40
41 * nspawn: support time namespaces
42
43 * pid1: Move to tracking of main pid/control pid of units per pidfd
44
45 * pid1: support new clone3() fork-into-cgroup feature
46
47 * pid1: also remove PID files of a service when the service starts, not just
48 when it exits
49
50 * make "systemd-dissect" an official supported tool, i.e. move to /usr/bin/ and
51 provide man page. Given that we now have a tool that can generate images like
52 this, it's useful to have one that can dump contents of them, too.
53
54 * All tools that support --root= should also learn --image= so that they can
55 operate on disk images directly. Specifically: bootctl, firstboot, tmpfiles,
56 sysusers, systemctl, repart, journalctl, coredumpctl.
57
58 * per-service credential system. Specifically: add LoadCredential= (for loading
59 cred from file), AcquireCredential= (for asking user for cred, via
60 ask-password), PassCredential= (for passing on credential systemd itself
61 got). Then, place credentials in a per-service, immutable ramfs instance (so
62 that it cannot be swapped out), destroy after use. Also pass via keyring
63 (with graceful fallback to cover for containers). Define CredentialPath= for
64 defining subdir of /run/credentials/ where to place it. Set $CREDENTIAL_PATH
65 env var for services to the result. Also pass via fd passing (optionally).
66
67 * homed: add native recovery key support. use 48 lowercase modhex characters
68 (192bit), show qr code of it, include pattern expression in user record.
69
70 * homed: introduce "degraded" state for home directories that weren't cleanly
71 unmounted (use xattr we add and remove on the loop back file)
72
73 * homed: during login resize fs automatically towards size goal. Specifically,
74 resize to diskSize if possible, but leave a certain amount (configured by a
75 new value diskLeaveFreeSize) of space free on the backing fs.
76
77 * homed: permit multiple user record signing keys to be used locally, and pick
78 the right one for signing records automatically depending on a pre-existing
79 signature
80
81 * homed: add a way to "adopt" a home directory, i.e. strip foreign signatures
82 and insert a local signature instead.
83
84 * homed: as an extension to the directory+subvolume backend: if located on
85 especially marked fs, then sync down password into LUKS header of that fs,
86 and always verify passwords against it too. Bootstrapping is a problem
87 though: if no one is logged in (or no other user even exists yet), how do you
88 unlock the volume in order to create the first user and add the first pw.
89
90 * homed: support new FS_IOC_ADD_ENCRYPTION_KEY ioctl for setting up fscrypt
91
92 * busctl: maybe expose a verb "ping" for pinging a dbus service to see if it
93 exists and responds.
94
95 * homed: maybe pre-create ~/.cache as subvol so that it can have separate quota
96 easily?
97
98 * when systemd-nspawn and suchlike dissect an OS image, and there are multiple
99 root partitions, do an strverscmp() on the partition label and boot
100 first. That is inspired how sd-boot figures out which kernel to boot, and
101 thus allows defining OS images which can be A/B updated and we default to the
102 newest version automatically, both in nspawn and in sd-boot
103
104 * cryptsetup/homed: also support FIDO2 HMAC password logic for unlocking
105 devices. (see: https://github.com/mjec/fido2-hmac-secret)
106
107 * systemd-gpt-auto should probably set x-systemd.growfs on the mounts it
108 creates
109
110 * homed/userdb: distinguish passwords and recovery keys in the records, since
111 we probably want to use different PBKDF algorithms/settings for them:
112 passwords have low entropy but recovery keys should have good entropy key
113 hence we can make them quicker to work.
114
115 * bootctl:
116 - teach it to prepare an ESP wholesale, i.e. with mkfs.vfat invocation
117 - teach it to copy in unified kernel images and maybe type #1 boot loader spec entries from host
118 - make it operate on loopback files, dissecting enough to find ESP to operate on
119
120 * by default, in systemd --user service bump the OOMAdjust to 100, as privs
121 allow so that systemd survives
122
123 * cryptsetup: allow encoding key directly in /etc/crypttab, maybe with a
124 "base64:" prefix. Useful in particular for pkcs11 mode.
125
126 * cryptsetup: reimplement the mkswap/mke2fs in cryptsetup-generator to use
127 systemd-makefs.service instead.
128
129 * socket units: allow creating a udev monitor socket with ListenDevices= or so,
130 with matches, then activate app through that passing socket over
131
132 * unify on openssl:
133 - port sd_id128_get_machine_app_specific() over from khash
134 - port resolved over from libgcrypt (DNSSEC code)
135 - port journald + fsprg over from libgcrypt
136 - port importd over from libgcrypt
137 - when that's done: kill khash.c
138 - when that's done: kill gnutls support in resolved
139
140 * kill zenata, all hail weblate?
141
142 * when we resize disks (homed?) always round up to 4K sectors, not 512K
143
144 * add growvol and makevol options for /etc/crypttab, similar to
145 x-systemd.growfs and x-systemd-makefs.
146
147 * hook up the TPM to /etc/crypttab, with a new option that is similar to the
148 new PKCS#11 option in crypttab, and allows unlocking a LUKS volume via a key
149 unsealed from the TPM. Optionally, if TPM is not available fall back to
150 TPM-less mode, and set up linear DM mapping instead (inspired by kpartx), so
151 that the device paths stay the same, regardless if crypto is used or not.
152
153 * systemd-repart: by default generate minimized partition tables (i.e. tables
154 that only cover the space actually used, excluding any free space at the
155 end), in order to maximize dd'ability. Requires libfdisk work, see
156 https://github.com/karelzak/util-linux/issues/907
157
158 * systemd-repart: optionally, allow specifying a path to initialize new
159 partitions from, i.e. an fs image file or a source device node. This would
160 then turn systemd-repart into a simple installer: with a few .repart files
161 you could replicate the host system on another device. a full installer would
162 then be: "systemd-repart /dev/sda && bootctl install /dev/sda &&
163 systemd-firstboot --image= …"
164
165 * systemd-repart: MBR partition table support. Care needs to be taken regarding
166 Type=, so that partition definitions can sanely apply to both the GPT and the
167 MBR case. Idea: accept syntax "Type=gpt:home mbr:0x83" for setting the types
168 for the two partition types explicitly. And provide an internal mapping so
169 that "Type=linux-generic" maps to the right types for both partition tables
170 automatically.
171
172 * systemd-repart: allow sizing partitions as factor of available RAM, so that
173 we can reasonably size swap partitions for hibernation.
174
175 * systemd-repart: allow running mkfs before making partitions pop up +
176 encryption via LUKS to allow booting into an empty root with only /usr mounted in
177
178 * systemd-repart: allow managing the gpt read-only partition flag + auto-mount flag
179
180 * systemd-repart: allow disabling growing of specific partitions, or making
181 them (think ESP: we don't ever want to grow it, since we cannot resize vfat)
182
183 * systemd-repart: add specifier expansion, add especifier that refers to root
184 device node of current system, /usr device node, and matching verity, so that
185 an installer can be made a "copy" installer of the booted OS
186
187 * systemd-repart: make it a static checker during early boot for existence and
188 absence of other partitions for trusted boot environments
189
190 * systemd-repart: allow config of partition uuid
191
192 * userdb: allow username prefix searches in varlink API, allow realname and
193 realname substr searches in varlink API
194
195 * userdb: allow uid/gid range checks
196
197 * userdb: allow existence checks
198
199 * pid1: activation by journal search expression
200
201 * when switching root from initrd to host, set the machine_id env var so that
202 if the host has no machine ID set yet we continue to use the random one the
203 initrd had set.
204
205 * sd-event: add native support for P_ALL waitid() watching, then move PID 1 to
206 it fo reaping assigned but unknown children. This needs to some special care
207 to operate somewhat sensibly in light of priorities: P_ALL will return
208 arbitrary processes, regardless of the priority we want to watch them with,
209 hence on each event loop iteration check all processes which we shall watch
210 with higher prio explicitly, and then watch the entire rest with P_ALL.
211
212 * tweak sd-event's child watching: keep a prioq of children to watch and use
213 waitid() only on the children with the highest priority until one is waitable
214 and ignore all lower-prio ones from that point on
215
216 * maybe introduce xattrs that can be set on the root dir of the root fs
217 partition that declare the volatility mode to use the image in. Previously I
218 thought marking this via GPT partition flags but that's not ideal since
219 that's outside of the LUKS encryption/verity verification, and we probably
220 shouldn't operate in a volatile mode unless we got told so from a trusted
221 source.
222
223 * figure out automatic partition discovery when combining writable root dir
224 with immutable /usr
225
226 * coredump: maybe when coredumping read a new xattr from /proc/$PID/exe that
227 may be used to mark a whole binary as non-coredumpable. Would fix:
228 https://bugs.freedesktop.org/show_bug.cgi?id=69447
229
230 * teach parse_timestamp() timezones like the calendar spec already knows it
231
232 * beef up hibernation to optionally do swapon/swapoff immediately before/after
233 the hibernation
234
235 * beef up s2h to implement a battery watch loop: instead of entering
236 hibernation unconditionally after coming back from resume make a decision
237 based on the battery load level: if battery level is above a specific
238 threshold, go to suspend again, only hibernate if below it. This means we'd
239 stick to suspend usually, but fall back to hibernation only when battery runs
240 empty (well, subject to our sampling interval). Related to this, check if we
241 can make ACPI _BTP (i.e. /sys/class/power_supply/*/alarm) work for us too,
242 i.e. see if it can wake up machines from suspend, so that we could resume
243 automatically when the system is low on power and move automatically to
244 hibernation mode. (see
245 https://uefi.org/sites/default/files/resources/ACPI%206_2_A_Sept29.pdf
246 section 10.2.2.8 and
247 https://docs.microsoft.com/en-us/windows-hardware/design/device-experiences/modern-standby-wake-sources
248 at the end).
249
250 * add an explicit "vertical" mode to format-table, so that "systemctl
251 status"-like outputs (i.e. with a series of field names left and values
252 right) become genuine first class citizens, and we gain automatic, sane JSON
253 output for them.
254
255 * systemd-firstboot: teach it dissector magic, so that you can point it to some
256 disk image and it will just set everything in it all behind the scenes.
257
258 * systemd-firstboot: add --force mode that replaces existing configuration.
259
260 * We should probably replace /var/log/README, /etc/rc.d/README with symlinks
261 that are linked to these places instead of copied. After all they are
262 constant vendor data.
263
264 * maybe add kernel cmdline params: to force random seed crediting
265
266 * nspawn: on cgroupsv1 issue cgroup empty handler process based on host events,
267 so that we make cgroup agent logic safe
268
269 * nspawn/machined: add API to invoke binary in container, then use that as
270 fallback in "machinectl shell"
271
272 * logind: rework pam_logind to also do a bus call in case of invocation from
273 user@.service, which returns the XDG_RUNTIME_DIR value, and make this
274 behaviour selectable via pam module option.
275
276 * homed:
277 - when user tries to log into record signed by unrecognized key, automatically add key to our chain after polkit auth
278 - hook up machined/nspawn users with a varlink user query interface
279 - rollback when resize fails mid-operation
280 - GNOME's side for forget key on suspend (requires rework so that lock screen runs outside of uid)
281 - resize on login?
282 - shrink fs on logout?
283 - update LUKS password on login if we find there's a password that unlocks the JSON record but not the LUKS device.
284 - create on activate?
285 - properties: icon url?, preferred session type?, administrator bool (which translates to 'wheel' membership)?, address?, telephone?, vcard?, samba stuff?, parental controls?
286 - communicate clearly when usb stick is safe to remove. probably involves
287 beefing up logind to make pam session close hook synchronous and wait until
288 systemd --user is shut down.
289 - logind: maybe keep a "busy fd" as long as there's a non-released session around or the user@.service
290 - maybe make automatic, read-only, time-based reflink-copies of LUKS disk images (think: time machine)
291 - distinguish destroy / remove (i.e. currently we can unregister a user, unregister+remove their home directory, but not just remove their home directory)
292 - in systemd's PAMName= logic: query passwords with ssh-askpassword, so that we can make "loginctl set-linger" mode work
293 - fingerprint authentication, pattern authentication, …
294 - make sure "classic" user records can also be managed by homed
295 - description field for groups
296 - make size of $XDG_RUNTIME_DIR configurable in user record
297 - reuse pwquality magic in firstboot
298 - query password from kernel keyring first
299 - update even if record is "absent"
300 - add a "access mode" + "fstype" field to the "status" section of json identity records reflecting the actually used access mode and fstype, even on non-luks backends
301 - move acct mgmt stuff from pam_systemd_home to pam_systemd?
302 - when "homectl --pkcs11-token-uri=" is used, synthesize ssh-authorized-keys records for all keys we have private keys on the stick for
303 - make slice for users configurable (requires logind rework)
304 - logind: populate auto-login list bus property from PKCS#11 token
305 - when determining state of a LUKS home directory, check DM suspended sysfs file
306
307 * introduce a new per-process uuid, similar to the boot id, the machine id, the
308 invocation id, that is derived from process creds, specifically a hashed
309 combination of AT_RANDOM + getpid() + the starttime from
310 /proc/self/status. Then add these ids implicitly when logging. Deriving this
311 uuid from these three things has the benefit that it can be derived easily
312 from /proc/$PID/ in a stable, and unique way that changes on both fork() and
313 exec().
314
315 * let's not GC a unit while its ratelimits are still pending
316
317 * when killing due to service watchdog timeout maybe detect whether target
318 process is under ptracing and then log loudly and continue instead.
319
320 * introduce a new group to own TPM devices
321
322 * make rfkill uaccess controllable by default, i.e. steal rule from
323 gnome-bluetooth and friends
324
325 * tweak journald context caching. In addition to caching per-process attributes
326 keyed by PID, cache per-cgroup attributes (i.e. the various xattrs we read)
327 keyed by cgroup path, and guarded by ctime changes. This should provide us
328 with a nice speed-up on services that have many processes running in the same
329 cgroup.
330
331 * make MAINPID= message reception checks even stricter: if service uses User=,
332 then check sending UID and ignore message if it doesn't match the user or
333 root.
334
335 * maybe trigger a uevent "change" on a device if "systemctl reload xyz.device"
336 is issued.
337
338 * when importing an fs tree with machined, optionally apply userns-rec-chown
339
340 * when importing an fs tree with machined, complain if image is not an OS
341
342 * Maybe introduce a helper safe_exec() or so, which is to execve() which
343 safe_fork() is to fork(). And then make revert the RLIMIT_NOFILE soft limit
344 to 1K implicitly, unless explicitly opted-out.
345
346 * rework seccomp/nnp logic that even if User= is used in combination with
347 a seccomp option we don't have to set NNP. For that, change uid first whil
348 keeping CAP_SYS_ADMIN, then apply seccomp, the drop cap.
349
350 * add a concept for automatically loading per-unit secrets off disk and
351 inserting them into the kernel keyring. Maybe SecretsDirectory= similar to
352 ConfigurationDirectory=.
353
354 * when no locale is configured, default to UEFI's PlatformLang variable
355
356 * bootctl,sd-boot: actually honour the "architecture" key
357
358 * add a new syscall group "@esoteric" for more esoteric stuff such as bpf() and
359 usefaultd() and make systemd-analyze check for it.
360
361 * paranoia: whenever we process passwords, call mlock() on the memory
362 first. i.e. look for all places we use free_and_erasep() and
363 augment them with mlock(). Also use MADV_DONTDUMP.
364
365 * Move RestrictAddressFamily= to the new cgroup create socket
366
367 * support the bind/connect/sendmsg cgroup stuff for sandboxing, and possibly
368 patching around
369
370 * maybe implicitly attach monotonic+realtime timestamps to outgoing messages in
371 log.c and sd-journal-send
372
373 * optionally: turn on cgroup delegation for per-session scope units
374
375 * introduce per-unit (i.e. per-slice, per-service) journal log size limits.
376
377 * sd-boot: automatically load EFI modules from some drop-in dir, so that people
378 can add in file system drivers and such
379
380 * sd-boot: optionally, show boot menu when previous default boot item has
381 non-zero "tries done" count
382
383 * introduce an option (or replacement) for "systemctl show" that outputs all
384 properties as JSON, similar to busctl's new JSON output. In contrast to that
385 it should skip the variant type string though.
386
387 * augment CODE_FILE=, CODE_LINE= with something like CODE_BASE= or so which
388 contains some identifier for the project, which allows us to include
389 clickable links to source files generating these log messages. The identifier
390 could be some abberviated URL prefix or so (taking inspiration from Go
391 imports). For example, for systemd we could use
392 CODE_BASE=github.com/systemd/systemd/blob/98b0b1123cc or so which is
393 sufficient to build a link by prefixing "http://" and suffixing the
394 CODE_FILE.
395
396 * Augment MESSAGE_ID with MESSAGE_BASE, in a similar fashion so that we can
397 make clickable links from log messages carrying a MESSAGE_ID, that lead to
398 some explanatory text online.
399
400 * maybe extend .path units to expose fanotify() per-mount change events
401
402 * Add a "systemctl list-units --by-slice" mode or so, which rearranges the
403 output of "systemctl list-units" slightly by showing the tree structure of
404 the slices, and the units attached to them.
405
406 * the a-posteriori stopping of units bound to units that disappeared logic
407 should be reworked: there should be a queue of units, and we should only
408 enqueue stop jobs from a defer event that processes queue instead of
409 right-away when we find a unit that is bound to one that doesn't exist
410 anymore. (similar to how the stop-unneeded queue has been reworked the same
411 way)
412
413 * nspawn: make nspawn suitable for shell pipelines: instead of triggering a
414 hangup when input is finished, send ^D, which synthesizes an EOF. Then wait
415 for hangup or ^D before passing on the EOF.
416
417 * When reloading configuration PID 1 should reset all its properties to the
418 original defaults before calling parse_config()
419
420 * nspawn: greater control over selinux label?
421
422 * hibernate/s2h: make this robust and safe to enable in Fedora by default.
423 Specifically:
424
425 1. add resume_offset support to the resume code (i.e. support swap files
426 properly)
427 2. check if swap is on weird storage and refuse if so
428 3. add auto-detection of hibernation images
429
430 * cgroups: use inotify to get notified when somebody else modifies cgroups
431 owned by us, then log a friendly warning.
432
433 * beef up log.c with support for stripping ANSI sequences from strings, so that
434 it is OK to include them in log strings. This would be particularly useful so
435 that our log messages could contain clickable links for example for unit
436 files and suchlike we operate on.
437
438 * add support for "portablectl attach http://foobar.com/waaa.raw (i.e. importd integration)
439
440 * add attach --enable and attach --now (for attach+enable+start)
441
442 * sync dynamic uids/gids between host+portable srvice (i.e. if DynamicUser=1 is set for a service, make sure that the
443 selected user is resolvable in the service even if it ships its own /etc/passwd)
444
445 * Fix DECIMAL_STR_MAX or DECIMAL_STR_WIDTH. One includes a trailing NUL, the
446 other doesn't. What a disaster. Probably to exclude it. Also
447 DECIMAL_STR_WIDTH should probably add an extra "-" into account for negative
448 numbers.
449
450 * Check that users of inotify's IN_DELETE_SELF flag are using it properly, as
451 usually IN_ATTRIB is the right way to watch deleted files, as the former only
452 fires when a file is actually removed from disk, i.e. the link count drops to
453 zero and is not open anymore, while the latter happens when a file is
454 unlinked from any dir.
455
456 * port systemctl, busctl, … over to format-table.[ch]'s table formatters
457
458 * pid1: lock image configured with RootDirectory=/RootImage= using the usual nspawn semantics while the unit is up
459
460 * add --vacuum-xyz options to coredumpctl, matching those journalctl already has.
461
462 * introduce Ephemeral= unit file switch, that creates an ephemeral copy of all
463 files and directories that are left writable for a unit, and which are
464 removed after the unit goes down again. A bit like --ephemeral for
465 systemd-nspawn but for system services. If used together with RootImage= this
466 should reflink the image file itself.
467
468 Related: add Ephemeral=<path1> <path2> … which would allow marking
469 specific paths only like this.
470
471 * add CopyFile= or so as unit file setting that may be used to copy files or
472 directory trees from the host to the services RootImage= and RootDirectory=
473 environment. Which we can use for /etc/machine-id and in particular
474 /etc/resolv.conf. Should be smart and do something useful on read-only
475 images, for example fallback to read-only bind mounting the file instead.
476
477 * show invocation ID in systemd-run output
478
479 * bypass SIGTERM state in unit files if KillSignal is SIGKILL
480
481 * add proper dbus APIs for the various sd_notify() commands, such as MAINPID=1
482 and so on, which would mean we could report errors and such.
483
484 * teach tmpfiles.d q/Q logic something sensible in the context of XFS/ext4
485 project quota
486
487 * introduce DefaultSlice= or so in system.conf that allows changing where we
488 place our units by default, i.e. change system.slice to something
489 else. Similar, ManagerSlice= should exist so that PID1's own scope unit could
490 be moved somewhere else too. Finally machined and logind should get similar
491 options so that it is possible to move user session scopes and machines to a
492 different slice too by default. Usecase: people who want to put resources on
493 the entire system, with the exception of one specific service. See:
494 https://lists.freedesktop.org/archives/systemd-devel/2018-February/040369.html
495
496 * maybe rework get_user_creds() to query the user database if $SHELL is used
497 for root, but only then.
498
499 * be stricter with fds we receive for the fdstore: close them asynchronously
500
501 * calenderspec: add support for week numbers and day numbers within a
502 year. This would allow us to define "bi-weekly" triggers safely.
503
504 * sd-bus: add vtable flag, that may be used to request client creds implicitly
505 and asynchronously before dispatching the operation
506
507 * make use of ethtool veth peer info in machined, for automatically finding out
508 host-side interface pointing to the container.
509
510 * add some special mode to LogsDirectory=/StateDirectory=… that allows
511 declaring these directories without necessarily pulling in deps for them, or
512 creating them when starting up. That way, we could declare that
513 systemd-journald writes to /var/log/journal, which could be useful when we
514 doing disk usage calculations and so on.
515
516 * taint systemd if there are fewer than 65536 users assigned (userns) to the system.
517
518 * deprecate RootDirectoryStartOnly= in favour of a new ExecStart= prefix char
519
520 * add a new RuntimeDirectoryPreserve= mode that defines a similar lifecycle for
521 the runtime dir as we maintain for the fdstore: i.e. keep it around as long
522 as the unit is running or has a job queued.
523
524 * support projid-based quota in machinectl for containers
525
526 * add a way to lock down cgroup migration: a boolean, which when set for a unit
527 makes sure the processes in it can never migrate out of it
528
529 * blog about fd store and restartable services
530
531 * document Environment=SYSTEMD_LOG_LEVEL=debug drop-in in debugging document
532
533 * rework ExecOutput and ExecInput enums so that EXEC_OUTPUT_NULL loses its
534 magic meaning and is no longer upgraded to something else if set explicitly.
535
536 * in the long run: permit a system with /etc/machine-id linked to /dev/null, to
537 make it lose its identity, i.e. be anonymous. For this we'd have to patch
538 through the whole tree to make all code deal with the case where no machine
539 ID is available.
540
541 * optionally, collect cgroup resource data, and store it in per-unit RRD files,
542 suitable for processing with rrdtool. Add bus API to access this data, and
543 possibly implement a CPULoad property based on it.
544
545 * beef up pam_systemd to take unit file settings such as cgroups properties as
546 parameters
547
548 * maybe hook of xfs/ext4 quotactl() with services? i.e. automatically manage
549 the quota of a the user indicated in User= via unit file settings, like the
550 other resource management concepts. Would mix nicely with DynamicUser=1. Or
551 alternatively, do this with projids, so that we can also cover services
552 running as root. Quota should probably cover all the special dirs such as
553 StateDirectory=, LogsDirectory=, CacheDirectory=, as well as RootDirectory= if it
554 is set, plus the whole disk space any image configured with RootImage=.
555
556 * In DynamicUser= mode: before selecting a UID, use disk quota APIs on relevant
557 disks to see if the UID is already in use.
558
559 * add "systemctl wait" or so, which does what "systemd-run --wait" does, but
560 for all units. It should be both a way to pin units into memory as well as a
561 wait to retrieve their exit data.
562
563 * expose IO accounting data on the bus, show it in systemd-run --wait and log
564 about it in the resource log message
565
566 * show whether a service has out-of-date configuration in "systemctl status" by
567 using mtime data of ConfigurationDirectory=.
568
569 * Add AddUser= setting to unit files, similar to DynamicUser=1 which however
570 creates a static, persistent user rather than a dynamic, transient user. We
571 can leverage code from sysusers.d for this.
572
573 * add some optional flag to ReadWritePaths= and friends, that has the effect
574 that we create the dir in question when the service is started. Example:
575
576 ReadWritePaths=:/var/lib/foobar
577
578 * maybe add call sd_journal_set_block_timeout() or so to set SO_SNDTIMEO for
579 the sd-journal logging socket, and, if the timeout is set to 0, sets
580 O_NONBLOCK on it. That way people can control if and when to block for
581 logging.
582
583 * hostnamed: populate form factor data from a new hwdb database, so that old
584 yogas can be recognized as "convertible" too, even if they predate the DMI
585 "convertible" form factor
586
587 * Maybe add PrivatePIDs= as new unit setting, and do minimal PID namespacing
588 after all. Be strict however, only support the equivalent of nspawn's
589 --as-pid2 switch, and sanely proxy sd_notify() messages dropping stuff such
590 as MAINPID.
591
592 * Add ExecMonitor= setting. May be used multiple times. Forks off a process in
593 the service cgroup, which is supposed to monitor the service, and when it
594 exits the service is considered failed by its monitor.
595
596 * track the per-service PAM process properly (i.e. as an additional control
597 process), so that it may be queried on the bus and everything.
598
599 * add a new "debug" job mode, that is propagated to unit_start() and for
600 services results in two things: we raise SIGSTOP right before invoking
601 execve() and turn off watchdog support. Then, use that to implement
602 "systemd-gdb" for attaching to the start-up of any system service in its
603 natural habitat.
604
605 * gpt-auto logic: related to the above, maybe support a "secondary" root
606 partition, that is mounted to / and is writable, and where the actual root's
607 /usr is mounted into.
608
609 * 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
610
611 * drop nss-myhostname in favour of nss-resolve?
612
613 * add a percentage syntax for TimeoutStopSec=, e.g. TimeoutStopSec=150%, and
614 then use that for the setting used in user@.service. It should be understood
615 relative to the configured default value.
616
617 * in networkd, when matching device types, fix up DEVTYPE rubbish the kernel passes to us
618
619 * enable LockMLOCK to take a percentage value relative to physical memory
620
621 * Permit masking specific netlink APIs with RestrictAddressFamily=
622
623 * nspawn: support that /proc, /sys/, /dev are pre-mounted
624
625 * define gpt header bits to select volatility mode
626
627 * ProtectClock= (drops CAP_SYS_TIMES, adds seecomp filters for settimeofday, adjtimex), sets DeviceAllow o /dev/rtc
628
629 * ProtectTracing= (drops CAP_SYS_PTRACE, blocks ptrace syscall, makes /sys/kernel/tracing go away)
630
631 * ProtectMount= (drop mount/umount/pivot_root from seccomp, disallow fuse via DeviceAllow, imply Mountflags=slave)
632
633 * ProtectKeyRing= to take keyring calls away
634
635 * RemoveKeyRing= to remove all keyring entries of the specified user
636
637 * ProtectReboot= that masks reboot() and kexec_load() syscalls, prohibits kill
638 on PID 1 with the relevant signals, and makes relevant files in /sys and
639 /proc (such as the sysrq stuff) unavailable
640
641 * make sure the ratelimit object can deal with USEC_INFINITY as way to turn off things
642
643 * journalctl: make sure -f ends when the container indicated by -M terminates
644
645 * mount: automatically search for "main" partition of an image has multiple
646 partitions
647
648 * in nss-systemd, if we run inside of RootDirectory= with PrivateUsers= set,
649 find a way to map the User=/Group= of the service to the right name. This way
650 a user/group for a service only has to exist on the host for the right
651 mapping to work.
652
653 * add bus API for creating unit files in /etc, reusing the code for transient units
654
655 * add bus API to remove unit files from /etc
656
657 * add bus API to retrieve current unit file contents (i.e. implement "systemctl cat" on the bus only)
658
659 * rework fopen_temporary() to make use of open_tmpfile_linkable() (problem: the
660 kernel doesn't support linkat() that replaces existing files, currently)
661
662 * transient units: don't bother with actually setting unit properties, we
663 reload the unit file anyway
664
665 * journald: sigbus API via a signal-handler safe function that people may call
666 from the SIGBUS handler
667
668 * optionally, also require WATCHDOG=1 notifications during service start-up and shutdown
669
670 * delay activation of logind until somebody logs in, or when /dev/tty0 pulls it
671 in or lingering is on (so that containers don't bother with it until PAM is used). also exit-on-idle
672
673 * cache sd_event_now() result from before the first iteration...
674
675 * PID1: find a way how we can reload unit file configuration for
676 specific units only, without reloading the whole of systemd
677
678 * add an explicit parser for LimitRTPRIO= that verifies
679 the specified range and generates sane error messages for incorrect
680 specifications.
681
682 * when we detect that there are waiting jobs but no running jobs, do something
683
684 * push CPUAffinity= also into the "cpuset" cgroup controller
685
686 * PID 1 should send out sd_notify("WATCHDOG=1") messages (for usage in the --user mode, and when run via nspawn)
687
688 * there's probably something wrong with having user mounts below /sys,
689 as we have for debugfs. for example, src/core/mount.c handles mounts
690 prefixed with /sys generally special.
691 http://lists.freedesktop.org/archives/systemd-devel/2015-June/032962.html
692
693 * fstab-generator: default to tmpfs-as-root if only usr= is specified on the kernel cmdline
694
695 * docs: bring http://www.freedesktop.org/wiki/Software/systemd/MyServiceCantGetRealtime up to date
696
697 * add a job mode that will fail if a transaction would mean stopping
698 running units. Use this in timedated to manage the NTP service
699 state.
700 http://lists.freedesktop.org/archives/systemd-devel/2015-April/030229.html
701
702 * The udev blkid built-in should expose a property that reflects
703 whether media was sensed in USB CF/SD card readers. This should then
704 be used to control SYSTEMD_READY=1/0 so that USB card readers aren't
705 picked up by systemd unless they contain a medium. This would mirror
706 the behaviour we already have for CD drives.
707
708 * networkd/udev: implement SR_IOV configuration in .link files:
709 http://lists.freedesktop.org/archives/systemd-devel/2015-January/027451.html
710
711 * hostnamectl: show root image uuid
712
713 * Find a solution for SMACK capabilities stuff:
714 http://lists.freedesktop.org/archives/systemd-devel/2014-December/026188.html
715
716 * "systemctl preset-all" should probably order the unit files it
717 operates on lexicographically before starting to work, in order to
718 ensure deterministic behaviour if two unit files conflict (like DMs
719 do, for example)
720
721 * synchronize console access with BSD locks:
722 http://lists.freedesktop.org/archives/systemd-devel/2014-October/024582.html
723
724 * as soon as we have sender timestamps, revisit coalescing multiple parallel daemon reloads:
725 http://lists.freedesktop.org/archives/systemd-devel/2014-December/025862.html
726
727 * figure out when we can use the coarse timers
728
729 * add "systemctl start -v foobar.service" that shows logs of a service
730 while the start command runs. This is non-trivial to do without
731 races though, since we should flush out all journal messages before
732 returning from the "systemctl stop".
733
734 * 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
735
736 * maybe add support for specifier expansion in user.conf, specifically DefaultEnvironment=
737
738 * maybe allow timer units with an empty Units= setting, so that they
739 can be used for resuming the system but nothing else.
740
741 * what to do about udev db binary stability for apps? (raw access is not an option)
742
743 * systemctl: if some operation fails, show log output?
744
745 * systemctl edit: use equivalent of cat() to insert existing config as a comment, prepended with #.
746 Upon editor exit, lines with one # are removed, lines with two # are left with one #, etc.
747
748 * exponential backoff in timesyncd when we cannot reach a server
749
750 * timesyncd: add ugly bus calls to set NTP servers per-interface, for usage by NM
751
752 * merge ~/.local/share and ~/.local/lib into one similar /usr/lib and /usr/share....
753
754 * add systemd.abort_on_kill or some other such flag to send SIGABRT instead of SIGKILL
755 (throughout the codebase, not only PID1)
756
757 * resolved:
758 - mDNS/DNS-SD
759 - service registration
760 - service/domain/types browsing
761 - avahi compat
762 - DNS-SD service registration from socket units
763 - resolved should optionally register additional per-interface LLMNR
764 names, so that for the container case we can establish the same name
765 (maybe "host") for referencing the server, everywhere.
766 - allow clients to request DNSSEC for a single lookup even if DNSSEC is off (?)
767 - hook up resolved with machined-based address resolution
768
769 * refcounting in sd-resolve is borked
770
771 * Add a new verb "systemctl top"
772
773 * add new gpt type for btrfs volumes
774
775 * support empty /etc boots nicely:
776 - nspawn/gpt-generator: introduce new gpt partition type for /usr
777
778 * generator that automatically discovers btrfs subvolumes, identifies their purpose based on some xattr on them.
779
780 * a way for container managers to turn off getty starting via $container_headless= or so...
781
782 * figure out a nice way how we can let the admin know what child/sibling unit causes cgroup membership for a specific unit
783
784 * For timer units: add some mechanisms so that timer units that trigger immediately on boot do not have the services
785 they run added to the initial transaction and thus confuse Type=idle.
786
787 * add bus api to query unit file's X fields.
788
789 * gpt-auto-generator:
790 - Define new partition type for encrypted swap? Support probed LUKS for encrypted swap?
791 - Make /home automount rather than mount?
792
793 * add generator that pulls in systemd-network from containers when
794 CAP_NET_ADMIN is set, more than the loopback device is defined, even
795 when it is otherwise off
796
797 * MessageQueueMessageSize= (and suchlike) should use parse_iec_size().
798
799 * implement Distribute= in socket units to allow running multiple
800 service instances processing the listening socket, and open this up
801 for ReusePort=
802
803 * introduce bus call FreezeUnit(s, b), as well as "systemctl freeze
804 $UNIT" and "systemctl thaw $UNIT" as wrappers around this. The calls
805 should SIGSTOP all unit processes in a loop until all processes of
806 it are fully stopped. This can later be used for app management by
807 desktop UIs such as gnome-shell to freeze apps that are not visible
808 on screen, not unlike how job control works on the shell
809
810 * cgroups:
811 - implement per-slice CPUFairScheduling=1 switch
812 - introduce high-level settings for RT budget, swappiness
813 - how to reset dynamically changed unit cgroup attributes sanely?
814 - when reloading configuration, apply new cgroup configuration
815 - when recursively showing the cgroup hierarchy, optionally also show
816 the hierarchies of child processes
817
818 * transient units:
819 - add field to transient units that indicate whether systemd or somebody else saves/restores its settings, for integration with libvirt
820
821 * when we detect low battery and no AC on boot, show pretty splash and refuse boot
822
823 * libsystemd-journal, libsystemd-login, libudev: add calls to easily attach these objects to sd-event event loops
824
825 * be more careful what we export on the bus as (usec_t) 0 and (usec_t) -1
826
827 * 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
828
829 * After coming back from hibernation reset hibernation swap partition using the /dev/snapshot ioctl APIs
830
831 * If we try to find a unit via a dangling symlink, generate a clean
832 error. Currently, we just ignore it and read the unit from the search
833 path anyway.
834
835 * refuse boot if /usr/lib/os-release is missing or /etc/machine-id cannot be set up
836
837 * man: the documentation of Restart= currently is very misleading and suggests the tools from ExecStartPre= might get restarted.
838
839 * load .d/*.conf dropins for device units
840
841 * sd-bus:
842 - EBADSLT handling
843 - GetAllProperties() on a non-existing object does not result in a failure currently
844 - port to sd-resolve for connecting to TCP dbus servers
845 - 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
846 - see if we can drop more message validation on the sending side
847 - add API to clone sd_bus_message objects
848 - longer term: priority inheritance
849 - dbus spec updates:
850 - NameLost/NameAcquired obsolete
851 - GVariant
852 - path escaping
853 - update systemd.special(7) to mention that dbus.socket is only about the compatibility socket now
854
855 * sd-event
856 - allow multiple signal handlers per signal?
857 - document chaining of signal handler for SIGCHLD and child handlers
858 - define more intervals where we will shift wakeup intervals around in, 1h, 6h, 24h, ...
859
860 * investigate endianness issues of UUID vs. GUID
861
862 * dbus: when a unit failed to load (i.e. is in UNIT_ERROR state), we
863 should be able to safely try another attempt when the bus call LoadUnit() is invoked.
864
865 * add a pam module that passes the hdd passphrase into the PAM stack and then expires it, for usage by gdm auto-login.
866
867 * add a pam module that on password changes updates any LUKS slot where the password matches
868
869 * test/:
870 - add unit tests for config_parse_device_allow()
871
872 * seems that when we follow symlinks to units we prefer the symlink
873 destination path over /etc and /usr. We should not do that. Instead
874 /etc should always override /run+/usr and also any symlink
875 destination.
876
877 * when isolating, try to figure out a way how we implicitly can order
878 all units we stop before the isolating unit...
879
880 * teach ConditionKernelCommandLine= globs or regexes (in order to match foobar={no,0,off})
881
882 * Add ConditionDirectoryNotEmpty= handle non-absoute paths as a search path or add
883 ConditionConfigSearchPathNotEmpty= or different syntax? See the discussion starting at
884 https://github.com/systemd/systemd/pull/15109#issuecomment-607740136.
885
886 * BootLoaderSpec: Clarify that the kernel has to be in $BOOT. Clarify
887 that the boot loader should be installed to the ESP. Define a way
888 how an installer can figure out whether a BLS compliant boot loader
889 is installed.
890
891 * think about requeuing jobs when daemon-reload is issued? usecase:
892 the initrd issues a reload after fstab from the host is accessible
893 and we might want to requeue the mounts local-fs acquired through
894 that automatically.
895
896 * systemd-inhibit: make taking delay locks useful: support sending SIGINT or SIGTERM on PrepareForSleep()
897
898 * remove any syslog support from log.c — we probably cannot do this before split-off udev is gone for good
899
900 * shutdown logging: store to EFI var, and store to USB stick?
901
902 * merge unit_kill_common() and unit_kill_context()
903
904 * EFI:
905 - honor language efi variables for default language selection (if there are any?)
906 - honor timezone efi variables for default timezone selection (if there are any?)
907 - change bootctl to be backed by systemd-bootd to control temporary and persistent default boot goal plus efi variables
908 * bootctl
909 - recognize the case when not booted on EFI
910
911 * maybe do not install getty@tty1.service symlink in /etc but in /usr?
912
913 * print a nicer explanation if people use variable/specifier expansion in ExecStart= for the first word
914
915 * mount: turn dependency information from /proc/self/mountinfo into dependency information between systemd units.
916
917 * logind:
918 - logind: optionally, ignore idle-hint logic for autosuspend, block suspend as long as a session is around
919 - logind: wakelock/opportunistic suspend support
920 - Add pretty name for seats in logind
921 - logind: allow showing logout dialog from system?
922 - add Suspend() bus calls which take timestamps to fix double suspend issues when somebody hits suspend and closes laptop quickly.
923 - if pam_systemd is invoked by su from a process that is outside of a
924 any session we should probably just become a NOP, since that's
925 usually not a real user session but just some system code that just
926 needs setuid().
927 - logind: make the Suspend()/Hibernate() bus calls wait for the for
928 the job to be completed. before returning, so that clients can wait
929 for "systemctl suspend" to finish to know when the suspending is
930 complete.
931 - logind: when the power button is pressed short, just popup a
932 logout dialog. If it is pressed for 1s, do the usual
933 shutdown. Inspiration are Macs here.
934 - expose "Locked" property on logind sesison objects
935 - maybe allow configuration of the StopTimeout for session scopes
936 - rename session scope so that it includes the UID. THat way
937 the session scope can be arranged freely in slices and we don't have
938 make assumptions about their slice anymore.
939 - follow PropertiesChanged state more closely, to deal with quick logouts and
940 relogins
941 - (optionally?) spawn seat-manager@$SEAT.service whenever a seat shows up that as CanGraphical set
942
943 * journal:
944 - consider introducing implicit _TTY= + _PPID= + _EUID= + _EGID= + _FSUID= + _FSGID= fields
945 - journald: also get thread ID from client, plus thread name
946 - journal: when waiting for journal additions in the client always sleep at least 1s or so, in order to minimize wakeups
947 - add API to close/reopen/get fd for journal client fd in libsystemd-journal.
948 - fallback to /dev/log based logging in libsystemd-journal, if we cannot log natively?
949 - declare the local journal protocol stable in the wiki interface chart
950 - sd-journal: speed up sd_journal_get_data() with transparent hash table in bg
951 - journald: when dropping msgs due to ratelimit make sure to write
952 "dropped %u messages" not only when we are about to print the next
953 message that works, but already after a short timeout
954 - check if we can make journalctl by default use --follow mode inside of less if called without args?
955 - maybe add API to send pairs of iovecs via sd_journal_send
956 - journal: add a setgid "systemd-journal" utility to invoke from libsystemd-journal, which passes fds via STDOUT and does PK access
957 - journactl: support negative filtering, i.e. FOOBAR!="waldo",
958 and !FOOBAR for events without FOOBAR.
959 - journal: store timestamp of journal_file_set_offline() int he header,
960 so it is possible to display when the file was last synced.
961 - 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.
962 - journal: find a way to allow dropping history early, based on priority, other rules
963 - journal: When used on NFS, check payload hashes
964 - journald: add kernel cmdline option to disable ratelimiting for debug purposes
965 - refuse taking lower-case variable names in sd_journal_send() and friends.
966 - journald: we currently rotate only after MaxUse+MaxFilesize has been reached.
967 - journal: deal nicely with byte-by-byte copied files, especially regards header
968 - journal: sanely deal with entries which are larger than the individual file size, but where the components would fit
969 - Replace utmp, wtmp, btmp, and lastlog completely with journal
970 - journalctl: instead --after-cursor= maybe have a --cursor=XYZ+1 syntax?
971 - when a kernel driver logs in a tight loop, we should ratelimit that too.
972 - journald: optionally, log debug messages to /run but everything else to /var
973 - journald: when we drop syslog messages because the syslog socket is
974 full, make sure to write how many messages are lost as first thing
975 to syslog when it works again.
976 - journald: allow per-priority and per-service retention times when rotating/vacuuming
977 - journald: make use of uid-range.h to managed uid ranges to split
978 journals in.
979 - journalctl: add the ability to look for the most recent process of a binary. journalctl /usr/bin/X11 --pid=-1 or so...
980 - improve journalctl performance by loading journal files
981 lazily. Encode just enough information in the file name, so that we
982 do not have to open it to know that it is not interesting for us, for
983 the most common operations.
984 - man: document that corrupted journal files is nothing to act on
985 - rework journald sigbus stuff to use mutex
986 - Set RLIMIT_NPROC for systemd-journal-xyz, and all other of our
987 services that run under their own user ids, and use User= (but only
988 in a world where userns is ubiquitous since otherwise we cannot
989 invoke those daemons on the host AND in a container anymore). Also,
990 if LimitNPROC= is used without User= we should warn and refuse
991 operation.
992 - journalctl --verify: don't show files that are currently being
993 written to as FAIL, but instead show that their are being written to.
994 - add journalctl -H that talks via ssh to a remote peer and passes through
995 binary logs data
996 - add a version of --merge which also merges /var/log/journal/remote
997 - journalctl: -m should access container journals directly by enumerating
998 them via machined, and also watch containers coming and going.
999 Benefit: nspawn --ephemeral would start working nicely with the journal.
1000 - assign MESSAGE_ID to log messages about failed services
1001
1002 * add a test if all entries in the catalog are properly formatted.
1003 (Adding dashes in a catalog entry currently results in the catalog entry
1004 being silently skipped. journalctl --update-catalog must warn about this,
1005 and we should also have a unit test to check that all our message are OK.)
1006
1007 * document:
1008 - document that deps in [Unit] sections ignore Alias= fields in
1009 [Install] units of other units, unless those units are disabled
1010 - man: clarify that time-sync.target is not only sysv compat but also useful otherwise. Same for similar targets
1011 - document that service reload may be implemented as service reexec
1012 - add a man page containing packaging guidelines and recommending usage of things like Documentation=, PrivateTmp=, PrivateNetwork= and ReadOnlyDirectories=/etc /usr.
1013 - document systemd-journal-flush.service properly
1014 - documentation: recommend to connect the timer units of a service to the service via Also= in [Install]
1015 - man: document the very specific env the shutdown drop-in tools live in
1016 - man: add more examples to man pages
1017 - man: maybe sort directives in man pages, and take sections from --help and apply them to man too
1018 - document root=gpt-auto properly
1019
1020 * systemctl:
1021 - add systemctl switch to dump transaction without executing it
1022 - Add a verbose mode to "systemctl start" and friends that explains what is being done or not done
1023 - "systemctl disable" on a static unit prints no message and does
1024 nothing. "systemctl enable" does nothing, and gives a bad message
1025 about it. Should fix both to print nice actionable messages.
1026 - print nice message from systemctl --failed if there are no entries shown, and hook that into ExecStartPre of rescue.service/emergency.service
1027 - add new command to systemctl: "systemctl system-reexec" which reexecs as many daemons as virtually possible
1028 - systemctl enable: fail if target to alias into does not exist? maybe show how many units are enabled afterwards?
1029 - systemctl: "Journal has been rotated since unit was started." message is misleading
1030 - systemctl status output should include list of triggering units and their status
1031
1032 * unit install:
1033 - "systemctl mask" should find all names by which a unit is accessible
1034 (i.e. by scanning for symlinks to it) and link them all to /dev/null
1035
1036 * timer units:
1037 - timer units should get the ability to trigger when:
1038 o DST changes
1039 - Modulate timer frequency based on battery state
1040
1041 * add libsystemd-password or so to query passwords during boot using the password agent logic
1042
1043 * clean up date formatting and parsing so that all absolute/relative timestamps we format can also be parsed
1044
1045 * on shutdown: move utmp, wall, audit logic all into PID 1 (or logind?), get rid of systemd-update-utmp-runlevel
1046
1047 * make repeated alt-ctrl-del presses printing a dump
1048
1049 * hostnamed: before returning information from /etc/machine-info.conf check the modification data and reread. Similar for localed, ...
1050
1051 * currently x-systemd.timeout is lost in the initrd, since crypttab is copied into dracut, but fstab is not
1052
1053 * nspawn:
1054 - emulate /dev/kmsg using CUSE and turn off the syslog syscall
1055 with seccomp. That should provide us with a useful log buffer that
1056 systemd can log to during early boot, and disconnect container logs
1057 from the kernel's logs.
1058 - as soon as networkd has a bus interface, hook up --network-interface=,
1059 --network-bridge= with networkd, to trigger netdev creation should an
1060 interface be missing
1061 - a nice way to boot up without machine id set, so that it is set at boot
1062 automatically for supporting --ephemeral. Maybe hash the host machine id
1063 together with the machine name to generate the machine id for the container
1064 - fix logic always print a final newline on output.
1065 https://github.com/systemd/systemd/pull/272#issuecomment-113153176
1066 - should optionally support receiving WATCHDOG=1 messages from its payload
1067 PID 1...
1068 - optionally automatically add FORWARD rules to iptables whenever nspawn is
1069 running, remove them when shut down.
1070
1071 * dissect
1072 - refuse mounting over a mount point
1073 - automatically discover .roothash files in dissect, similarly to nspawn
1074
1075 * machined:
1076 - add an API so that libvirt-lxc can inform us about network interfaces being
1077 removed or added to an existing machine
1078 - "machinectl migrate" or similar to copy a container from or to a
1079 difference host, via ssh
1080 - introduce systemd-nspawn-ephemeral@.service, and hook it into
1081 "machinectl start" with a new --ephemeral switch
1082 - "machinectl status" should also show internal logs of the container in
1083 question
1084 - "machinectl history"
1085 - "machinectl diff"
1086 - "machinectl commit" that takes a writable snapshot of a tree, invokes a
1087 shell in it, and marks it read-only after use
1088
1089 * cryptsetup:
1090 - cryptsetup-generator: allow specification of passwords in crypttab itself
1091 - support rd.luks.allow-discards= kernel cmdline params in cryptsetup generator
1092
1093 * hw watchdog: optionally try to use the preset watchdog timeout instead of always overriding it
1094 https://bugs.freedesktop.org/show_bug.cgi?id=54712
1095
1096 * add a dependency on standard-conf.xml and other included files to man pages
1097
1098 * MountFlags=shared acts as MountFlags=slave right now.
1099
1100 * properly handle loop back mounts via fstab, especially regards to fsck/passno
1101
1102 * initialize the hostname from the fs label of /, if /etc/hostname does not exist?
1103
1104 * udev:
1105 - move to LGPL
1106 - kill scsi_id
1107 - add trigger --subsystem-match=usb/usb_device device
1108 - reimport udev db after MOVE events for devices without dev_t
1109
1110 * There's currently no way to cancel fsck (used to be possible via C-c or c on the console)
1111
1112 * add option to sockets to avoid activation. Instead just drop packets/connections, see http://cyberelk.net/tim/2012/02/15/portreserve-systemd-solution/
1113
1114 * coredump:
1115 - save coredump in Windows/Mozilla minidump format
1116 - when truncating coredumps, also log the full size that the process had, and make a metadata field so we can report truncated coredumps
1117
1118 * support crash reporting operation modes (https://live.gnome.org/GnomeOS/Design/Whiteboards/ProblemReporting)
1119
1120 * be able to specify a forced restart of service A where service B depends on, in case B
1121 needs to be auto-respawned?
1122
1123 * tmpfiles:
1124 - apply "x" on "D" too (see patch from William Douglas)
1125 - instead of ignoring unknown fields, reject them.
1126 - creating new directories/subvolumes/fifos/device nodes
1127 should not follow symlinks. None of the other adjustment or creation
1128 calls follow symlinks.
1129
1130 * make sure systemd-ask-password-wall does not shutdown systemd-ask-password-console too early
1131
1132 * verify that the AF_UNIX sockets of a service in the fs still exist
1133 when we start a service in order to avoid confusion when a user
1134 assumes starting a service is enough to make it accessible
1135
1136 * Make it possible to set the keymap independently from the font on
1137 the kernel cmdline. Right now setting one resets also the other.
1138
1139 * and a dbus call to generate target from current state
1140
1141 * write blog stories about:
1142 - hwdb: what belongs into it, lsusb
1143 - enabling dbus services
1144 - how to make changes to sysctl and sysfs attributes
1145 - remote access
1146 - how to pass throw-away units to systemd, or dynamically change properties of existing units
1147 - testing with Harald's awesome test kit
1148 - auto-restart
1149 - how to develop against journal browsing APIs
1150 - the journal HTTP iface
1151 - non-cgroup resource management
1152 - dynamic resource management with cgroups
1153 - refreshed, longer missions statement
1154 - calendar time events
1155 - init=/bin/sh vs. "emergency" mode, vs. "rescue" mode, vs. "multi-user" mode, vs. "graphical" mode, and the debug shell
1156 - how to create your own target
1157 - instantiated apache, dovecot and so on
1158 - hooking a script into various stages of shutdown/rearly booot
1159
1160 * investigate whether the gnome pty helper should be moved into systemd, to provide cgroup support.
1161
1162 * dot output for --test showing the 'initial transaction'
1163
1164 * pid1:
1165 - When logging about multiple units (stopping BoundTo units, conflicts, etc.),
1166 log both units as UNIT=, so that journalctl -u triggers on both.
1167 - generate better errors when people try to set transient properties
1168 that are not supported...
1169 http://lists.freedesktop.org/archives/systemd-devel/2015-February/028076.html
1170 - maybe introduce WantsMountsFor=? Usecase:
1171 http://lists.freedesktop.org/archives/systemd-devel/2015-January/027729.html
1172 - recreate systemd's D-Bus private socket file on SIGUSR2
1173 - move PAM code into its own binary
1174 - when we automatically restart a service, ensure we restart its rdeps, too.
1175 - hide PAM options in fragment parser when compile time disabled
1176 - Support --test based on current system state
1177 - 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().
1178 - after deserializing sockets in socket.c we should reapply sockopts and things
1179 - drop PID 1 reloading, only do reexecing (difficult: Reload()
1180 currently is properly synchronous, Reexec() is weird, because we
1181 cannot delay the response properly until we are back, so instead of
1182 being properly synchronous we just keep open the fd and close it
1183 when done. That means clients do not get a successful method reply,
1184 but much rather a disconnect on success.
1185 - when breaking cycles drop sysv services first, then services from /run, then from /etc, then from /usr
1186 - when a bus name of a service disappears from the bus make sure to queue further activation requests
1187
1188 * unit files:
1189 - allow port=0 in .socket units
1190 - maybe introduce ExecRestartPre=
1191 - add ReloadSignal= for configuring a reload signal to use
1192 - implement Register= switch in .socket units to enable registration
1193 in Avahi, RPC and other socket registration services.
1194 - allow Type=simple with PIDFile=
1195 https://bugzilla.redhat.com/show_bug.cgi?id=723942
1196 - allow writing multiple conditions in unit files on one line
1197 - introduce Type=pid-file
1198 - introduce mix of BindTo and Requisite
1199 - add a concept of RemainAfterExit= to scope units
1200 - Allow multiple ExecStart= for all Type= settings, so that we can cover rescue.service nicely
1201 - add verification of [Install] section to systemd-analyze verify
1202
1203 * udev-link-config:
1204 - Make sure ID_PATH is always exported and complete for
1205 network devices where possible, so we can safely rely
1206 on Path= matching
1207
1208 * sd-rtnl:
1209 - add support for more attribute types
1210 - inbuilt piping support (essentially degenerate async)? see loopback-setup.c and other places
1211
1212 * networkd:
1213 - add more keys to [Route] and [Address] sections
1214 - add support for more DHCPv4 options (and, longer term, other kinds of dynamic config)
1215 - add reduced [Link] support to .network files
1216 - properly handle routerless dhcp leases
1217 - work with non-Ethernet devices
1218 - dhcp: do we allow configuring dhcp routes on interfaces that are not the one we got the dhcp info from?
1219 - the DHCP lease data (such as NTP/DNS) is still made available when
1220 a carrier is lost on a link. It should be removed instantly.
1221 - expose in the API the following bits:
1222 - option 15, domain name
1223 - option 12, hostname and/or option 81, fqdn
1224 - option 123, 144, geolocation
1225 - option 252, configure http proxy (PAC/wpad)
1226 - provide a way to define a per-network interface default metric value
1227 for all routes to it. possibly a second default for DHCP routes.
1228 - allow Name= to be specified repeatedly in the [Match] section. Maybe also
1229 support Name=foo*|bar*|baz ?
1230 - duplicate address check for static IPs (like ARPCHECK in network-scripts)
1231 - whenever uplink info changes, make DHCP server send out FORCERENEW
1232
1233 * dhcp:
1234 - figure out how much we can increase Maximum Message Size
1235
1236 * dhcp6:
1237 - add functions to set previously stored IPv6 addresses on startup and get
1238 them at shutdown; store them in client->ia_na
1239 - write more test cases
1240 - implement reconfigure support, see 5.3., 15.11. and 22.20.
1241 - implement support for temporary adressess (IA_TA)
1242 - implement dhcpv6 authentication
1243 - investigate the usefulness of Confirm messages; i.e. are there any
1244 situations where the link changes without any loss in carrier detection
1245 or interface down
1246 - some servers don't do rapid commit without a filled in IA_NA, verify
1247 this behavior
1248 - RouteTable= ?
1249
1250 External:
1251
1252 * dbus:
1253 - natively watch for dbus-*.service symlinks (PENDING)
1254 - teach dbus to activate all services it finds in /etc/systemd/services/org-*.service
1255
1256 * make cryptsetup lower --iter-time
1257
1258 * kernel: add device_type = "fb", "fbcon" to class "graphics"
1259
1260 * /usr/bin/service should actually show the new command line
1261
1262 * 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=
1263
1264 * neither pkexec nor sudo initialize environ[] from the PAM environment?
1265
1266 * fedora: update policy to declare access mode and ownership of unit files to root:root 0644, and add an rpmlint check for it
1267
1268 * register catalog database signature as file magic
1269
1270 * zsh shell completion:
1271 - <command> <verb> -<TAB> should complete options, but currently does not
1272 - systemctl add-wants,add-requires
1273
1274 * systemctl status should know about 'systemd-analyze calendar ... --iterations='
1275 * If timer has just OnInactiveSec=..., it should fire after a specified time
1276 after being started.
1277
1278 Regularly:
1279
1280 * look for close() vs. close_nointr() vs. close_nointr_nofail()
1281
1282 * check for strerror(r) instead of strerror(-r)
1283
1284 * pahole
1285
1286 * set_put(), hashmap_put() return values check. i.e. == 0 does not free()!
1287
1288 * use secure_getenv() instead of getenv() where appropriate
1289
1290 * link up selected blog stories from man pages and unit files Documentation= fields
1291 String is not UTF-8 clean, ignoring assignment