man: fix incorrectly placed full stop
[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 Janitorial Clean-ups:
12
13 * Rearrange tests so that the various test-xyz.c match a specific src/basic/xyz.c again
14
15 * rework mount.c and swap.c to follow proper state enumeration/deserialization
16 semantics, like we do for device.c now
17
18 Features:
19
20 * make use of the new statx mountid and rootmount fields in path_get_mnt_id()
21 and fd_is_mount_point()
22
23 * make use of new loopback setup ioctl to setup loopback devices in one atomic
24 ioctl
25
26 * nspawn: move "incoming mount" directory to /run/host, move "inaccessible"
27 nodes to /run/host, move notify socket (for sd_notify() between payload and
28 container manager)
29
30 * make use of new glibc 2.32 APIs sigabbrev_np() and strerrorname_np().
31
32 * cryptsetup: if keyfile specified in crypttab is AF_UNIX socket, connect to it
33 and read from it (like we do elsewhere with READ_FULL_FILE_CONNECT_SOCKET)
34
35 * when main nspawn supervisor process gets suspended due to SIGSTOP/SIGTTOU or
36 so, freeze the payload too.
37
38 * repart: support setting up dm-integrity with HMAC
39
40 * add /etc/integritytab, to support dm-integrity setups. In particular those
41 with HMAC as hash function, so that we can have a protected /home without
42 encryption (leaving encryption to the individual dirs/homed).
43
44 * complement root=, rootflags=, rootfstype= with rootsubdir= which allows
45 mounting a subdir of the root fs as actual root. This can be used as
46 fstype-agnostic version of btrfs' rootflags=subvol=foobar.
47
48 * add --copy-from and --copy-to command to systemd-dissect which copies stuff
49 in and out of a disk image
50
51 * Support ProtectProc= or so, using: https://patchwork.kernel.org/cover/11310197/
52
53 * if /usr/bin/swapoff fails due to OOM, log a friendly explanatory message about it
54
55 * build short web pages out of each catalog entry, build them along with man
56 pages, and include hyperlinks to them in the journal output
57
58 * machined: add API to acquire UID range. add API to mount/dissect loopback
59 file. Both protected by PK. Then make nspawn use these APIs to run
60 unprivileged containers. i.e. push the truly privileged bits into machined,
61 so that the client side can remain entirely unprivileged, with SUID or
62 anything like that.
63
64 * journald: do journal file writing out-of-process, with one writer process per
65 client UID, so that synthetic hash table collisions can slow down a specific
66 user's journal stream down but not the others.
67
68 * add "throttling" to sd-event event sources: optionally, when we wake up too
69 often for one, let's turn it off entirely for a while. Use that for the
70 /proc/self/mountinfo logic.
71
72 * move our systemd-user PAM snippet to /usr/, which PAM appears to support
73 these days
74
75 * nspawn: support time namespaces
76
77 * systemd-firstboot: make sure to always use chase_symlinks() before
78 reading/writing files
79
80 * add ConditionSecurity=tpm2
81
82 * Remove any support for booting without /usr pre-mounted in the initrd entirely.
83 Update INITRD_INTERFACE.md accordingly.
84
85 * pid1: Move to tracking of main pid/control pid of units per pidfd
86
87 * pid1: support new clone3() fork-into-cgroup feature
88
89 * pid1: also remove PID files of a service when the service starts, not just
90 when it exits
91
92 * make us use dynamically fewer deps for containers in general purpose distros:
93 o turn into dlopen() deps:
94 - libpwquality (always) - only relevant for homed, and maybe soon
95 firstboot
96 - elfutils (always)
97 - p11-kit-trust (always)
98 - kmod-libs (only when called from PID 1)
99 - cryptsetup-libs (only in RootImage= handling in PID 1, but not in systemd-cryptsetup)
100 - similar: libblkid
101 - libpam (only when called from PID 1)
102 - bzip2, xz, lz4 (always — gzip and zstd should probably stay static deps the way they are,
103 since they are so basic and our defaults)
104 o move into separate libsystemd-shared-iptables.so .so
105 - iptables-libs (only used by nspawn + networkd)
106
107 * seccomp: when SystemCallArchitectures=native is set then don't install any
108 other seccomp filters for any of the other archs, in order to reduce the
109 number of seccomp filters we install needlessly.
110
111 * seccomp: maybe use seccomp_merge() to merge our filters per-arch if we can.
112 Apparently kernel performance is much better with fewer larger seccomp
113 filters than with more smaller seccomp filters.
114
115 * systemd-path: add ESP and XBOOTLDR path. Add "private" runtime/state/cache dir enum,
116 mapping to $RUNTIME_DIRECTORY, $STATE_DIRECTORY and such
117
118 * make "systemd-dissect" an official supported tool, i.e. move to /usr/bin/ and
119 provide man page. Given that we now have a tool that can generate images like
120 this, it's useful to have one that can dump contents of them, too.
121
122 * All tools that support --root= should also learn --image= so that they can
123 operate on disk images directly. Specifically: bootctl, systemctl,
124 coredumpctl. (Already done: systemd-nspawn, systemd-firstboot,
125 systemd-repart, systemd-tmpfiles, systemd-sysusers, journalctl)
126
127 * seccomp: by default mask x32 ABI system wide on x86-64. it's on its way out
128
129 * seccomp: don't install filters for ABIs that are masked anyway for the
130 specific service
131
132 * seccomp: maybe merge all filters we install into one with that libseccomp API that allows merging.
133
134 * per-service credential system. Specifically: add LoadCredential= (for loading
135 cred from file), AcquireCredential= (for asking user for cred, via
136 ask-password), PassCredential= (for passing on credential systemd itself
137 got). Then, place credentials in a per-service, immutable ramfs instance (so
138 that it cannot be swapped out), destroy after use. Also pass via keyring
139 (with graceful fallback to cover for containers). Define CredentialPath= for
140 defining subdir of /run/credentials/ where to place it. Set $CREDENTIAL_PATH
141 env var for services to the result. Also pass via fd passing (optionally).
142
143 * homed: add native recovery key support. use 48 lowercase modhex characters
144 (192bit), show qr code of it, include pattern expression in user record.
145
146 * homed: introduce "degraded" state for home directories that weren't cleanly
147 unmounted (use xattr we add and remove on the loop back file)
148
149 * homed: during login resize fs automatically towards size goal. Specifically,
150 resize to diskSize if possible, but leave a certain amount (configured by a
151 new value diskLeaveFreeSize) of space free on the backing fs.
152
153 * homed: permit multiple user record signing keys to be used locally, and pick
154 the right one for signing records automatically depending on a pre-existing
155 signature
156
157 * homed: add a way to "adopt" a home directory, i.e. strip foreign signatures
158 and insert a local signature instead.
159
160 * homed: as an extension to the directory+subvolume backend: if located on
161 especially marked fs, then sync down password into LUKS header of that fs,
162 and always verify passwords against it too. Bootstrapping is a problem
163 though: if no one is logged in (or no other user even exists yet), how do you
164 unlock the volume in order to create the first user and add the first pw.
165
166 * homed: support new FS_IOC_ADD_ENCRYPTION_KEY ioctl for setting up fscrypt
167
168 * homed: maybe pre-create ~/.cache as subvol so that it can have separate quota
169 easily?
170
171 * busctl: maybe expose a verb "ping" for pinging a dbus service to see if it
172 exists and responds.
173
174 * when systemd-nspawn and suchlike dissect an OS image, and there are multiple
175 root partitions, do an strverscmp() on the partition label and boot
176 first. That is inspired how sd-boot figures out which kernel to boot, and
177 thus allows defining OS images which can be A/B updated and we default to the
178 newest version automatically, both in nspawn and in sd-boot
179
180 * cryptsetup: support FIDO2 tokens for deriving keys (i.e. do what homed can do
181 also in plain cryptsetup)
182
183 * systemd-gpt-auto should probably set x-systemd.growfs on the mounts it
184 creates
185
186 * homed/userdb: distinguish passwords and recovery keys in the records, since
187 we probably want to use different PBKDF algorithms/settings for them:
188 passwords have low entropy but recovery keys should have good entropy key
189 hence we can make them quicker to work.
190
191 * bootctl:
192 - teach it to prepare an ESP wholesale, i.e. with mkfs.vfat invocation
193 - teach it to copy in unified kernel images and maybe type #1 boot loader spec entries from host
194 - make it operate on loopback files, dissecting enough to find ESP to operate on
195
196 * Maybe add a separate GPT partition type to the discoverable partition spec
197 for "hibernate" partitions, that are exactly like swap partitions but only
198 activated right before hibernation and thus never used for regular swapping.
199
200 * by default, in systemd --user service bump the OOMAdjust to 100, as privs
201 allow so that systemd survives
202
203 * cryptsetup: allow encoding key directly in /etc/crypttab, maybe with a
204 "base64:" prefix. Useful in particular for pkcs11 mode.
205
206 * cryptsetup: reimplement the mkswap/mke2fs in cryptsetup-generator to use
207 systemd-makefs.service instead.
208
209 * socket units: allow creating a udev monitor socket with ListenDevices= or so,
210 with matches, then activate app through that passing socket over
211
212 * unify on openssl (as soon as OpenSSL 3.0 is out, and the Debian license
213 confusion is gone)
214 - port sd_id128_get_machine_app_specific() over from khash
215 - port resolved over from libgcrypt (DNSSEC code)
216 - port journald + fsprg over from libgcrypt
217 - port importd over from libgcrypt
218 - when that's done: kill khash.c
219 - when that's done: kill gnutls support in resolved
220
221 * kill zenata, all hail weblate?
222
223 * when we resize disks (homed?) always round up to 4K sectors, not 512K
224
225 * add growvol and makevol options for /etc/crypttab, similar to
226 x-systemd.growfs and x-systemd-makefs.
227
228 * hook up the TPM to /etc/crypttab, with a new option that is similar to the
229 new PKCS#11 option in crypttab, and allows unlocking a LUKS volume via a key
230 unsealed from the TPM. Optionally, if TPM is not available fall back to
231 TPM-less mode, and set up linear DM mapping instead (inspired by kpartx), so
232 that the device paths stay the same, regardless if crypto is used or not.
233
234 * systemd-repart: by default generate minimized partition tables (i.e. tables
235 that only cover the space actually used, excluding any free space at the
236 end), in order to maximize dd'ability. Requires libfdisk work, see
237 https://github.com/karelzak/util-linux/issues/907
238
239 * systemd-repart: MBR partition table support. Care needs to be taken regarding
240 Type=, so that partition definitions can sanely apply to both the GPT and the
241 MBR case. Idea: accept syntax "Type=gpt:home mbr:0x83" for setting the types
242 for the two partition types explicitly. And provide an internal mapping so
243 that "Type=linux-generic" maps to the right types for both partition tables
244 automatically.
245
246 * systemd-repart: allow sizing partitions as factor of available RAM, so that
247 we can reasonably size swap partitions for hibernation.
248
249 * systemd-repart: allow running mkfs before making partitions pop up +
250 encryption via LUKS to allow booting into an empty root with only /usr mounted in
251
252 * systemd-repart: allow managing the gpt read-only partition flag + auto-mount flag
253
254 * systemd-repart: allow boolean option that ensures that if existing partition
255 doesn't exist within the configured size bounds the whole command fails. This
256 is useful to implement ESP vs. XBOOTLDR schemes in installers: have one set
257 of repart files for the case where ESP is large enough and one where it isn't
258 and XBOOTLDR is added in instead. Then apply the former first, and if it
259 fails to apply use the latter.
260
261 * systemd-repart: add per-partition option to never reuse existing partition
262 and always create anew even if matching partition already exists.
263
264 * systemd-repart: add per-partition option to fail if partition already exist,
265 i.e. is not added new. Similar, add option to fail if partition does not exist yet.
266
267 * systemd-repart: add --size=auto for generating/resizing images of minimal
268 size, i.e. where the image file is sized exactly as large as necessary taking
269 SizeMin= into account, but not a single byte larger.
270
271 * systemd-repart: allow disabling growing of specific partitions, or making
272 them (think ESP: we don't ever want to grow it, since we cannot resize vfat)
273
274 * systemd-repart: make it a static checker during early boot for existence and
275 absence of other partitions for trusted boot environments
276
277 * userdb: allow username prefix searches in varlink API, allow realname and
278 realname substr searches in varlink API
279
280 * userdb: allow uid/gid range checks
281
282 * userdb: allow existence checks
283
284 * pid1: activation by journal search expression
285
286 * when switching root from initrd to host, set the machine_id env var so that
287 if the host has no machine ID set yet we continue to use the random one the
288 initrd had set.
289
290 * sd-event: add native support for P_ALL waitid() watching, then move PID 1 to
291 it fo reaping assigned but unknown children. This needs to some special care
292 to operate somewhat sensibly in light of priorities: P_ALL will return
293 arbitrary processes, regardless of the priority we want to watch them with,
294 hence on each event loop iteration check all processes which we shall watch
295 with higher prio explicitly, and then watch the entire rest with P_ALL.
296
297 * tweak sd-event's child watching: keep a prioq of children to watch and use
298 waitid() only on the children with the highest priority until one is waitable
299 and ignore all lower-prio ones from that point on
300
301 * maybe introduce xattrs that can be set on the root dir of the root fs
302 partition that declare the volatility mode to use the image in. Previously I
303 thought marking this via GPT partition flags but that's not ideal since
304 that's outside of the LUKS encryption/verity verification, and we probably
305 shouldn't operate in a volatile mode unless we got told so from a trusted
306 source.
307
308 * figure out automatic partition discovery when combining writable root dir
309 with immutable /usr
310
311 * coredump: maybe when coredumping read a new xattr from /proc/$PID/exe that
312 may be used to mark a whole binary as non-coredumpable. Would fix:
313 https://bugs.freedesktop.org/show_bug.cgi?id=69447
314
315 * teach parse_timestamp() timezones like the calendar spec already knows it
316
317 * beef up hibernation to optionally do swapon/swapoff immediately before/after
318 the hibernation
319
320 * beef up s2h to implement a battery watch loop: instead of entering
321 hibernation unconditionally after coming back from resume make a decision
322 based on the battery load level: if battery level is above a specific
323 threshold, go to suspend again, only hibernate if below it. This means we'd
324 stick to suspend usually, but fall back to hibernation only when battery runs
325 empty (well, subject to our sampling interval). Related to this, check if we
326 can make ACPI _BTP (i.e. /sys/class/power_supply/*/alarm) work for us too,
327 i.e. see if it can wake up machines from suspend, so that we could resume
328 automatically when the system is low on power and move automatically to
329 hibernation mode. (see
330 https://uefi.org/sites/default/files/resources/ACPI%206_2_A_Sept29.pdf
331 section 10.2.2.8 and
332 https://docs.microsoft.com/en-us/windows-hardware/design/device-experiences/modern-standby-wake-sources
333 at the end).
334
335 * add an explicit "vertical" mode to format-table, so that "systemctl
336 status"-like outputs (i.e. with a series of field names left and values
337 right) become genuine first class citizens, and we gain automatic, sane JSON
338 output for them.
339
340 * We should probably replace /var/log/README, /etc/rc.d/README with symlinks
341 that are linked to these places instead of copied. After all they are
342 constant vendor data.
343
344 * maybe add kernel cmdline params: to force random seed crediting
345
346 * nspawn: on cgroupsv1 issue cgroup empty handler process based on host events,
347 so that we make cgroup agent logic safe
348
349 * nspawn/machined: add API to invoke binary in container, then use that as
350 fallback in "machinectl shell"
351
352 * logind: rework pam_logind to also do a bus call in case of invocation from
353 user@.service, which returns the XDG_RUNTIME_DIR value, and make this
354 behaviour selectable via pam module option.
355
356 * homed:
357 - when user tries to log into record signed by unrecognized key, automatically add key to our chain after polkit auth
358 - rollback when resize fails mid-operation
359 - GNOME's side for forget key on suspend (requires rework so that lock screen runs outside of uid)
360 - resize on login?
361 - shrink fs on logout?
362 - update LUKS password on login if we find there's a password that unlocks the JSON record but not the LUKS device.
363 - create on activate?
364 - properties: icon url?, preferred session type?, administrator bool (which translates to 'wheel' membership)?, address?, telephone?, vcard?, samba stuff?, parental controls?
365 - communicate clearly when usb stick is safe to remove. probably involves
366 beefing up logind to make pam session close hook synchronous and wait until
367 systemd --user is shut down.
368 - logind: maybe keep a "busy fd" as long as there's a non-released session around or the user@.service
369 - maybe make automatic, read-only, time-based reflink-copies of LUKS disk
370 images (and btrfs snapshots of subvolumes) (think: time machine)
371 - distinguish destroy / remove (i.e. currently we can unregister a user, unregister+remove their home directory, but not just remove their home directory)
372 - in systemd's PAMName= logic: query passwords with ssh-askpassword, so that we can make "loginctl set-linger" mode work
373 - fingerprint authentication, pattern authentication, …
374 - make sure "classic" user records can also be managed by homed
375 - make size of $XDG_RUNTIME_DIR configurable in user record
376 - reuse pwquality magic in firstboot
377 - query password from kernel keyring first
378 - update even if record is "absent"
379 - add a "access mode" + "fstype" field to the "status" section of json identity records reflecting the actually used access mode and fstype, even on non-luks backends
380 - move acct mgmt stuff from pam_systemd_home to pam_systemd?
381 - when "homectl --pkcs11-token-uri=" is used, synthesize ssh-authorized-keys records for all keys we have private keys on the stick for
382 - make slice for users configurable (requires logind rework)
383 - logind: populate auto-login list bus property from PKCS#11 token
384 - when determining state of a LUKS home directory, check DM suspended sysfs file
385 - introduce API for "making room", that grows/shrinks home directory
386 according to elastic parameters, discards blocks, and removes additional snapshots. Call it
387 either from UI when disk space gets low
388
389 * introduce a new per-process uuid, similar to the boot id, the machine id, the
390 invocation id, that is derived from process creds, specifically a hashed
391 combination of AT_RANDOM + getpid() + the starttime from
392 /proc/self/status. Then add these ids implicitly when logging. Deriving this
393 uuid from these three things has the benefit that it can be derived easily
394 from /proc/$PID/ in a stable, and unique way that changes on both fork() and
395 exec().
396
397 * let's not GC a unit while its ratelimits are still pending
398
399 * when killing due to service watchdog timeout maybe detect whether target
400 process is under ptracing and then log loudly and continue instead.
401
402 * introduce a new group to own TPM devices
403
404 * make rfkill uaccess controllable by default, i.e. steal rule from
405 gnome-bluetooth and friends
406
407 * tweak journald context caching. In addition to caching per-process attributes
408 keyed by PID, cache per-cgroup attributes (i.e. the various xattrs we read)
409 keyed by cgroup path, and guarded by ctime changes. This should provide us
410 with a nice speed-up on services that have many processes running in the same
411 cgroup.
412
413 * make MAINPID= message reception checks even stricter: if service uses User=,
414 then check sending UID and ignore message if it doesn't match the user or
415 root.
416
417 * maybe trigger a uevent "change" on a device if "systemctl reload xyz.device"
418 is issued.
419
420 * when importing an fs tree with machined, optionally apply userns-rec-chown
421
422 * when importing an fs tree with machined, complain if image is not an OS
423
424 * Maybe introduce a helper safe_exec() or so, which is to execve() which
425 safe_fork() is to fork(). And then make revert the RLIMIT_NOFILE soft limit
426 to 1K implicitly, unless explicitly opted-out.
427
428 * rework seccomp/nnp logic that even if User= is used in combination with
429 a seccomp option we don't have to set NNP. For that, change uid first whil
430 keeping CAP_SYS_ADMIN, then apply seccomp, the drop cap.
431
432 * add a concept for automatically loading per-unit secrets off disk and
433 inserting them into the kernel keyring. Maybe SecretsDirectory= similar to
434 ConfigurationDirectory=.
435
436 * when no locale is configured, default to UEFI's PlatformLang variable
437
438 * bootctl,sd-boot: actually honour the "architecture" key
439
440 * add a new syscall group "@esoteric" for more esoteric stuff such as bpf() and
441 usefaultd() and make systemd-analyze check for it.
442
443 * paranoia: whenever we process passwords, call mlock() on the memory
444 first. i.e. look for all places we use free_and_erasep() and
445 augment them with mlock(). Also use MADV_DONTDUMP.
446
447 * Move RestrictAddressFamily= to the new cgroup create socket
448
449 * support the bind/connect/sendmsg cgroup stuff for sandboxing, and possibly
450 patching around
451
452 * maybe implicitly attach monotonic+realtime timestamps to outgoing messages in
453 log.c and sd-journal-send
454
455 * optionally: turn on cgroup delegation for per-session scope units
456
457 * introduce per-unit (i.e. per-slice, per-service) journal log size limits.
458
459 * sd-boot: automatically load EFI modules from some drop-in dir, so that people
460 can add in file system drivers and such
461
462 * sd-boot: optionally, show boot menu when previous default boot item has
463 non-zero "tries done" count
464
465 * introduce an option (or replacement) for "systemctl show" that outputs all
466 properties as JSON, similar to busctl's new JSON output. In contrast to that
467 it should skip the variant type string though.
468
469 * augment CODE_FILE=, CODE_LINE= with something like CODE_BASE= or so which
470 contains some identifier for the project, which allows us to include
471 clickable links to source files generating these log messages. The identifier
472 could be some abberviated URL prefix or so (taking inspiration from Go
473 imports). For example, for systemd we could use
474 CODE_BASE=github.com/systemd/systemd/blob/98b0b1123cc or so which is
475 sufficient to build a link by prefixing "http://" and suffixing the
476 CODE_FILE.
477
478 * Augment MESSAGE_ID with MESSAGE_BASE, in a similar fashion so that we can
479 make clickable links from log messages carrying a MESSAGE_ID, that lead to
480 some explanatory text online.
481
482 * maybe extend .path units to expose fanotify() per-mount change events
483
484 * Add a "systemctl list-units --by-slice" mode or so, which rearranges the
485 output of "systemctl list-units" slightly by showing the tree structure of
486 the slices, and the units attached to them.
487
488 * the a-posteriori stopping of units bound to units that disappeared logic
489 should be reworked: there should be a queue of units, and we should only
490 enqueue stop jobs from a defer event that processes queue instead of
491 right-away when we find a unit that is bound to one that doesn't exist
492 anymore. (similar to how the stop-unneeded queue has been reworked the same
493 way)
494
495 * nspawn: make nspawn suitable for shell pipelines: instead of triggering a
496 hangup when input is finished, send ^D, which synthesizes an EOF. Then wait
497 for hangup or ^D before passing on the EOF.
498
499 * When reloading configuration PID 1 should reset all its properties to the
500 original defaults before calling parse_config()
501
502 * nspawn: greater control over selinux label?
503
504 * hibernate/s2h: make this robust and safe to enable in Fedora by default.
505 Specifically:
506
507 1. add resume_offset support to the resume code (i.e. support swap files
508 properly)
509 2. check if swap is on weird storage and refuse if so
510 3. add auto-detection of hibernation images
511
512 * cgroups: use inotify to get notified when somebody else modifies cgroups
513 owned by us, then log a friendly warning.
514
515 * beef up log.c with support for stripping ANSI sequences from strings, so that
516 it is OK to include them in log strings. This would be particularly useful so
517 that our log messages could contain clickable links for example for unit
518 files and suchlike we operate on.
519
520 * add support for "portablectl attach http://foobar.com/waaa.raw (i.e. importd integration)
521
522 * add attach --enable and attach --now (for attach+enable+start)
523
524 * sync dynamic uids/gids between host+portable srvice (i.e. if DynamicUser=1 is set for a service, make sure that the
525 selected user is resolvable in the service even if it ships its own /etc/passwd)
526
527 * Fix DECIMAL_STR_MAX or DECIMAL_STR_WIDTH. One includes a trailing NUL, the
528 other doesn't. What a disaster. Probably to exclude it. Also
529 DECIMAL_STR_WIDTH should probably add an extra "-" into account for negative
530 numbers.
531
532 * Check that users of inotify's IN_DELETE_SELF flag are using it properly, as
533 usually IN_ATTRIB is the right way to watch deleted files, as the former only
534 fires when a file is actually removed from disk, i.e. the link count drops to
535 zero and is not open anymore, while the latter happens when a file is
536 unlinked from any dir.
537
538 * port systemctl, busctl, … over to format-table.[ch]'s table formatters
539
540 * pid1: lock image configured with RootDirectory=/RootImage= using the usual nspawn semantics while the unit is up
541
542 * add --vacuum-xyz options to coredumpctl, matching those journalctl already has.
543
544 * introduce Ephemeral= unit file switch, that creates an ephemeral copy of all
545 files and directories that are left writable for a unit, and which are
546 removed after the unit goes down again. A bit like --ephemeral for
547 systemd-nspawn but for system services. If used together with RootImage= this
548 should reflink the image file itself.
549
550 Related: add Ephemeral=<path1> <path2> … which would allow marking
551 specific paths only like this.
552
553 * add CopyFile= or so as unit file setting that may be used to copy files or
554 directory trees from the host to the services RootImage= and RootDirectory=
555 environment. Which we can use for /etc/machine-id and in particular
556 /etc/resolv.conf. Should be smart and do something useful on read-only
557 images, for example fallback to read-only bind mounting the file instead.
558
559 * show invocation ID in systemd-run output
560
561 * bypass SIGTERM state in unit files if KillSignal is SIGKILL
562
563 * add proper dbus APIs for the various sd_notify() commands, such as MAINPID=1
564 and so on, which would mean we could report errors and such.
565
566 * teach tmpfiles.d q/Q logic something sensible in the context of XFS/ext4
567 project quota
568
569 * introduce DefaultSlice= or so in system.conf that allows changing where we
570 place our units by default, i.e. change system.slice to something
571 else. Similar, ManagerSlice= should exist so that PID1's own scope unit could
572 be moved somewhere else too. Finally machined and logind should get similar
573 options so that it is possible to move user session scopes and machines to a
574 different slice too by default. Usecase: people who want to put resources on
575 the entire system, with the exception of one specific service. See:
576 https://lists.freedesktop.org/archives/systemd-devel/2018-February/040369.html
577
578 * maybe rework get_user_creds() to query the user database if $SHELL is used
579 for root, but only then.
580
581 * be stricter with fds we receive for the fdstore: close them asynchronously
582
583 * calenderspec: add support for week numbers and day numbers within a
584 year. This would allow us to define "bi-weekly" triggers safely.
585
586 * sd-bus: add vtable flag, that may be used to request client creds implicitly
587 and asynchronously before dispatching the operation
588
589 * make use of ethtool veth peer info in machined, for automatically finding out
590 host-side interface pointing to the container.
591
592 * add some special mode to LogsDirectory=/StateDirectory=… that allows
593 declaring these directories without necessarily pulling in deps for them, or
594 creating them when starting up. That way, we could declare that
595 systemd-journald writes to /var/log/journal, which could be useful when we
596 doing disk usage calculations and so on.
597
598 * taint systemd if there are fewer than 65536 users assigned (userns) to the system.
599
600 * deprecate RootDirectoryStartOnly= in favour of a new ExecStart= prefix char
601
602 * add a new RuntimeDirectoryPreserve= mode that defines a similar lifecycle for
603 the runtime dir as we maintain for the fdstore: i.e. keep it around as long
604 as the unit is running or has a job queued.
605
606 * support projid-based quota in machinectl for containers
607
608 * add a way to lock down cgroup migration: a boolean, which when set for a unit
609 makes sure the processes in it can never migrate out of it
610
611 * blog about fd store and restartable services
612
613 * document Environment=SYSTEMD_LOG_LEVEL=debug drop-in in debugging document
614
615 * rework ExecOutput and ExecInput enums so that EXEC_OUTPUT_NULL loses its
616 magic meaning and is no longer upgraded to something else if set explicitly.
617
618 * in the long run: permit a system with /etc/machine-id linked to /dev/null, to
619 make it lose its identity, i.e. be anonymous. For this we'd have to patch
620 through the whole tree to make all code deal with the case where no machine
621 ID is available.
622
623 * optionally, collect cgroup resource data, and store it in per-unit RRD files,
624 suitable for processing with rrdtool. Add bus API to access this data, and
625 possibly implement a CPULoad property based on it.
626
627 * beef up pam_systemd to take unit file settings such as cgroups properties as
628 parameters
629
630 * maybe hook of xfs/ext4 quotactl() with services? i.e. automatically manage
631 the quota of a the user indicated in User= via unit file settings, like the
632 other resource management concepts. Would mix nicely with DynamicUser=1. Or
633 alternatively, do this with projids, so that we can also cover services
634 running as root. Quota should probably cover all the special dirs such as
635 StateDirectory=, LogsDirectory=, CacheDirectory=, as well as RootDirectory= if it
636 is set, plus the whole disk space any image configured with RootImage=.
637
638 * In DynamicUser= mode: before selecting a UID, use disk quota APIs on relevant
639 disks to see if the UID is already in use.
640
641 * add "systemctl wait" or so, which does what "systemd-run --wait" does, but
642 for all units. It should be both a way to pin units into memory as well as a
643 wait to retrieve their exit data.
644
645 * expose IO accounting data on the bus, show it in systemd-run --wait and log
646 about it in the resource log message
647
648 * show whether a service has out-of-date configuration in "systemctl status" by
649 using mtime data of ConfigurationDirectory=.
650
651 * Add AddUser= setting to unit files, similar to DynamicUser=1 which however
652 creates a static, persistent user rather than a dynamic, transient user. We
653 can leverage code from sysusers.d for this.
654
655 * add some optional flag to ReadWritePaths= and friends, that has the effect
656 that we create the dir in question when the service is started. Example:
657
658 ReadWritePaths=:/var/lib/foobar
659
660 * maybe add call sd_journal_set_block_timeout() or so to set SO_SNDTIMEO for
661 the sd-journal logging socket, and, if the timeout is set to 0, sets
662 O_NONBLOCK on it. That way people can control if and when to block for
663 logging.
664
665 * hostnamed: populate form factor data from a new hwdb database, so that old
666 yogas can be recognized as "convertible" too, even if they predate the DMI
667 "convertible" form factor
668
669 * Maybe add PrivatePIDs= as new unit setting, and do minimal PID namespacing
670 after all. Be strict however, only support the equivalent of nspawn's
671 --as-pid2 switch, and sanely proxy sd_notify() messages dropping stuff such
672 as MAINPID.
673
674 * Add ExecMonitor= setting. May be used multiple times. Forks off a process in
675 the service cgroup, which is supposed to monitor the service, and when it
676 exits the service is considered failed by its monitor.
677
678 * track the per-service PAM process properly (i.e. as an additional control
679 process), so that it may be queried on the bus and everything.
680
681 * add a new "debug" job mode, that is propagated to unit_start() and for
682 services results in two things: we raise SIGSTOP right before invoking
683 execve() and turn off watchdog support. Then, use that to implement
684 "systemd-gdb" for attaching to the start-up of any system service in its
685 natural habitat.
686
687 * gpt-auto logic: related to the above, maybe support a "secondary" root
688 partition, that is mounted to / and is writable, and where the actual root's
689 /usr is mounted into.
690
691 * 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
692
693 * drop nss-myhostname in favour of nss-resolve?
694
695 * add a percentage syntax for TimeoutStopSec=, e.g. TimeoutStopSec=150%, and
696 then use that for the setting used in user@.service. It should be understood
697 relative to the configured default value.
698
699 * in networkd, when matching device types, fix up DEVTYPE rubbish the kernel passes to us
700
701 * enable LockMLOCK to take a percentage value relative to physical memory
702
703 * Permit masking specific netlink APIs with RestrictAddressFamily=
704
705 * nspawn: support that /proc, /sys/, /dev are pre-mounted
706
707 * define gpt header bits to select volatility mode
708
709 * ProtectClock= (drops CAP_SYS_TIMES, adds seecomp filters for settimeofday, adjtimex), sets DeviceAllow o /dev/rtc
710
711 * ProtectTracing= (drops CAP_SYS_PTRACE, blocks ptrace syscall, makes /sys/kernel/tracing go away)
712
713 * ProtectMount= (drop mount/umount/pivot_root from seccomp, disallow fuse via DeviceAllow, imply Mountflags=slave)
714
715 * ProtectKeyRing= to take keyring calls away
716
717 * RemoveKeyRing= to remove all keyring entries of the specified user
718
719 * ProtectReboot= that masks reboot() and kexec_load() syscalls, prohibits kill
720 on PID 1 with the relevant signals, and makes relevant files in /sys and
721 /proc (such as the sysrq stuff) unavailable
722
723 * make sure the ratelimit object can deal with USEC_INFINITY as way to turn off things
724
725 * journalctl: make sure -f ends when the container indicated by -M terminates
726
727 * mount: automatically search for "main" partition of an image has multiple
728 partitions
729
730 * in nss-systemd, if we run inside of RootDirectory= with PrivateUsers= set,
731 find a way to map the User=/Group= of the service to the right name. This way
732 a user/group for a service only has to exist on the host for the right
733 mapping to work.
734
735 * add bus API for creating unit files in /etc, reusing the code for transient units
736
737 * add bus API to remove unit files from /etc
738
739 * add bus API to retrieve current unit file contents (i.e. implement "systemctl cat" on the bus only)
740
741 * rework fopen_temporary() to make use of open_tmpfile_linkable() (problem: the
742 kernel doesn't support linkat() that replaces existing files, currently)
743
744 * transient units: don't bother with actually setting unit properties, we
745 reload the unit file anyway
746
747 * journald: sigbus API via a signal-handler safe function that people may call
748 from the SIGBUS handler
749
750 * optionally, also require WATCHDOG=1 notifications during service start-up and shutdown
751
752 * delay activation of logind until somebody logs in, or when /dev/tty0 pulls it
753 in or lingering is on (so that containers don't bother with it until PAM is used). also exit-on-idle
754
755 * cache sd_event_now() result from before the first iteration...
756
757 * PID1: find a way how we can reload unit file configuration for
758 specific units only, without reloading the whole of systemd
759
760 * add an explicit parser for LimitRTPRIO= that verifies
761 the specified range and generates sane error messages for incorrect
762 specifications.
763
764 * when we detect that there are waiting jobs but no running jobs, do something
765
766 * push CPUAffinity= also into the "cpuset" cgroup controller
767
768 * PID 1 should send out sd_notify("WATCHDOG=1") messages (for usage in the --user mode, and when run via nspawn)
769
770 * there's probably something wrong with having user mounts below /sys,
771 as we have for debugfs. for example, src/core/mount.c handles mounts
772 prefixed with /sys generally special.
773 http://lists.freedesktop.org/archives/systemd-devel/2015-June/032962.html
774
775 * fstab-generator: default to tmpfs-as-root if only usr= is specified on the kernel cmdline
776
777 * docs: bring http://www.freedesktop.org/wiki/Software/systemd/MyServiceCantGetRealtime up to date
778
779 * add a job mode that will fail if a transaction would mean stopping
780 running units. Use this in timedated to manage the NTP service
781 state.
782 http://lists.freedesktop.org/archives/systemd-devel/2015-April/030229.html
783
784 * The udev blkid built-in should expose a property that reflects
785 whether media was sensed in USB CF/SD card readers. This should then
786 be used to control SYSTEMD_READY=1/0 so that USB card readers aren't
787 picked up by systemd unless they contain a medium. This would mirror
788 the behaviour we already have for CD drives.
789
790 * networkd/udev: implement SR_IOV configuration in .link files:
791 http://lists.freedesktop.org/archives/systemd-devel/2015-January/027451.html
792
793 * hostnamectl: show root image uuid
794
795 * Find a solution for SMACK capabilities stuff:
796 http://lists.freedesktop.org/archives/systemd-devel/2014-December/026188.html
797
798 * "systemctl preset-all" should probably order the unit files it
799 operates on lexicographically before starting to work, in order to
800 ensure deterministic behaviour if two unit files conflict (like DMs
801 do, for example)
802
803 * synchronize console access with BSD locks:
804 http://lists.freedesktop.org/archives/systemd-devel/2014-October/024582.html
805
806 * as soon as we have sender timestamps, revisit coalescing multiple parallel daemon reloads:
807 http://lists.freedesktop.org/archives/systemd-devel/2014-December/025862.html
808
809 * figure out when we can use the coarse timers
810
811 * add "systemctl start -v foobar.service" that shows logs of a service
812 while the start command runs. This is non-trivial to do without
813 races though, since we should flush out all journal messages before
814 returning from the "systemctl stop".
815
816 * 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
817
818 * maybe add support for specifier expansion in user.conf, specifically DefaultEnvironment=
819
820 * maybe allow timer units with an empty Units= setting, so that they
821 can be used for resuming the system but nothing else.
822
823 * what to do about udev db binary stability for apps? (raw access is not an option)
824
825 * systemctl: if some operation fails, show log output?
826
827 * systemctl edit: use equivalent of cat() to insert existing config as a comment, prepended with #.
828 Upon editor exit, lines with one # are removed, lines with two # are left with one #, etc.
829
830 * exponential backoff in timesyncd when we cannot reach a server
831
832 * timesyncd: add ugly bus calls to set NTP servers per-interface, for usage by NM
833
834 * merge ~/.local/share and ~/.local/lib into one similar /usr/lib and /usr/share....
835
836 * add systemd.abort_on_kill or some other such flag to send SIGABRT instead of SIGKILL
837 (throughout the codebase, not only PID1)
838
839 * resolved:
840 - mDNS/DNS-SD
841 - service registration
842 - service/domain/types browsing
843 - avahi compat
844 - DNS-SD service registration from socket units
845 - resolved should optionally register additional per-interface LLMNR
846 names, so that for the container case we can establish the same name
847 (maybe "host") for referencing the server, everywhere.
848 - allow clients to request DNSSEC for a single lookup even if DNSSEC is off (?)
849 - hook up resolved with machined-based address resolution
850
851 * refcounting in sd-resolve is borked
852
853 * Add a new verb "systemctl top"
854
855 * add new gpt type for btrfs volumes
856
857 * support empty /etc boots nicely:
858 - nspawn/gpt-generator: introduce new gpt partition type for /usr
859
860 * generator that automatically discovers btrfs subvolumes, identifies their purpose based on some xattr on them.
861
862 * a way for container managers to turn off getty starting via $container_headless= or so...
863
864 * figure out a nice way how we can let the admin know what child/sibling unit causes cgroup membership for a specific unit
865
866 * For timer units: add some mechanisms so that timer units that trigger immediately on boot do not have the services
867 they run added to the initial transaction and thus confuse Type=idle.
868
869 * add bus api to query unit file's X fields.
870
871 * gpt-auto-generator:
872 - Define new partition type for encrypted swap? Support probed LUKS for encrypted swap?
873 - Make /home automount rather than mount?
874
875 * add generator that pulls in systemd-network from containers when
876 CAP_NET_ADMIN is set, more than the loopback device is defined, even
877 when it is otherwise off
878
879 * MessageQueueMessageSize= (and suchlike) should use parse_iec_size().
880
881 * implement Distribute= in socket units to allow running multiple
882 service instances processing the listening socket, and open this up
883 for ReusePort=
884
885 * introduce bus call FreezeUnit(s, b), as well as "systemctl freeze
886 $UNIT" and "systemctl thaw $UNIT" as wrappers around this. The calls
887 should SIGSTOP all unit processes in a loop until all processes of
888 it are fully stopped. This can later be used for app management by
889 desktop UIs such as gnome-shell to freeze apps that are not visible
890 on screen, not unlike how job control works on the shell
891
892 * cgroups:
893 - implement per-slice CPUFairScheduling=1 switch
894 - introduce high-level settings for RT budget, swappiness
895 - how to reset dynamically changed unit cgroup attributes sanely?
896 - when reloading configuration, apply new cgroup configuration
897 - when recursively showing the cgroup hierarchy, optionally also show
898 the hierarchies of child processes
899
900 * transient units:
901 - add field to transient units that indicate whether systemd or somebody else saves/restores its settings, for integration with libvirt
902
903 * when we detect low battery and no AC on boot, show pretty splash and refuse boot
904
905 * libsystemd-journal, libsystemd-login, libudev: add calls to easily attach these objects to sd-event event loops
906
907 * be more careful what we export on the bus as (usec_t) 0 and (usec_t) -1
908
909 * 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
910
911 * After coming back from hibernation reset hibernation swap partition using the /dev/snapshot ioctl APIs
912
913 * If we try to find a unit via a dangling symlink, generate a clean
914 error. Currently, we just ignore it and read the unit from the search
915 path anyway.
916
917 * refuse boot if /usr/lib/os-release is missing or /etc/machine-id cannot be set up
918
919 * man: the documentation of Restart= currently is very misleading and suggests the tools from ExecStartPre= might get restarted.
920
921 * load .d/*.conf dropins for device units
922
923 * sd-bus:
924 - EBADSLT handling
925 - GetAllProperties() on a non-existing object does not result in a failure currently
926 - port to sd-resolve for connecting to TCP dbus servers
927 - 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
928 - see if we can drop more message validation on the sending side
929 - add API to clone sd_bus_message objects
930 - longer term: priority inheritance
931 - dbus spec updates:
932 - NameLost/NameAcquired obsolete
933 - GVariant
934 - path escaping
935 - update systemd.special(7) to mention that dbus.socket is only about the compatibility socket now
936
937 * sd-event
938 - allow multiple signal handlers per signal?
939 - document chaining of signal handler for SIGCHLD and child handlers
940 - define more intervals where we will shift wakeup intervals around in, 1h, 6h, 24h, ...
941 - maybe support iouring as backend, so that we allow hooking read and write
942 operations instead of IO ready events into event loops. See considerations
943 here:
944 http://blog.vmsplice.net/2020/07/rethinking-event-loop-integration-for.html
945
946 * investigate endianness issues of UUID vs. GUID
947
948 * dbus: when a unit failed to load (i.e. is in UNIT_ERROR state), we
949 should be able to safely try another attempt when the bus call LoadUnit() is invoked.
950
951 * add a pam module that passes the hdd passphrase into the PAM stack and then expires it, for usage by gdm auto-login.
952
953 * add a pam module that on password changes updates any LUKS slot where the password matches
954
955 * test/:
956 - add unit tests for config_parse_device_allow()
957
958 * seems that when we follow symlinks to units we prefer the symlink
959 destination path over /etc and /usr. We should not do that. Instead
960 /etc should always override /run+/usr and also any symlink
961 destination.
962
963 * when isolating, try to figure out a way how we implicitly can order
964 all units we stop before the isolating unit...
965
966 * teach ConditionKernelCommandLine= globs or regexes (in order to match foobar={no,0,off})
967
968 * Add ConditionDirectoryNotEmpty= handle non-absoute paths as a search path or add
969 ConditionConfigSearchPathNotEmpty= or different syntax? See the discussion starting at
970 https://github.com/systemd/systemd/pull/15109#issuecomment-607740136.
971
972 * BootLoaderSpec: Clarify that the kernel has to be in $BOOT. Clarify
973 that the boot loader should be installed to the ESP. Define a way
974 how an installer can figure out whether a BLS compliant boot loader
975 is installed.
976
977 * think about requeuing jobs when daemon-reload is issued? usecase:
978 the initrd issues a reload after fstab from the host is accessible
979 and we might want to requeue the mounts local-fs acquired through
980 that automatically.
981
982 * systemd-inhibit: make taking delay locks useful: support sending SIGINT or SIGTERM on PrepareForSleep()
983
984 * remove any syslog support from log.c — we probably cannot do this before split-off udev is gone for good
985
986 * shutdown logging: store to EFI var, and store to USB stick?
987
988 * merge unit_kill_common() and unit_kill_context()
989
990 * EFI:
991 - honor language efi variables for default language selection (if there are any?)
992 - honor timezone efi variables for default timezone selection (if there are any?)
993 - change bootctl to be backed by systemd-bootd to control temporary and persistent default boot goal plus efi variables
994 * bootctl
995 - recognize the case when not booted on EFI
996
997 * maybe do not install getty@tty1.service symlink in /etc but in /usr?
998
999 * print a nicer explanation if people use variable/specifier expansion in ExecStart= for the first word
1000
1001 * mount: turn dependency information from /proc/self/mountinfo into dependency information between systemd units.
1002
1003 * logind:
1004 - logind: optionally, ignore idle-hint logic for autosuspend, block suspend as long as a session is around
1005 - logind: wakelock/opportunistic suspend support
1006 - Add pretty name for seats in logind
1007 - logind: allow showing logout dialog from system?
1008 - add Suspend() bus calls which take timestamps to fix double suspend issues when somebody hits suspend and closes laptop quickly.
1009 - if pam_systemd is invoked by su from a process that is outside of a
1010 any session we should probably just become a NOP, since that's
1011 usually not a real user session but just some system code that just
1012 needs setuid().
1013 - logind: make the Suspend()/Hibernate() bus calls wait for the for
1014 the job to be completed. before returning, so that clients can wait
1015 for "systemctl suspend" to finish to know when the suspending is
1016 complete.
1017 - logind: when the power button is pressed short, just popup a
1018 logout dialog. If it is pressed for 1s, do the usual
1019 shutdown. Inspiration are Macs here.
1020 - expose "Locked" property on logind sesison objects
1021 - maybe allow configuration of the StopTimeout for session scopes
1022 - rename session scope so that it includes the UID. THat way
1023 the session scope can be arranged freely in slices and we don't have
1024 make assumptions about their slice anymore.
1025 - follow PropertiesChanged state more closely, to deal with quick logouts and
1026 relogins
1027 - (optionally?) spawn seat-manager@$SEAT.service whenever a seat shows up that as CanGraphical set
1028
1029 * journal:
1030 - consider introducing implicit _TTY= + _PPID= + _EUID= + _EGID= + _FSUID= + _FSGID= fields
1031 - journald: also get thread ID from client, plus thread name
1032 - journal: when waiting for journal additions in the client always sleep at least 1s or so, in order to minimize wakeups
1033 - add API to close/reopen/get fd for journal client fd in libsystemd-journal.
1034 - fallback to /dev/log based logging in libsystemd-journal, if we cannot log natively?
1035 - declare the local journal protocol stable in the wiki interface chart
1036 - sd-journal: speed up sd_journal_get_data() with transparent hash table in bg
1037 - journald: when dropping msgs due to ratelimit make sure to write
1038 "dropped %u messages" not only when we are about to print the next
1039 message that works, but already after a short timeout
1040 - check if we can make journalctl by default use --follow mode inside of less if called without args?
1041 - maybe add API to send pairs of iovecs via sd_journal_send
1042 - journal: add a setgid "systemd-journal" utility to invoke from libsystemd-journal, which passes fds via STDOUT and does PK access
1043 - journactl: support negative filtering, i.e. FOOBAR!="waldo",
1044 and !FOOBAR for events without FOOBAR.
1045 - journal: store timestamp of journal_file_set_offline() in the header,
1046 so it is possible to display when the file was last synced.
1047 - 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.
1048 - journal: find a way to allow dropping history early, based on priority, other rules
1049 - journal: When used on NFS, check payload hashes
1050 - journald: add kernel cmdline option to disable ratelimiting for debug purposes
1051 - refuse taking lower-case variable names in sd_journal_send() and friends.
1052 - journald: we currently rotate only after MaxUse+MaxFilesize has been reached.
1053 - journal: deal nicely with byte-by-byte copied files, especially regards header
1054 - journal: sanely deal with entries which are larger than the individual file size, but where the components would fit
1055 - Replace utmp, wtmp, btmp, and lastlog completely with journal
1056 - journalctl: instead --after-cursor= maybe have a --cursor=XYZ+1 syntax?
1057 - when a kernel driver logs in a tight loop, we should ratelimit that too.
1058 - journald: optionally, log debug messages to /run but everything else to /var
1059 - journald: when we drop syslog messages because the syslog socket is
1060 full, make sure to write how many messages are lost as first thing
1061 to syslog when it works again.
1062 - journald: allow per-priority and per-service retention times when rotating/vacuuming
1063 - journald: make use of uid-range.h to managed uid ranges to split
1064 journals in.
1065 - journalctl: add the ability to look for the most recent process of a binary. journalctl /usr/bin/X11 --pid=-1 or so...
1066 - improve journalctl performance by loading journal files
1067 lazily. Encode just enough information in the file name, so that we
1068 do not have to open it to know that it is not interesting for us, for
1069 the most common operations.
1070 - man: document that corrupted journal files is nothing to act on
1071 - rework journald sigbus stuff to use mutex
1072 - Set RLIMIT_NPROC for systemd-journal-xyz, and all other of our
1073 services that run under their own user ids, and use User= (but only
1074 in a world where userns is ubiquitous since otherwise we cannot
1075 invoke those daemons on the host AND in a container anymore). Also,
1076 if LimitNPROC= is used without User= we should warn and refuse
1077 operation.
1078 - journalctl --verify: don't show files that are currently being
1079 written to as FAIL, but instead show that their are being written to.
1080 - add journalctl -H that talks via ssh to a remote peer and passes through
1081 binary logs data
1082 - add a version of --merge which also merges /var/log/journal/remote
1083 - journalctl: -m should access container journals directly by enumerating
1084 them via machined, and also watch containers coming and going.
1085 Benefit: nspawn --ephemeral would start working nicely with the journal.
1086 - assign MESSAGE_ID to log messages about failed services
1087 - check if loop in decompress_blob_xz() is necessary
1088
1089 * add a test if all entries in the catalog are properly formatted.
1090 (Adding dashes in a catalog entry currently results in the catalog entry
1091 being silently skipped. journalctl --update-catalog must warn about this,
1092 and we should also have a unit test to check that all our message are OK.)
1093
1094 * document:
1095 - document that deps in [Unit] sections ignore Alias= fields in
1096 [Install] units of other units, unless those units are disabled
1097 - man: clarify that time-sync.target is not only sysv compat but also useful otherwise. Same for similar targets
1098 - document that service reload may be implemented as service reexec
1099 - add a man page containing packaging guidelines and recommending usage of things like Documentation=, PrivateTmp=, PrivateNetwork= and ReadOnlyDirectories=/etc /usr.
1100 - document systemd-journal-flush.service properly
1101 - documentation: recommend to connect the timer units of a service to the service via Also= in [Install]
1102 - man: document the very specific env the shutdown drop-in tools live in
1103 - man: add more examples to man pages,
1104 - in particular an example how to do the equivalent of switching runlevels
1105 - man: maybe sort directives in man pages, and take sections from --help and apply them to man too
1106 - document root=gpt-auto properly
1107
1108 * systemctl:
1109 - add systemctl switch to dump transaction without executing it
1110 - Add a verbose mode to "systemctl start" and friends that explains what is being done or not done
1111 - "systemctl disable" on a static unit prints no message and does
1112 nothing. "systemctl enable" does nothing, and gives a bad message
1113 about it. Should fix both to print nice actionable messages.
1114 - print nice message from systemctl --failed if there are no entries shown, and hook that into ExecStartPre of rescue.service/emergency.service
1115 - add new command to systemctl: "systemctl system-reexec" which reexecs as many daemons as virtually possible
1116 - systemctl enable: fail if target to alias into does not exist? maybe show how many units are enabled afterwards?
1117 - systemctl: "Journal has been rotated since unit was started." message is misleading
1118 - systemctl status output should include list of triggering units and their status
1119
1120 * unit install:
1121 - "systemctl mask" should find all names by which a unit is accessible
1122 (i.e. by scanning for symlinks to it) and link them all to /dev/null
1123
1124 * timer units:
1125 - timer units should get the ability to trigger when:
1126 o DST changes
1127 - Modulate timer frequency based on battery state
1128
1129 * add libsystemd-password or so to query passwords during boot using the password agent logic
1130
1131 * clean up date formatting and parsing so that all absolute/relative timestamps we format can also be parsed
1132
1133 * on shutdown: move utmp, wall, audit logic all into PID 1 (or logind?), get rid of systemd-update-utmp-runlevel
1134
1135 * make repeated alt-ctrl-del presses printing a dump
1136
1137 * hostnamed: before returning information from /etc/machine-info.conf check the modification data and reread. Similar for localed, ...
1138
1139 * currently x-systemd.timeout is lost in the initrd, since crypttab is copied into dracut, but fstab is not
1140
1141 * nspawn:
1142 - emulate /dev/kmsg using CUSE and turn off the syslog syscall
1143 with seccomp. That should provide us with a useful log buffer that
1144 systemd can log to during early boot, and disconnect container logs
1145 from the kernel's logs.
1146 - as soon as networkd has a bus interface, hook up --network-interface=,
1147 --network-bridge= with networkd, to trigger netdev creation should an
1148 interface be missing
1149 - a nice way to boot up without machine id set, so that it is set at boot
1150 automatically for supporting --ephemeral. Maybe hash the host machine id
1151 together with the machine name to generate the machine id for the container
1152 - fix logic always print a final newline on output.
1153 https://github.com/systemd/systemd/pull/272#issuecomment-113153176
1154 - should optionally support receiving WATCHDOG=1 messages from its payload
1155 PID 1...
1156 - optionally automatically add FORWARD rules to iptables whenever nspawn is
1157 running, remove them when shut down.
1158
1159 * dissect
1160 - refuse mounting over a mount point
1161 - automatically discover .roothash files in dissect, similarly to nspawn
1162
1163 * machined:
1164 - add an API so that libvirt-lxc can inform us about network interfaces being
1165 removed or added to an existing machine
1166 - "machinectl migrate" or similar to copy a container from or to a
1167 difference host, via ssh
1168 - introduce systemd-nspawn-ephemeral@.service, and hook it into
1169 "machinectl start" with a new --ephemeral switch
1170 - "machinectl status" should also show internal logs of the container in
1171 question
1172 - "machinectl history"
1173 - "machinectl diff"
1174 - "machinectl commit" that takes a writable snapshot of a tree, invokes a
1175 shell in it, and marks it read-only after use
1176
1177 * cryptsetup:
1178 - cryptsetup-generator: allow specification of passwords in crypttab itself
1179 - support rd.luks.allow-discards= kernel cmdline params in cryptsetup generator
1180
1181 * hw watchdog: optionally try to use the preset watchdog timeout instead of always overriding it
1182 https://bugs.freedesktop.org/show_bug.cgi?id=54712
1183
1184 * add a dependency on standard-conf.xml and other included files to man pages
1185
1186 * MountFlags=shared acts as MountFlags=slave right now.
1187
1188 * properly handle loop back mounts via fstab, especially regards to fsck/passno
1189
1190 * initialize the hostname from the fs label of /, if /etc/hostname does not exist?
1191
1192 * udev:
1193 - move to LGPL
1194 - kill scsi_id
1195 - add trigger --subsystem-match=usb/usb_device device
1196 - reimport udev db after MOVE events for devices without dev_t
1197
1198 * There's currently no way to cancel fsck (used to be possible via C-c or c on the console)
1199
1200 * add option to sockets to avoid activation. Instead just drop packets/connections, see http://cyberelk.net/tim/2012/02/15/portreserve-systemd-solution/
1201
1202 * coredump:
1203 - save coredump in Windows/Mozilla minidump format
1204 - when truncating coredumps, also log the full size that the process had, and make a metadata field so we can report truncated coredumps
1205
1206 * support crash reporting operation modes (https://live.gnome.org/GnomeOS/Design/Whiteboards/ProblemReporting)
1207
1208 * be able to specify a forced restart of service A where service B depends on, in case B
1209 needs to be auto-respawned?
1210
1211 * tmpfiles:
1212 - apply "x" on "D" too (see patch from William Douglas)
1213 - instead of ignoring unknown fields, reject them.
1214 - creating new directories/subvolumes/fifos/device nodes
1215 should not follow symlinks. None of the other adjustment or creation
1216 calls follow symlinks.
1217
1218 * make sure systemd-ask-password-wall does not shutdown systemd-ask-password-console too early
1219
1220 * verify that the AF_UNIX sockets of a service in the fs still exist
1221 when we start a service in order to avoid confusion when a user
1222 assumes starting a service is enough to make it accessible
1223
1224 * Make it possible to set the keymap independently from the font on
1225 the kernel cmdline. Right now setting one resets also the other.
1226
1227 * and a dbus call to generate target from current state
1228
1229 * write blog stories about:
1230 - hwdb: what belongs into it, lsusb
1231 - enabling dbus services
1232 - how to make changes to sysctl and sysfs attributes
1233 - remote access
1234 - how to pass throw-away units to systemd, or dynamically change properties of existing units
1235 - testing with Harald's awesome test kit
1236 - auto-restart
1237 - how to develop against journal browsing APIs
1238 - the journal HTTP iface
1239 - non-cgroup resource management
1240 - dynamic resource management with cgroups
1241 - refreshed, longer missions statement
1242 - calendar time events
1243 - init=/bin/sh vs. "emergency" mode, vs. "rescue" mode, vs. "multi-user" mode, vs. "graphical" mode, and the debug shell
1244 - how to create your own target
1245 - instantiated apache, dovecot and so on
1246 - hooking a script into various stages of shutdown/rearly booot
1247
1248 * investigate whether the gnome pty helper should be moved into systemd, to provide cgroup support.
1249
1250 * dot output for --test showing the 'initial transaction'
1251
1252 * pid1:
1253 - When logging about multiple units (stopping BoundTo units, conflicts, etc.),
1254 log both units as UNIT=, so that journalctl -u triggers on both.
1255 - generate better errors when people try to set transient properties
1256 that are not supported...
1257 http://lists.freedesktop.org/archives/systemd-devel/2015-February/028076.html
1258 - maybe introduce WantsMountsFor=? Usecase:
1259 http://lists.freedesktop.org/archives/systemd-devel/2015-January/027729.html
1260 - recreate systemd's D-Bus private socket file on SIGUSR2
1261 - move PAM code into its own binary
1262 - when we automatically restart a service, ensure we restart its rdeps, too.
1263 - hide PAM options in fragment parser when compile time disabled
1264 - Support --test based on current system state
1265 - 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().
1266 - after deserializing sockets in socket.c we should reapply sockopts and things
1267 - drop PID 1 reloading, only do reexecing (difficult: Reload()
1268 currently is properly synchronous, Reexec() is weird, because we
1269 cannot delay the response properly until we are back, so instead of
1270 being properly synchronous we just keep open the fd and close it
1271 when done. That means clients do not get a successful method reply,
1272 but much rather a disconnect on success.
1273 - when breaking cycles drop sysv services first, then services from /run, then from /etc, then from /usr
1274 - when a bus name of a service disappears from the bus make sure to queue further activation requests
1275
1276 * unit files:
1277 - allow port=0 in .socket units
1278 - maybe introduce ExecRestartPre=
1279 - add ReloadSignal= for configuring a reload signal to use
1280 - implement Register= switch in .socket units to enable registration
1281 in Avahi, RPC and other socket registration services.
1282 - allow Type=simple with PIDFile=
1283 https://bugzilla.redhat.com/show_bug.cgi?id=723942
1284 - allow writing multiple conditions in unit files on one line
1285 - introduce Type=pid-file
1286 - introduce mix of BindTo and Requisite
1287 - add a concept of RemainAfterExit= to scope units
1288 - Allow multiple ExecStart= for all Type= settings, so that we can cover rescue.service nicely
1289 - add verification of [Install] section to systemd-analyze verify
1290
1291 * udev-link-config:
1292 - Make sure ID_PATH is always exported and complete for
1293 network devices where possible, so we can safely rely
1294 on Path= matching
1295
1296 * sd-rtnl:
1297 - add support for more attribute types
1298 - inbuilt piping support (essentially degenerate async)? see loopback-setup.c and other places
1299
1300 * networkd:
1301 - add more keys to [Route] and [Address] sections
1302 - add support for more DHCPv4 options (and, longer term, other kinds of dynamic config)
1303 - add reduced [Link] support to .network files
1304 - properly handle routerless dhcp leases
1305 - work with non-Ethernet devices
1306 - dhcp: do we allow configuring dhcp routes on interfaces that are not the one we got the dhcp info from?
1307 - the DHCP lease data (such as NTP/DNS) is still made available when
1308 a carrier is lost on a link. It should be removed instantly.
1309 - expose in the API the following bits:
1310 - option 15, domain name
1311 - option 12, hostname and/or option 81, fqdn
1312 - option 123, 144, geolocation
1313 - option 252, configure http proxy (PAC/wpad)
1314 - provide a way to define a per-network interface default metric value
1315 for all routes to it. possibly a second default for DHCP routes.
1316 - allow Name= to be specified repeatedly in the [Match] section. Maybe also
1317 support Name=foo*|bar*|baz ?
1318 - duplicate address check for static IPs (like ARPCHECK in network-scripts)
1319 - whenever uplink info changes, make DHCP server send out FORCERENEW
1320
1321 * Figure out how to do unittests of networkd's state serialization
1322
1323 * dhcp:
1324 - figure out how much we can increase Maximum Message Size
1325
1326 * dhcp6:
1327 - add functions to set previously stored IPv6 addresses on startup and get
1328 them at shutdown; store them in client->ia_na
1329 - write more test cases
1330 - implement reconfigure support, see 5.3., 15.11. and 22.20.
1331 - implement support for temporary adressess (IA_TA)
1332 - implement dhcpv6 authentication
1333 - investigate the usefulness of Confirm messages; i.e. are there any
1334 situations where the link changes without any loss in carrier detection
1335 or interface down
1336 - some servers don't do rapid commit without a filled in IA_NA, verify
1337 this behavior
1338 - RouteTable= ?
1339
1340 External:
1341
1342 * dbus:
1343 - natively watch for dbus-*.service symlinks (PENDING)
1344 - teach dbus to activate all services it finds in /etc/systemd/services/org-*.service
1345
1346 * make cryptsetup lower --iter-time
1347
1348 * kernel: add device_type = "fb", "fbcon" to class "graphics"
1349
1350 * /usr/bin/service should actually show the new command line
1351
1352 * 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=
1353
1354 * neither pkexec nor sudo initialize environ[] from the PAM environment?
1355
1356 * fedora: update policy to declare access mode and ownership of unit files to root:root 0644, and add an rpmlint check for it
1357
1358 * register catalog database signature as file magic
1359
1360 * zsh shell completion:
1361 - <command> <verb> -<TAB> should complete options, but currently does not
1362 - systemctl add-wants,add-requires
1363
1364 * systemctl status should know about 'systemd-analyze calendar ... --iterations='
1365 * If timer has just OnInactiveSec=..., it should fire after a specified time
1366 after being started.
1367
1368 Regularly:
1369
1370 * look for close() vs. close_nointr() vs. close_nointr_nofail()
1371
1372 * check for strerror(r) instead of strerror(-r)
1373
1374 * pahole
1375
1376 * set_put(), hashmap_put() return values check. i.e. == 0 does not free()!
1377
1378 * use secure_getenv() instead of getenv() where appropriate
1379
1380 * link up selected blog stories from man pages and unit files Documentation= fields
1381 String is not UTF-8 clean, ignoring assignment