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