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.
7 * userdbctl: "Password OK: yes" is shown even when there are no passwords
8 or the password is locked.
10 * Get rid of nftw(). We should refuse to use such useless APIs on principle.
14 * Fedora: add an rpmlint check that verifies that all unit files in the RPM are listed in %systemd_post macros.
18 * Rearrange tests so that the various test-xyz.c match a specific src/basic/xyz.c again
20 * rework mount.c and swap.c to follow proper state enumeration/deserialization
21 semantics, like we do for device.c now
25 * maybe add a tool that displays most recent journal logs as QR code to scan
26 off screen and run it automatically on boot failures, emergency logs and
27 such. Use DRM APIs directly, see
28 https://github.com/dvdhrm/docs/blob/master/drm-howto/modeset.c for an example
31 * pass systemd-detect-virt result to generators as env var. Modifying behaviour
32 based on whether we are virtualized or not is a pretty common thing, hence
33 maybe just pass that info along for free in an env var. We cache the result
34 anyway, so it's basically free.
36 * systemd-repart: read LUKS encryption key from $CREDENTIALS_PATH
38 * introduce /dev/disk/root/* symlinks that allow referencing partitions on the
39 disk the rootfs is on in a reasonably secure way.
41 * systemd-repart: add a switch to factory reset the partition table without
42 immediately applying the new configuration again. i.e. --factory-reset=leave
43 or so. (this is useful to factory reset an image, then putting it into
44 another machine, ensuring that luks key is generated on new machine, not old)
46 * move logind udev rules to top-level rule.d/ directory
48 * move multiseat vid/pid matches from logind udev rule to hwdb
50 * nspawn: default to 1:1 userns
52 * Provide a reasonably bespoke solution for mounting host $HOME directories
54 • add new option --mount-user=$USER for mounting $HOME of the user into the
55 container at the same place
56 • check /etc/passwd for UID or user name clashes. If UID clash pick a different
57 UID in container, and map via userns. If user name clash, refuse. If
58 matching user already exists use that.
59 • otherwise: write user record of specified user into /run/host/passwd or so
60 • in nss-systemd pick up user record from there and make available to system
61 With all that in place if nspawn host and container payload are up-to-date
62 enough we have a very simple way to make host users available in containers.
64 * whenever we receive fds via SCM_RIGHTS make sure none got dropped due to the
65 reception limit the kernel silently enforces.
67 * add an Open= setting to service unit files that can open arbitrary file
68 system paths at service startup time and pass them to the service process via
69 our usual socket activation protocol. If passed path refers to AF_UNIX
70 socket: connect() to it.
72 * add a ConnectSocket= setting to service unit files, that may reference a
73 socket unit, and which will connect to the socket defined therein, and pass
74 the resulting fd to the service program via socket activation proto.
76 * Add a concept of ListenStream=anonymous to socket units: listen on a socket
77 that is deleted in the fs. Usecase would be with ConnectSocket= above.
79 * Hook up journald's FSS logic with TPM2: seal the verification disk by
80 time-based policy, so that the verification key can remain on host and ve
83 * sd-boot: define a drop-in dir in the ESP that may contain X.509
84 certificates. If the firmware is detected to be in setup mode, automatically
85 enroll them as PK/KEK/db, turn off setup mode and proceed. Optionally,
86 instead of auto-enrolling them add them to the sd-boot menu, giving the user
87 the option to manually enroll them, after selecting the menu entry. This way,
88 installer images can just drop the certfiicates in the ESP, and on first boot
89 can easily enroll the keys without ever booting up.
91 * efi stub: optionally, load initrd from disk as a separate file, HMAC check it
92 with key from TPM, bound to PCR, refusing if failing. This would then allow
93 traditional distros that generate initrds locally to secure them with TPM:
94 after generating the initrd, do the HMAC calculation, put result in initrd
95 filename, done. This would then bind the validity of the initrd to the local
96 host, and used kernel, and means people cannot change initrd or kernel
97 without booting the kernel + initrd.
99 * importd: add ability download images for portabled + sysext
101 * importd: support image signature verification with PKCS#7 + OpenBSD signify
102 logic, as alternative to crummy gpg
104 * sd-boot: add service that automatically runs "bootctl update" on every boot,
105 in a graceful way, so that updated /usr trees automatically propagate into
106 updated boot loaders on reboot.
108 * sysext: optionally, if the merged trees allow it use bind mounts instead of
111 * nspawn: add support for sysext extensions, too. i.e. a new --extension=
112 switch that takes one or more arguments, and applies the extensions already
115 * add "systemd-analyze debug" + AttachDebugger= in unit files: The former
116 specifies a command to execute; the latter specifies that an already running
117 "systemd-analyze debug" instance shall be contacted and execution paused
118 until it gives an OK. That way, tools like gdb or strace can be safely be
119 invoked on processes forked off PID 1.
121 * expose MS_NOSYMFOLLOW in various places
123 * tpm2: support a PIN policy, i.e. allowing windows-style short authentication
124 passwords by using the TPM2 to enforce ratelimiting and such, use for
127 * Add concept for upgrading TPM2 enrollments, maybe a new switch
128 --pcrs=4:<hash> or so, i.e. select a PCR to include in the hash, and then
131 * homed: store PKCS#11 + FIDO2 token info in LUKS2 header, compatible with
132 systemd-cryptsetup, so that it can unlock homed volumes
134 * cryptenroll: politely refuse enrolling new keys to homed volumes, since we
135 we cannot update identity info
137 * TPM2: auto-reenroll in cryptsetup, as fallback for hosed firmware upgrades
140 * cryptsetup: if only recovery keys are registered and no regular passphrases,
141 ask user for "recovery key", not "passphrase"
143 * cyptsetup: add option for automatically removing empty password slot on boot
145 * cryptsetup: optionally, when run during boot-up and password is never
146 entered, and we are on battery power (or so), power off machine again
148 * cryptsetup: when FIDO2/PKCS#11/TPM2 token/chip didn't show up after some
149 time, abort the attempt, fallback to asking for pw
151 * cryptsetup: when waiting for FIDO2/PKCS#11 token, tell plymouth that, and
152 allow plymouth to abort the waiting and enter pw instead
154 * when configuring loopback netif, and it fails due to EPERM, eat up error if
155 it happens to be set up alright already.
157 * at boot: check if battery above some threshold, if not power off again after explanation
159 * userdb: add field for ambient caps, so that a user can have CAP_WAKE_ALARM
160 for example. And add code that resets ambient caps for all services by
163 * homed: try to unmount in regular intervals when home dir was busy when we
166 * sd-bus: when connecting to some dbus server socker, set originating AF_UNIX
167 socket name in abstract namespace to include "description" string, and pick
168 it up from there in sd_bus_creds logic. i.e. we can use the socket peer
169 address as conduit for some minimal connection metainfo, and use it to
170 restore the "description" logic that kdbus used to have.
172 * teach LoadCredential= the ability to load all files from a specified dir as
175 * systemd-analyze netif that explains predictable interface (or networkctl)
177 * port selinux code from mallinfo() to mallinfo2() once added to glibc
179 * Add service setting to run a service within the specified VRF. i.e. do the
180 equivalent of "ip vrf exec".
182 * export action of device object on sd-device, so that monitor becomes useful
184 * add root=tmpfs that mounts a tmpfs to /sysroot (to be used in combination
185 with usr=…, for a similar effect as systemd.volatile=yes but without the
186 "hide-out" effect). Also, add root=gpt-auto-late support or so, that is like
187 root=gpt-auto but initially mounts a tmpfs to /sysroot, and then revisits
188 later after systemd-repart ran. Usecase: let's ship images with only /usr
189 partition, then on first boot create the root partition. In this case we want
190 to read the repart data from /usr before the root partition exists. Add
191 usr=gpt-auto that automatically finds a /usr partition.
193 * change SwitchRoot() implementation in PID 1 to use pivot_root(".", "."), as
194 documented in the pivot_root(2) man page, so that we can drop the /oldroot
197 * special case some calls of chase_symlinks() to use openat2() internally, so
198 that the kernel does what we otherwise do.
200 * homed: keep an fd to the homedir open at all times, to keep the fs pinned
201 (autofs and such) while user is logged in.
203 * nss-systemd: also synthesize shadow records for users/groups
205 * make use of new glibc 2.32 APIs sigabbrev_np() and strerrorname_np().
207 * when main nspawn supervisor process gets suspended due to SIGSTOP/SIGTTOU or
208 so, freeze the payload too.
210 * repart: support setting up dm-integrity with HMAC
212 * add /etc/integritytab, to support dm-integrity setups. In particular those
213 with HMAC as hash function, so that we can have a protected /home without
214 encryption (leaving encryption to the individual dirs/homed).
216 * complement root=, rootflags=, rootfstype= with rootsubdir= which allows
217 mounting a subdir of the root fs as actual root. This can be used as
218 fstype-agnostic version of btrfs' rootflags=subvol=foobar.
220 * if /usr/bin/swapoff fails due to OOM, log a friendly explanatory message about it
222 * build short web pages out of each catalog entry, build them along with man
223 pages, and include hyperlinks to them in the journal output
225 * machined: add API to acquire UID range. add API to mount/dissect loopback
226 file. Both protected by PK. Then make nspawn use these APIs to run
227 unprivileged containers. i.e. push the truly privileged bits into machined,
228 so that the client side can remain entirely unprivileged, with SUID or
231 * journald: do journal file writing out-of-process, with one writer process per
232 client UID, so that synthetic hash table collisions can slow down a specific
233 user's journal stream down but not the others.
235 * nspawn: support time namespaces
237 * systemd-firstboot: make sure to always use chase_symlinks() before
238 reading/writing files
240 * Remove any support for booting without /usr pre-mounted in the initrd entirely.
241 Update INITRD_INTERFACE.md accordingly.
243 * pid1: Move to tracking of main pid/control pid of units per pidfd
245 * pid1: support new clone3() fork-into-cgroup feature
247 * pid1: also remove PID files of a service when the service starts, not just
250 * make us use dynamically fewer deps for containers in general purpose distros:
251 o turn into dlopen() deps:
253 - p11-kit-trust (always)
254 - kmod-libs (only when called from PID 1)
255 - libblkid (only in RootImage= handling in PID 1, but not elsewhere)
256 - libpam (only when called from PID 1)
257 - bzip2, xz, lz4 (always — gzip and zstd should probably stay static deps the way they are,
258 since they are so basic and our defaults)
259 o move into separate libsystemd-shared-iptables.so .so
260 - iptables-libs (only used by nspawn + networkd)
262 * seccomp: maybe use seccomp_merge() to merge our filters per-arch if we can.
263 Apparently kernel performance is much better with fewer larger seccomp
264 filters than with more smaller seccomp filters.
266 * systemd-path: add ESP and XBOOTLDR path. Add "private" runtime/state/cache dir enum,
267 mapping to $RUNTIME_DIRECTORY, $STATE_DIRECTORY and such
269 * All tools that support --root= should also learn --image= so that they can
270 operate on disk images directly. Specifically: bootctl, systemctl,
271 coredumpctl. (Already done: systemd-nspawn, systemd-firstboot,
272 systemd-repart, systemd-tmpfiles, systemd-sysusers, journalctl)
274 * seccomp: by default mask x32 ABI system wide on x86-64. it's on its way out
276 * seccomp: don't install filters for ABIs that are masked anyway for the
279 * seccomp: maybe merge all filters we install into one with that libseccomp API that allows merging.
281 * credentials system:
282 - maybe add AcquireCredential= for querying a cred via ask-password
283 - maybe try to acquire creds via keyring?
284 - maybe try to pass creds via keyring?
285 - maybe optionally pass creds via memfd
286 - maybe add support for decrypting creds via TPM
287 - maybe add support for decrypting/importing creds via pkcs11
288 - make systemd-cryptsetup acquire pw via creds logic
289 - make PAMName= acquire pw via creds logic
290 - make macsec/wireguard code in networkd read key via creds logic
291 - make gatwayd/remote read key via creds logic
292 - add sd_notify() command for flushing out creds not needed anymore
294 * homed: during login resize fs automatically towards size goal. Specifically,
295 resize to diskSize if possible, but leave a certain amount (configured by a
296 new value diskLeaveFreeSize) of space free on the backing fs.
298 * homed: permit multiple user record signing keys to be used locally, and pick
299 the right one for signing records automatically depending on a pre-existing
302 * homed: add a way to "adopt" a home directory, i.e. strip foreign signatures
303 and insert a local signature instead.
305 * homed: as an extension to the directory+subvolume backend: if located on
306 especially marked fs, then sync down password into LUKS header of that fs,
307 and always verify passwords against it too. Bootstrapping is a problem
308 though: if no one is logged in (or no other user even exists yet), how do you
309 unlock the volume in order to create the first user and add the first pw.
311 * homed: support new FS_IOC_ADD_ENCRYPTION_KEY ioctl for setting up fscrypt
313 * homed: maybe pre-create ~/.cache as subvol so that it can have separate quota
316 * busctl: maybe expose a verb "ping" for pinging a dbus service to see if it
319 * systemd-gpt-auto should probably set x-systemd.growfs on the mounts it
323 - teach it to prepare an ESP wholesale, i.e. with mkfs.vfat invocation
324 - teach it to copy in unified kernel images and maybe type #1 boot loader spec entries from host
325 - make it operate on loopback files, dissecting enough to find ESP to operate on
327 * Maybe add a separate GPT partition type to the discoverable partition spec
328 for "hibernate" partitions, that are exactly like swap partitions but only
329 activated right before hibernation and thus never used for regular swapping.
331 * by default, in systemd --user service bump the OOMAdjust to 100, as privs
332 allow so that systemd survives
334 * cryptsetup: allow encoding key directly in /etc/crypttab, maybe with a
335 "base64:" prefix. Useful in particular for pkcs11 mode.
337 * cryptsetup: reimplement the mkswap/mke2fs in cryptsetup-generator to use
338 systemd-makefs.service instead.
340 * socket units: allow creating a udev monitor socket with ListenDevices= or so,
341 with matches, then activate app through that passing socket over
343 * unify on openssl (as soon as OpenSSL 3.0 is out, and the Debian license
345 - port sd_id128_get_machine_app_specific() over from khash
346 - port resolved over from libgcrypt (DNSSEC code)
347 - port journald + fsprg over from libgcrypt
348 - port importd over from libgcrypt
349 - when that's done: kill khash.c
350 - when that's done: kill gnutls support in resolved
352 * when we resize disks (homed?) always round up to 4K sectors, not 512K
354 * add growvol and makevol options for /etc/crypttab, similar to
355 x-systemd.growfs and x-systemd-makefs.
357 * systemd-repart: by default generate minimized partition tables (i.e. tables
358 that only cover the space actually used, excluding any free space at the
359 end), in order to maximize dd'ability. Requires libfdisk work, see
360 https://github.com/karelzak/util-linux/issues/907
362 * systemd-repart: MBR partition table support. Care needs to be taken regarding
363 Type=, so that partition definitions can sanely apply to both the GPT and the
364 MBR case. Idea: accept syntax "Type=gpt:home mbr:0x83" for setting the types
365 for the two partition types explicitly. And provide an internal mapping so
366 that "Type=linux-generic" maps to the right types for both partition tables
369 * systemd-repart: allow sizing partitions as factor of available RAM, so that
370 we can reasonably size swap partitions for hibernation.
372 * systemd-repart: allow managing the gpt read-only partition flag + auto-mount flag
374 * systemd-repart: allow boolean option that ensures that if existing partition
375 doesn't exist within the configured size bounds the whole command fails. This
376 is useful to implement ESP vs. XBOOTLDR schemes in installers: have one set
377 of repart files for the case where ESP is large enough and one where it isn't
378 and XBOOTLDR is added in instead. Then apply the former first, and if it
379 fails to apply use the latter.
381 * systemd-repart: add per-partition option to never reuse existing partition
382 and always create anew even if matching partition already exists.
384 * systemd-repart: add per-partition option to fail if partition already exist,
385 i.e. is not added new. Similar, add option to fail if partition does not exist yet.
387 * systemd-repart: allow disabling growing of specific partitions, or making
388 them (think ESP: we don't ever want to grow it, since we cannot resize vfat)
390 * systemd-repart: make it a static checker during early boot for existence and
391 absence of other partitions for trusted boot environments
393 * userdb: allow username prefix searches in varlink API, allow realname and
394 realname substr searches in varlink API
396 * userdb: allow uid/gid range checks
398 * userdb: allow existence checks
400 * pid1: activation by journal search expression
402 * when switching root from initrd to host, set the machine_id env var so that
403 if the host has no machine ID set yet we continue to use the random one the
406 * sd-event: add native support for P_ALL waitid() watching, then move PID 1 to
407 it for reaping assigned but unknown children. This needs to some special care
408 to operate somewhat sensibly in light of priorities: P_ALL will return
409 arbitrary processes, regardless of the priority we want to watch them with,
410 hence on each event loop iteration check all processes which we shall watch
411 with higher prio explicitly, and then watch the entire rest with P_ALL.
413 * tweak sd-event's child watching: keep a prioq of children to watch and use
414 waitid() only on the children with the highest priority until one is waitable
415 and ignore all lower-prio ones from that point on
417 * maybe introduce xattrs that can be set on the root dir of the root fs
418 partition that declare the volatility mode to use the image in. Previously I
419 thought marking this via GPT partition flags but that's not ideal since
420 that's outside of the LUKS encryption/verity verification, and we probably
421 shouldn't operate in a volatile mode unless we got told so from a trusted
424 * figure out automatic partition discovery when combining writable root dir
427 * coredump: maybe when coredumping read a new xattr from /proc/$PID/exe that
428 may be used to mark a whole binary as non-coredumpable. Would fix:
429 https://bugs.freedesktop.org/show_bug.cgi?id=69447
431 * teach parse_timestamp() timezones like the calendar spec already knows it
433 * beef up hibernation to optionally do swapon/swapoff immediately before/after
436 * beef up s2h to implement a battery watch loop: instead of entering
437 hibernation unconditionally after coming back from resume make a decision
438 based on the battery load level: if battery level is above a specific
439 threshold, go to suspend again, only hibernate if below it. This means we'd
440 stick to suspend usually, but fall back to hibernation only when battery runs
441 empty (well, subject to our sampling interval). Related to this, check if we
442 can make ACPI _BTP (i.e. /sys/class/power_supply/*/alarm) work for us too,
443 i.e. see if it can wake up machines from suspend, so that we could resume
444 automatically when the system is low on power and move automatically to
445 hibernation mode. (see
446 https://uefi.org/sites/default/files/resources/ACPI%206_2_A_Sept29.pdf
448 https://docs.microsoft.com/en-us/windows-hardware/design/device-experiences/modern-standby-wake-sources
451 * add an explicit "vertical" mode to format-table, so that "systemctl
452 status"-like outputs (i.e. with a series of field names left and values
453 right) become genuine first class citizens, and we gain automatic, sane JSON
456 * We should probably replace /var/log/README, /etc/rc.d/README with symlinks
457 that are linked to these places instead of copied. After all they are
458 constant vendor data.
460 * maybe add kernel cmdline params: to force random seed crediting
462 * nspawn: on cgroupsv1 issue cgroup empty handler process based on host events,
463 so that we make cgroup agent logic safe
465 * nspawn/machined: add API to invoke binary in container, then use that as
466 fallback in "machinectl shell"
468 * logind: rework pam_logind to also do a bus call in case of invocation from
469 user@.service, which returns the XDG_RUNTIME_DIR value, and make this
470 behaviour selectable via pam module option.
473 - when user tries to log into record signed by unrecognized key, automatically add key to our chain after polkit auth
474 - rollback when resize fails mid-operation
475 - GNOME's side for forget key on suspend (requires rework so that lock screen runs outside of uid)
477 - shrink fs on logout?
478 - update LUKS password on login if we find there's a password that unlocks the JSON record but not the LUKS device.
479 - create on activate?
480 - properties: icon url?, preferred session type?, administrator bool (which translates to 'wheel' membership)?, address?, telephone?, vcard?, samba stuff?, parental controls?
481 - communicate clearly when usb stick is safe to remove. probably involves
482 beefing up logind to make pam session close hook synchronous and wait until
483 systemd --user is shut down.
484 - logind: maybe keep a "busy fd" as long as there's a non-released session around or the user@.service
485 - maybe make automatic, read-only, time-based reflink-copies of LUKS disk
486 images (and btrfs snapshots of subvolumes) (think: time machine)
487 - distinguish destroy / remove (i.e. currently we can unregister a user, unregister+remove their home directory, but not just remove their home directory)
488 - in systemd's PAMName= logic: query passwords with ssh-askpassword, so that we can make "loginctl set-linger" mode work
489 - fingerprint authentication, pattern authentication, …
490 - make sure "classic" user records can also be managed by homed
491 - make size of $XDG_RUNTIME_DIR configurable in user record
492 - query password from kernel keyring first
493 - update even if record is "absent"
494 - add a "access mode" + "fstype" field to the "status" section of json identity records reflecting the actually used access mode and fstype, even on non-luks backends
495 - move acct mgmt stuff from pam_systemd_home to pam_systemd?
496 - when "homectl --pkcs11-token-uri=" is used, synthesize ssh-authorized-keys records for all keys we have private keys on the stick for
497 - make slice for users configurable (requires logind rework)
498 - logind: populate auto-login list bus property from PKCS#11 token
499 - when determining state of a LUKS home directory, check DM suspended sysfs file
500 - introduce API for "making room", that grows/shrinks home directory
501 according to elastic parameters, discards blocks, and removes additional snapshots. Call it
502 either from UI when disk space gets low
504 * introduce a new per-process uuid, similar to the boot id, the machine id, the
505 invocation id, that is derived from process creds, specifically a hashed
506 combination of AT_RANDOM + getpid() + the starttime from
507 /proc/self/status. Then add these ids implicitly when logging. Deriving this
508 uuid from these three things has the benefit that it can be derived easily
509 from /proc/$PID/ in a stable, and unique way that changes on both fork() and
512 * let's not GC a unit while its ratelimits are still pending
514 * when killing due to service watchdog timeout maybe detect whether target
515 process is under ptracing and then log loudly and continue instead.
517 * introduce a new group to own TPM devices
519 * make rfkill uaccess controllable by default, i.e. steal rule from
520 gnome-bluetooth and friends
522 * tweak journald context caching. In addition to caching per-process attributes
523 keyed by PID, cache per-cgroup attributes (i.e. the various xattrs we read)
524 keyed by cgroup path, and guarded by ctime changes. This should provide us
525 with a nice speed-up on services that have many processes running in the same
528 * make MAINPID= message reception checks even stricter: if service uses User=,
529 then check sending UID and ignore message if it doesn't match the user or
532 * maybe trigger a uevent "change" on a device if "systemctl reload xyz.device"
535 * when importing an fs tree with machined, optionally apply userns-rec-chown
537 * when importing an fs tree with machined, complain if image is not an OS
539 * Maybe introduce a helper safe_exec() or so, which is to execve() which
540 safe_fork() is to fork(). And then make revert the RLIMIT_NOFILE soft limit
541 to 1K implicitly, unless explicitly opted-out.
543 * rework seccomp/nnp logic that even if User= is used in combination with
544 a seccomp option we don't have to set NNP. For that, change uid first whil
545 keeping CAP_SYS_ADMIN, then apply seccomp, the drop cap.
547 * add a concept for automatically loading per-unit secrets off disk and
548 inserting them into the kernel keyring. Maybe SecretsDirectory= similar to
549 ConfigurationDirectory=.
551 * when no locale is configured, default to UEFI's PlatformLang variable
553 * bootctl,sd-boot: actually honour the "architecture" key
555 * add a new syscall group "@esoteric" for more esoteric stuff such as bpf() and
556 usefaultd() and make systemd-analyze check for it.
558 * paranoia: whenever we process passwords, call mlock() on the memory
559 first. i.e. look for all places we use free_and_erasep() and
560 augment them with mlock(). Also use MADV_DONTDUMP.
562 * Move RestrictAddressFamily= to the new cgroup create socket
564 * support the bind/connect/sendmsg cgroup stuff for sandboxing, and possibly
567 * maybe implicitly attach monotonic+realtime timestamps to outgoing messages in
568 log.c and sd-journal-send
570 * optionally: turn on cgroup delegation for per-session scope units
572 * introduce per-unit (i.e. per-slice, per-service) journal log size limits.
574 * sd-boot: automatically load EFI modules from some drop-in dir, so that people
575 can add in file system drivers and such
577 * sd-boot: optionally, show boot menu when previous default boot item has
578 non-zero "tries done" count
580 * introduce an option (or replacement) for "systemctl show" that outputs all
581 properties as JSON, similar to busctl's new JSON output. In contrast to that
582 it should skip the variant type string though.
584 * augment CODE_FILE=, CODE_LINE= with something like CODE_BASE= or so which
585 contains some identifier for the project, which allows us to include
586 clickable links to source files generating these log messages. The identifier
587 could be some abberviated URL prefix or so (taking inspiration from Go
588 imports). For example, for systemd we could use
589 CODE_BASE=github.com/systemd/systemd/blob/98b0b1123cc or so which is
590 sufficient to build a link by prefixing "http://" and suffixing the
593 * Augment MESSAGE_ID with MESSAGE_BASE, in a similar fashion so that we can
594 make clickable links from log messages carrying a MESSAGE_ID, that lead to
595 some explanatory text online.
597 * maybe extend .path units to expose fanotify() per-mount change events
599 * Add a "systemctl list-units --by-slice" mode or so, which rearranges the
600 output of "systemctl list-units" slightly by showing the tree structure of
601 the slices, and the units attached to them.
603 * the a-posteriori stopping of units bound to units that disappeared logic
604 should be reworked: there should be a queue of units, and we should only
605 enqueue stop jobs from a defer event that processes queue instead of
606 right-away when we find a unit that is bound to one that doesn't exist
607 anymore. (similar to how the stop-unneeded queue has been reworked the same
610 * nspawn: make nspawn suitable for shell pipelines: instead of triggering a
611 hangup when input is finished, send ^D, which synthesizes an EOF. Then wait
612 for hangup or ^D before passing on the EOF.
614 * When reloading configuration PID 1 should reset all its properties to the
615 original defaults before calling parse_config()
617 * nspawn: greater control over selinux label?
619 * hibernate/s2h: make this robust and safe to enable in Fedora by default.
622 1. add resume_offset support to the resume code (i.e. support swap files
624 2. check if swap is on weird storage and refuse if so
625 3. add auto-detection of hibernation images
627 * cgroups: use inotify to get notified when somebody else modifies cgroups
628 owned by us, then log a friendly warning.
630 * beef up log.c with support for stripping ANSI sequences from strings, so that
631 it is OK to include them in log strings. This would be particularly useful so
632 that our log messages could contain clickable links for example for unit
633 files and suchlike we operate on.
635 * add support for "portablectl attach http://foobar.com/waaa.raw (i.e. importd integration)
637 * add attach --enable and attach --now (for attach+enable+start)
639 * sync dynamic uids/gids between host+portable srvice (i.e. if DynamicUser=1 is set for a service, make sure that the
640 selected user is resolvable in the service even if it ships its own /etc/passwd)
642 * Fix DECIMAL_STR_MAX or DECIMAL_STR_WIDTH. One includes a trailing NUL, the
643 other doesn't. What a disaster. Probably to exclude it. Also
644 DECIMAL_STR_WIDTH should probably add an extra "-" into account for negative
647 * Check that users of inotify's IN_DELETE_SELF flag are using it properly, as
648 usually IN_ATTRIB is the right way to watch deleted files, as the former only
649 fires when a file is actually removed from disk, i.e. the link count drops to
650 zero and is not open anymore, while the latter happens when a file is
651 unlinked from any dir.
653 * port systemctl, busctl, … over to format-table.[ch]'s table formatters
655 * pid1: lock image configured with RootDirectory=/RootImage= using the usual nspawn semantics while the unit is up
657 * add --vacuum-xyz options to coredumpctl, matching those journalctl already has.
659 * introduce Ephemeral= unit file switch, that creates an ephemeral copy of all
660 files and directories that are left writable for a unit, and which are
661 removed after the unit goes down again. A bit like --ephemeral for
662 systemd-nspawn but for system services. If used together with RootImage= this
663 should reflink the image file itself.
665 Related: add Ephemeral=<path1> <path2> … which would allow marking
666 specific paths only like this.
668 * add CopyFile= or so as unit file setting that may be used to copy files or
669 directory trees from the host to the services RootImage= and RootDirectory=
670 environment. Which we can use for /etc/machine-id and in particular
671 /etc/resolv.conf. Should be smart and do something useful on read-only
672 images, for example fall back to read-only bind mounting the file instead.
674 * show invocation ID in systemd-run output
676 * bypass SIGTERM state in unit files if KillSignal is SIGKILL
678 * add proper dbus APIs for the various sd_notify() commands, such as MAINPID=1
679 and so on, which would mean we could report errors and such.
681 * teach tmpfiles.d q/Q logic something sensible in the context of XFS/ext4
684 * introduce DefaultSlice= or so in system.conf that allows changing where we
685 place our units by default, i.e. change system.slice to something
686 else. Similar, ManagerSlice= should exist so that PID1's own scope unit could
687 be moved somewhere else too. Finally machined and logind should get similar
688 options so that it is possible to move user session scopes and machines to a
689 different slice too by default. Usecase: people who want to put resources on
690 the entire system, with the exception of one specific service. See:
691 https://lists.freedesktop.org/archives/systemd-devel/2018-February/040369.html
693 * maybe rework get_user_creds() to query the user database if $SHELL is used
694 for root, but only then.
696 * be stricter with fds we receive for the fdstore: close them asynchronously
698 * calenderspec: add support for week numbers and day numbers within a
699 year. This would allow us to define "bi-weekly" triggers safely.
701 * sd-bus: add vtable flag, that may be used to request client creds implicitly
702 and asynchronously before dispatching the operation
704 * sd-bus: parse addresses given in sd_bus_set_addresses immediately and not
705 only when used. Add unit tests.
707 * make use of ethtool veth peer info in machined, for automatically finding out
708 host-side interface pointing to the container.
710 * add some special mode to LogsDirectory=/StateDirectory=… that allows
711 declaring these directories without necessarily pulling in deps for them, or
712 creating them when starting up. That way, we could declare that
713 systemd-journald writes to /var/log/journal, which could be useful when we
714 doing disk usage calculations and so on.
716 * taint systemd if there are fewer than 65536 users assigned (userns) to the system.
718 * deprecate RootDirectoryStartOnly= in favour of a new ExecStart= prefix char
720 * add a new RuntimeDirectoryPreserve= mode that defines a similar lifecycle for
721 the runtime dir as we maintain for the fdstore: i.e. keep it around as long
722 as the unit is running or has a job queued.
724 * support projid-based quota in machinectl for containers
726 * add a way to lock down cgroup migration: a boolean, which when set for a unit
727 makes sure the processes in it can never migrate out of it
729 * blog about fd store and restartable services
731 * document Environment=SYSTEMD_LOG_LEVEL=debug drop-in in debugging document
733 * rework ExecOutput and ExecInput enums so that EXEC_OUTPUT_NULL loses its
734 magic meaning and is no longer upgraded to something else if set explicitly.
736 * in the long run: permit a system with /etc/machine-id linked to /dev/null, to
737 make it lose its identity, i.e. be anonymous. For this we'd have to patch
738 through the whole tree to make all code deal with the case where no machine
741 * optionally, collect cgroup resource data, and store it in per-unit RRD files,
742 suitable for processing with rrdtool. Add bus API to access this data, and
743 possibly implement a CPULoad property based on it.
745 * beef up pam_systemd to take unit file settings such as cgroups properties as
748 * maybe hook of xfs/ext4 quotactl() with services? i.e. automatically manage
749 the quota of the user indicated in User= via unit file settings, like the
750 other resource management concepts. Would mix nicely with DynamicUser=1. Or
751 alternatively, do this with projids, so that we can also cover services
752 running as root. Quota should probably cover all the special dirs such as
753 StateDirectory=, LogsDirectory=, CacheDirectory=, as well as RootDirectory= if it
754 is set, plus the whole disk space any image configured with RootImage=.
756 * In DynamicUser= mode: before selecting a UID, use disk quota APIs on relevant
757 disks to see if the UID is already in use.
759 * add "systemctl wait" or so, which does what "systemd-run --wait" does, but
760 for all units. It should be both a way to pin units into memory as well as a
761 wait to retrieve their exit data.
763 * expose IO accounting data on the bus, show it in systemd-run --wait and log
764 about it in the resource log message
766 * show whether a service has out-of-date configuration in "systemctl status" by
767 using mtime data of ConfigurationDirectory=.
769 * Add AddUser= setting to unit files, similar to DynamicUser=1 which however
770 creates a static, persistent user rather than a dynamic, transient user. We
771 can leverage code from sysusers.d for this.
773 * add some optional flag to ReadWritePaths= and friends, that has the effect
774 that we create the dir in question when the service is started. Example:
776 ReadWritePaths=:/var/lib/foobar
778 * maybe add call sd_journal_set_block_timeout() or so to set SO_SNDTIMEO for
779 the sd-journal logging socket, and, if the timeout is set to 0, sets
780 O_NONBLOCK on it. That way people can control if and when to block for
783 * hostnamed: populate form factor data from a new hwdb database, so that old
784 yogas can be recognized as "convertible" too, even if they predate the DMI
785 "convertible" form factor
787 * Add ExecMonitor= setting. May be used multiple times. Forks off a process in
788 the service cgroup, which is supposed to monitor the service, and when it
789 exits the service is considered failed by its monitor.
791 * track the per-service PAM process properly (i.e. as an additional control
792 process), so that it may be queried on the bus and everything.
794 * add a new "debug" job mode, that is propagated to unit_start() and for
795 services results in two things: we raise SIGSTOP right before invoking
796 execve() and turn off watchdog support. Then, use that to implement
797 "systemd-gdb" for attaching to the start-up of any system service in its
800 * gpt-auto logic: related to the above, maybe support a "secondary" root
801 partition, that is mounted to / and is writable, and where the actual root's
802 /usr is mounted into.
804 * 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
806 * drop nss-myhostname in favour of nss-resolve?
808 * add a percentage syntax for TimeoutStopSec=, e.g. TimeoutStopSec=150%, and
809 then use that for the setting used in user@.service. It should be understood
810 relative to the configured default value.
812 * in networkd, when matching device types, fix up DEVTYPE rubbish the kernel passes to us
814 * enable LockMLOCK to take a percentage value relative to physical memory
816 * Permit masking specific netlink APIs with RestrictAddressFamily=
818 * nspawn: support that /proc, /sys/, /dev are pre-mounted
820 * define gpt header bits to select volatility mode
822 * ProtectClock= (drops CAP_SYS_TIMES, adds seecomp filters for settimeofday, adjtimex), sets DeviceAllow o /dev/rtc
824 * ProtectTracing= (drops CAP_SYS_PTRACE, blocks ptrace syscall, makes /sys/kernel/tracing go away)
826 * ProtectMount= (drop mount/umount/pivot_root from seccomp, disallow fuse via DeviceAllow, imply Mountflags=slave)
828 * ProtectKeyRing= to take keyring calls away
830 * RemoveKeyRing= to remove all keyring entries of the specified user
832 * ProtectReboot= that masks reboot() and kexec_load() syscalls, prohibits kill
833 on PID 1 with the relevant signals, and makes relevant files in /sys and
834 /proc (such as the sysrq stuff) unavailable
836 * make sure the ratelimit object can deal with USEC_INFINITY as way to turn off things
838 * journalctl: make sure -f ends when the container indicated by -M terminates
840 * in nss-systemd, if we run inside of RootDirectory= with PrivateUsers= set,
841 find a way to map the User=/Group= of the service to the right name. This way
842 a user/group for a service only has to exist on the host for the right
845 * add bus API for creating unit files in /etc, reusing the code for transient units
847 * add bus API to remove unit files from /etc
849 * add bus API to retrieve current unit file contents (i.e. implement "systemctl cat" on the bus only)
851 * rework fopen_temporary() to make use of open_tmpfile_linkable() (problem: the
852 kernel doesn't support linkat() that replaces existing files, currently)
854 * transient units: don't bother with actually setting unit properties, we
855 reload the unit file anyway
857 * journald: sigbus API via a signal-handler safe function that people may call
858 from the SIGBUS handler
860 * optionally, also require WATCHDOG=1 notifications during service start-up and shutdown
862 * delay activation of logind until somebody logs in, or when /dev/tty0 pulls it
863 in or lingering is on (so that containers don't bother with it until PAM is used). also exit-on-idle
865 * cache sd_event_now() result from before the first iteration...
867 * PID1: find a way how we can reload unit file configuration for
868 specific units only, without reloading the whole of systemd
870 * add an explicit parser for LimitRTPRIO= that verifies
871 the specified range and generates sane error messages for incorrect
874 * when we detect that there are waiting jobs but no running jobs, do something
876 * push CPUAffinity= also into the "cpuset" cgroup controller
878 * PID 1 should send out sd_notify("WATCHDOG=1") messages (for usage in the --user mode, and when run via nspawn)
880 * there's probably something wrong with having user mounts below /sys,
881 as we have for debugfs. for example, src/core/mount.c handles mounts
882 prefixed with /sys generally special.
883 http://lists.freedesktop.org/archives/systemd-devel/2015-June/032962.html
885 * fstab-generator: default to tmpfs-as-root if only usr= is specified on the kernel cmdline
887 * docs: bring http://www.freedesktop.org/wiki/Software/systemd/MyServiceCantGetRealtime up to date
889 * add a job mode that will fail if a transaction would mean stopping
890 running units. Use this in timedated to manage the NTP service
892 http://lists.freedesktop.org/archives/systemd-devel/2015-April/030229.html
894 * The udev blkid built-in should expose a property that reflects
895 whether media was sensed in USB CF/SD card readers. This should then
896 be used to control SYSTEMD_READY=1/0 so that USB card readers aren't
897 picked up by systemd unless they contain a medium. This would mirror
898 the behaviour we already have for CD drives.
900 * hostnamectl: show root image uuid
902 * Find a solution for SMACK capabilities stuff:
903 http://lists.freedesktop.org/archives/systemd-devel/2014-December/026188.html
905 * "systemctl preset-all" should probably order the unit files it
906 operates on lexicographically before starting to work, in order to
907 ensure deterministic behaviour if two unit files conflict (like DMs
910 * synchronize console access with BSD locks:
911 http://lists.freedesktop.org/archives/systemd-devel/2014-October/024582.html
913 * as soon as we have sender timestamps, revisit coalescing multiple parallel daemon reloads:
914 http://lists.freedesktop.org/archives/systemd-devel/2014-December/025862.html
916 * figure out when we can use the coarse timers
918 * add "systemctl start -v foobar.service" that shows logs of a service
919 while the start command runs. This is non-trivial to do without
920 races though, since we should flush out all journal messages before
921 returning from the "systemctl stop".
923 * 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
925 * maybe add support for specifier expansion in user.conf, specifically DefaultEnvironment=
927 * maybe allow timer units with an empty Units= setting, so that they
928 can be used for resuming the system but nothing else.
930 * what to do about udev db binary stability for apps? (raw access is not an option)
932 * systemctl: if some operation fails, show log output?
934 * exponential backoff in timesyncd when we cannot reach a server
936 * timesyncd: add ugly bus calls to set NTP servers per-interface, for usage by NM
938 * merge ~/.local/share and ~/.local/lib into one similar /usr/lib and /usr/share....
940 * add systemd.abort_on_kill or some other such flag to send SIGABRT instead of SIGKILL
941 (throughout the codebase, not only PID1)
945 - service registration
946 - service/domain/types browsing
948 - DNS-SD service registration from socket units
949 - resolved should optionally register additional per-interface LLMNR
950 names, so that for the container case we can establish the same name
951 (maybe "host") for referencing the server, everywhere.
952 - allow clients to request DNSSEC for a single lookup even if DNSSEC is off (?)
953 - hook up resolved with machined-based address resolution
955 * refcounting in sd-resolve is borked
957 * Add a new verb "systemctl top"
959 * add new gpt type for btrfs volumes
961 * generator that automatically discovers btrfs subvolumes, identifies their purpose based on some xattr on them.
963 * a way for container managers to turn off getty starting via $container_headless= or so...
965 * figure out a nice way how we can let the admin know what child/sibling unit causes cgroup membership for a specific unit
967 * For timer units: add some mechanisms so that timer units that trigger immediately on boot do not have the services
968 they run added to the initial transaction and thus confuse Type=idle.
970 * add bus api to query unit file's X fields.
972 * gpt-auto-generator:
973 - Define new partition type for encrypted swap? Support probed LUKS for encrypted swap?
974 - Make /home automount rather than mount?
976 * add generator that pulls in systemd-network from containers when
977 CAP_NET_ADMIN is set, more than the loopback device is defined, even
978 when it is otherwise off
980 * MessageQueueMessageSize= (and suchlike) should use parse_iec_size().
982 * implement Distribute= in socket units to allow running multiple
983 service instances processing the listening socket, and open this up
987 - implement per-slice CPUFairScheduling=1 switch
988 - introduce high-level settings for RT budget, swappiness
989 - how to reset dynamically changed unit cgroup attributes sanely?
990 - when reloading configuration, apply new cgroup configuration
991 - when recursively showing the cgroup hierarchy, optionally also show
992 the hierarchies of child processes
995 - add field to transient units that indicate whether systemd or somebody else saves/restores its settings, for integration with libvirt
997 * when we detect low battery and no AC on boot, show pretty splash and refuse boot
999 * libsystemd-journal, libsystemd-login, libudev: add calls to easily attach these objects to sd-event event loops
1001 * be more careful what we export on the bus as (usec_t) 0 and (usec_t) -1
1003 * 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
1005 * After coming back from hibernation reset hibernation swap partition using the /dev/snapshot ioctl APIs
1007 * If we try to find a unit via a dangling symlink, generate a clean
1008 error. Currently, we just ignore it and read the unit from the search
1011 * refuse boot if /usr/lib/os-release is missing or /etc/machine-id cannot be set up
1013 * man: the documentation of Restart= currently is very misleading and suggests the tools from ExecStartPre= might get restarted.
1015 * load .d/*.conf dropins for device units
1019 - GetAllProperties() on a non-existing object does not result in a failure currently
1020 - port to sd-resolve for connecting to TCP dbus servers
1021 - 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
1022 - see if we can drop more message validation on the sending side
1023 - add API to clone sd_bus_message objects
1024 - longer term: priority inheritance
1025 - dbus spec updates:
1026 - NameLost/NameAcquired obsolete
1029 - update systemd.special(7) to mention that dbus.socket is only about the compatibility socket now
1032 - allow multiple signal handlers per signal?
1033 - document chaining of signal handler for SIGCHLD and child handlers
1034 - define more intervals where we will shift wakeup intervals around in, 1h, 6h, 24h, ...
1035 - maybe support iouring as backend, so that we allow hooking read and write
1036 operations instead of IO ready events into event loops. See considerations
1038 http://blog.vmsplice.net/2020/07/rethinking-event-loop-integration-for.html
1040 * investigate endianness issues of UUID vs. GUID
1042 * dbus: when a unit failed to load (i.e. is in UNIT_ERROR state), we
1043 should be able to safely try another attempt when the bus call LoadUnit() is invoked.
1045 * add a pam module that passes the hdd passphrase into the PAM stack and then expires it, for usage by gdm auto-login.
1047 * add a pam module that on password changes updates any LUKS slot where the password matches
1050 - add unit tests for config_parse_device_allow()
1052 * seems that when we follow symlinks to units we prefer the symlink
1053 destination path over /etc and /usr. We should not do that. Instead
1054 /etc should always override /run+/usr and also any symlink
1057 * when isolating, try to figure out a way how we implicitly can order
1058 all units we stop before the isolating unit...
1060 * teach ConditionKernelCommandLine= globs or regexes (in order to match foobar={no,0,off})
1062 * Add ConditionDirectoryNotEmpty= handle non-absoute paths as a search path or add
1063 ConditionConfigSearchPathNotEmpty= or different syntax? See the discussion starting at
1064 https://github.com/systemd/systemd/pull/15109#issuecomment-607740136.
1066 * BootLoaderSpec: Clarify that the kernel has to be in $BOOT. Clarify
1067 that the boot loader should be installed to the ESP. Define a way
1068 how an installer can figure out whether a BLS compliant boot loader
1071 * think about requeuing jobs when daemon-reload is issued? usecase:
1072 the initrd issues a reload after fstab from the host is accessible
1073 and we might want to requeue the mounts local-fs acquired through
1076 * systemd-inhibit: make taking delay locks useful: support sending SIGINT or SIGTERM on PrepareForSleep()
1078 * remove any syslog support from log.c — we probably cannot do this before split-off udev is gone for good
1080 * shutdown logging: store to EFI var, and store to USB stick?
1082 * merge unit_kill_common() and unit_kill_context()
1085 - honor language efi variables for default language selection (if there are any?)
1086 - honor timezone efi variables for default timezone selection (if there are any?)
1087 - change bootctl to be backed by systemd-bootd to control temporary and persistent default boot goal plus efi variables
1089 - recognize the case when not booted on EFI
1091 * maybe do not install getty@tty1.service symlink in /etc but in /usr?
1093 * print a nicer explanation if people use variable/specifier expansion in ExecStart= for the first word
1095 * mount: turn dependency information from /proc/self/mountinfo into dependency information between systemd units.
1098 - logind: optionally, ignore idle-hint logic for autosuspend, block suspend as long as a session is around
1099 - logind: wakelock/opportunistic suspend support
1100 - Add pretty name for seats in logind
1101 - logind: allow showing logout dialog from system?
1102 - add Suspend() bus calls which take timestamps to fix double suspend issues when somebody hits suspend and closes laptop quickly.
1103 - if pam_systemd is invoked by su from a process that is outside of a
1104 any session we should probably just become a NOP, since that's
1105 usually not a real user session but just some system code that just
1107 - logind: make the Suspend()/Hibernate() bus calls wait for the for
1108 the job to be completed. before returning, so that clients can wait
1109 for "systemctl suspend" to finish to know when the suspending is
1111 - logind: when the power button is pressed short, just popup a
1112 logout dialog. If it is pressed for 1s, do the usual
1113 shutdown. Inspiration are Macs here.
1114 - expose "Locked" property on logind session objects
1115 - maybe allow configuration of the StopTimeout for session scopes
1116 - rename session scope so that it includes the UID. THat way
1117 the session scope can be arranged freely in slices and we don't have
1118 make assumptions about their slice anymore.
1119 - follow PropertiesChanged state more closely, to deal with quick logouts and
1121 - (optionally?) spawn seat-manager@$SEAT.service whenever a seat shows up that as CanGraphical set
1124 - consider introducing implicit _TTY= + _PPID= + _EUID= + _EGID= + _FSUID= + _FSGID= fields
1125 - journald: also get thread ID from client, plus thread name
1126 - journal: when waiting for journal additions in the client always sleep at least 1s or so, in order to minimize wakeups
1127 - add API to close/reopen/get fd for journal client fd in libsystemd-journal.
1128 - fall back to /dev/log based logging in libsystemd-journal, if we cannot log natively?
1129 - declare the local journal protocol stable in the wiki interface chart
1130 - sd-journal: speed up sd_journal_get_data() with transparent hash table in bg
1131 - journald: when dropping msgs due to ratelimit make sure to write
1132 "dropped %u messages" not only when we are about to print the next
1133 message that works, but already after a short timeout
1134 - check if we can make journalctl by default use --follow mode inside of less if called without args?
1135 - maybe add API to send pairs of iovecs via sd_journal_send
1136 - journal: add a setgid "systemd-journal" utility to invoke from libsystemd-journal, which passes fds via STDOUT and does PK access
1137 - journactl: support negative filtering, i.e. FOOBAR!="waldo",
1138 and !FOOBAR for events without FOOBAR.
1139 - journal: store timestamp of journal_file_set_offline() in the header,
1140 so it is possible to display when the file was last synced.
1141 - 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.
1142 - journal: find a way to allow dropping history early, based on priority, other rules
1143 - journal: When used on NFS, check payload hashes
1144 - journald: add kernel cmdline option to disable ratelimiting for debug purposes
1145 - refuse taking lower-case variable names in sd_journal_send() and friends.
1146 - journald: we currently rotate only after MaxUse+MaxFilesize has been reached.
1147 - journal: deal nicely with byte-by-byte copied files, especially regards header
1148 - journal: sanely deal with entries which are larger than the individual file size, but where the components would fit
1149 - Replace utmp, wtmp, btmp, and lastlog completely with journal
1150 - journalctl: instead --after-cursor= maybe have a --cursor=XYZ+1 syntax?
1151 - when a kernel driver logs in a tight loop, we should ratelimit that too.
1152 - journald: optionally, log debug messages to /run but everything else to /var
1153 - journald: when we drop syslog messages because the syslog socket is
1154 full, make sure to write how many messages are lost as first thing
1155 to syslog when it works again.
1156 - journald: allow per-priority and per-service retention times when rotating/vacuuming
1157 - journald: make use of uid-range.h to managed uid ranges to split
1159 - journalctl: add the ability to look for the most recent process of a binary. journalctl /usr/bin/X11 --pid=-1 or so...
1160 - improve journalctl performance by loading journal files
1161 lazily. Encode just enough information in the file name, so that we
1162 do not have to open it to know that it is not interesting for us, for
1163 the most common operations.
1164 - man: document that corrupted journal files is nothing to act on
1165 - rework journald sigbus stuff to use mutex
1166 - Set RLIMIT_NPROC for systemd-journal-xyz, and all other of our
1167 services that run under their own user ids, and use User= (but only
1168 in a world where userns is ubiquitous since otherwise we cannot
1169 invoke those daemons on the host AND in a container anymore). Also,
1170 if LimitNPROC= is used without User= we should warn and refuse
1172 - journalctl --verify: don't show files that are currently being
1173 written to as FAIL, but instead show that their are being written to.
1174 - add journalctl -H that talks via ssh to a remote peer and passes through
1176 - add a version of --merge which also merges /var/log/journal/remote
1177 - journalctl: -m should access container journals directly by enumerating
1178 them via machined, and also watch containers coming and going.
1179 Benefit: nspawn --ephemeral would start working nicely with the journal.
1180 - assign MESSAGE_ID to log messages about failed services
1181 - check if loop in decompress_blob_xz() is necessary
1183 * add a test if all entries in the catalog are properly formatted.
1184 (Adding dashes in a catalog entry currently results in the catalog entry
1185 being silently skipped. journalctl --update-catalog must warn about this,
1186 and we should also have a unit test to check that all our message are OK.)
1189 - document that deps in [Unit] sections ignore Alias= fields in
1190 [Install] units of other units, unless those units are disabled
1191 - man: clarify that time-sync.target is not only sysv compat but also useful otherwise. Same for similar targets
1192 - document that service reload may be implemented as service reexec
1193 - add a man page containing packaging guidelines and recommending usage of things like Documentation=, PrivateTmp=, PrivateNetwork= and ReadOnlyDirectories=/etc /usr.
1194 - document systemd-journal-flush.service properly
1195 - documentation: recommend to connect the timer units of a service to the service via Also= in [Install]
1196 - man: document the very specific env the shutdown drop-in tools live in
1197 - man: add more examples to man pages,
1198 - in particular an example how to do the equivalent of switching runlevels
1199 - man: maybe sort directives in man pages, and take sections from --help and apply them to man too
1200 - document root=gpt-auto properly
1203 - add systemctl switch to dump transaction without executing it
1204 - Add a verbose mode to "systemctl start" and friends that explains what is being done or not done
1205 - "systemctl disable" on a static unit prints no message and does
1206 nothing. "systemctl enable" does nothing, and gives a bad message
1207 about it. Should fix both to print nice actionable messages.
1208 - print nice message from systemctl --failed if there are no entries shown, and hook that into ExecStartPre of rescue.service/emergency.service
1209 - add new command to systemctl: "systemctl system-reexec" which reexecs as many daemons as virtually possible
1210 - systemctl enable: fail if target to alias into does not exist? maybe show how many units are enabled afterwards?
1211 - systemctl: "Journal has been rotated since unit was started." message is misleading
1212 - systemctl status output should include list of triggering units and their status
1215 - "systemctl mask" should find all names by which a unit is accessible
1216 (i.e. by scanning for symlinks to it) and link them all to /dev/null
1219 - timer units should get the ability to trigger when:
1221 - Modulate timer frequency based on battery state
1223 * add libsystemd-password or so to query passwords during boot using the password agent logic
1225 * clean up date formatting and parsing so that all absolute/relative timestamps we format can also be parsed
1227 * on shutdown: move utmp, wall, audit logic all into PID 1 (or logind?), get rid of systemd-update-utmp-runlevel
1229 * make repeated alt-ctrl-del presses printing a dump
1231 * hostnamed: before returning information from /etc/machine-info.conf check the modification data and reread. Similar for localed, ...
1233 * currently x-systemd.timeout is lost in the initrd, since crypttab is copied into dracut, but fstab is not
1236 - emulate /dev/kmsg using CUSE and turn off the syslog syscall
1237 with seccomp. That should provide us with a useful log buffer that
1238 systemd can log to during early boot, and disconnect container logs
1239 from the kernel's logs.
1240 - as soon as networkd has a bus interface, hook up --network-interface=,
1241 --network-bridge= with networkd, to trigger netdev creation should an
1242 interface be missing
1243 - a nice way to boot up without machine id set, so that it is set at boot
1244 automatically for supporting --ephemeral. Maybe hash the host machine id
1245 together with the machine name to generate the machine id for the container
1246 - fix logic always print a final newline on output.
1247 https://github.com/systemd/systemd/pull/272#issuecomment-113153176
1248 - should optionally support receiving WATCHDOG=1 messages from its payload
1250 - optionally automatically add FORWARD rules to iptables whenever nspawn is
1251 running, remove them when shut down.
1254 - add an API so that libvirt-lxc can inform us about network interfaces being
1255 removed or added to an existing machine
1256 - "machinectl migrate" or similar to copy a container from or to a
1257 difference host, via ssh
1258 - introduce systemd-nspawn-ephemeral@.service, and hook it into
1259 "machinectl start" with a new --ephemeral switch
1260 - "machinectl status" should also show internal logs of the container in
1262 - "machinectl history"
1264 - "machinectl commit" that takes a writable snapshot of a tree, invokes a
1265 shell in it, and marks it read-only after use
1268 - cryptsetup-generator: allow specification of passwords in crypttab itself
1269 - support rd.luks.allow-discards= kernel cmdline params in cryptsetup generator
1271 * hw watchdog: optionally try to use the preset watchdog timeout instead of always overriding it
1272 https://bugs.freedesktop.org/show_bug.cgi?id=54712
1274 * add a dependency on standard-conf.xml and other included files to man pages
1276 * MountFlags=shared acts as MountFlags=slave right now.
1278 * properly handle loop back mounts via fstab, especially regards to fsck/passno
1280 * initialize the hostname from the fs label of /, if /etc/hostname does not exist?
1285 - add trigger --subsystem-match=usb/usb_device device
1286 - reimport udev db after MOVE events for devices without dev_t
1288 * There's currently no way to cancel fsck (used to be possible via C-c or c on the console)
1290 * add option to sockets to avoid activation. Instead just drop packets/connections, see http://cyberelk.net/tim/2012/02/15/portreserve-systemd-solution/
1293 - save coredump in Windows/Mozilla minidump format
1294 - when truncating coredumps, also log the full size that the process had, and make a metadata field so we can report truncated coredumps
1296 * support crash reporting operation modes (https://live.gnome.org/GnomeOS/Design/Whiteboards/ProblemReporting)
1298 * be able to specify a forced restart of service A where service B depends on, in case B
1299 needs to be auto-respawned?
1302 - apply "x" on "D" too (see patch from William Douglas)
1303 - instead of ignoring unknown fields, reject them.
1304 - creating new directories/subvolumes/fifos/device nodes
1305 should not follow symlinks. None of the other adjustment or creation
1306 calls follow symlinks.
1309 * make sure systemd-ask-password-wall does not shutdown systemd-ask-password-console too early
1311 * verify that the AF_UNIX sockets of a service in the fs still exist
1312 when we start a service in order to avoid confusion when a user
1313 assumes starting a service is enough to make it accessible
1315 * Make it possible to set the keymap independently from the font on
1316 the kernel cmdline. Right now setting one resets also the other.
1318 * and a dbus call to generate target from current state
1320 * write blog stories about:
1321 - hwdb: what belongs into it, lsusb
1322 - enabling dbus services
1323 - how to make changes to sysctl and sysfs attributes
1325 - how to pass throw-away units to systemd, or dynamically change properties of existing units
1326 - testing with Harald's awesome test kit
1328 - how to develop against journal browsing APIs
1329 - the journal HTTP iface
1330 - non-cgroup resource management
1331 - dynamic resource management with cgroups
1332 - refreshed, longer missions statement
1333 - calendar time events
1334 - init=/bin/sh vs. "emergency" mode, vs. "rescue" mode, vs. "multi-user" mode, vs. "graphical" mode, and the debug shell
1335 - how to create your own target
1336 - instantiated apache, dovecot and so on
1337 - hooking a script into various stages of shutdown/rearly booot
1339 * investigate whether the gnome pty helper should be moved into systemd, to provide cgroup support.
1341 * dot output for --test showing the 'initial transaction'
1344 - When logging about multiple units (stopping BoundTo units, conflicts, etc.),
1345 log both units as UNIT=, so that journalctl -u triggers on both.
1346 - generate better errors when people try to set transient properties
1347 that are not supported...
1348 http://lists.freedesktop.org/archives/systemd-devel/2015-February/028076.html
1349 - maybe introduce WantsMountsFor=? Usecase:
1350 http://lists.freedesktop.org/archives/systemd-devel/2015-January/027729.html
1351 - recreate systemd's D-Bus private socket file on SIGUSR2
1352 - move PAM code into its own binary
1353 - when we automatically restart a service, ensure we restart its rdeps, too.
1354 - hide PAM options in fragment parser when compile time disabled
1355 - Support --test based on current system state
1356 - 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().
1357 - after deserializing sockets in socket.c we should reapply sockopts and things
1358 - drop PID 1 reloading, only do reexecing (difficult: Reload()
1359 currently is properly synchronous, Reexec() is weird, because we
1360 cannot delay the response properly until we are back, so instead of
1361 being properly synchronous we just keep open the fd and close it
1362 when done. That means clients do not get a successful method reply,
1363 but much rather a disconnect on success.
1364 - when breaking cycles drop sysv services first, then services from /run, then from /etc, then from /usr
1365 - when a bus name of a service disappears from the bus make sure to queue further activation requests
1368 - allow port=0 in .socket units
1369 - maybe introduce ExecRestartPre=
1370 - add ReloadSignal= for configuring a reload signal to use
1371 - implement Register= switch in .socket units to enable registration
1372 in Avahi, RPC and other socket registration services.
1373 - allow Type=simple with PIDFile=
1374 https://bugzilla.redhat.com/show_bug.cgi?id=723942
1375 - allow writing multiple conditions in unit files on one line
1376 - introduce Type=pid-file
1377 - introduce mix of BindTo and Requisite
1378 - add a concept of RemainAfterExit= to scope units
1379 - Allow multiple ExecStart= for all Type= settings, so that we can cover rescue.service nicely
1380 - add verification of [Install] section to systemd-analyze verify
1383 - Make sure ID_PATH is always exported and complete for
1384 network devices where possible, so we can safely rely
1388 - add support for more attribute types
1389 - inbuilt piping support (essentially degenerate async)? see loopback-setup.c and other places
1392 - add more keys to [Route] and [Address] sections
1393 - add support for more DHCPv4 options (and, longer term, other kinds of dynamic config)
1394 - add reduced [Link] support to .network files
1395 - properly handle routerless dhcp leases
1396 - work with non-Ethernet devices
1397 - dhcp: do we allow configuring dhcp routes on interfaces that are not the one we got the dhcp info from?
1398 - the DHCP lease data (such as NTP/DNS) is still made available when
1399 a carrier is lost on a link. It should be removed instantly.
1400 - expose in the API the following bits:
1401 - option 15, domain name
1402 - option 12, hostname and/or option 81, fqdn
1403 - option 123, 144, geolocation
1404 - option 252, configure http proxy (PAC/wpad)
1405 - provide a way to define a per-network interface default metric value
1406 for all routes to it. possibly a second default for DHCP routes.
1407 - allow Name= to be specified repeatedly in the [Match] section. Maybe also
1408 support Name=foo*|bar*|baz ?
1409 - whenever uplink info changes, make DHCP server send out FORCERENEW
1411 * Figure out how to do unittests of networkd's state serialization
1414 - figure out how much we can increase Maximum Message Size
1417 - add functions to set previously stored IPv6 addresses on startup and get
1418 them at shutdown; store them in client->ia_na
1419 - write more test cases
1420 - implement reconfigure support, see 5.3., 15.11. and 22.20.
1421 - implement support for temporary adressess (IA_TA)
1422 - implement dhcpv6 authentication
1423 - investigate the usefulness of Confirm messages; i.e. are there any
1424 situations where the link changes without any loss in carrier detection
1426 - some servers don't do rapid commit without a filled in IA_NA, verify
1433 - natively watch for dbus-*.service symlinks (PENDING)
1434 - teach dbus to activate all services it finds in /etc/systemd/services/org-*.service
1436 * make cryptsetup lower --iter-time
1438 * kernel: add device_type = "fb", "fbcon" to class "graphics"
1440 * /usr/bin/service should actually show the new command line
1442 * 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=
1444 * neither pkexec nor sudo initialize environ[] from the PAM environment?
1446 * fedora: update policy to declare access mode and ownership of unit files to root:root 0644, and add an rpmlint check for it
1448 * register catalog database signature as file magic
1450 * zsh shell completion:
1451 - <command> <verb> -<TAB> should complete options, but currently does not
1452 - systemctl add-wants,add-requires
1454 * systemctl status should know about 'systemd-analyze calendar ... --iterations='
1455 * If timer has just OnInactiveSec=..., it should fire after a specified time
1456 after being started.
1460 * look for close() vs. close_nointr() vs. close_nointr_nofail()
1462 * check for strerror(r) instead of strerror(-r)
1466 * set_put(), hashmap_put() return values check. i.e. == 0 does not free()!
1468 * use secure_getenv() instead of getenv() where appropriate
1470 * link up selected blog stories from man pages and unit files Documentation= fields
1471 String is not UTF-8 clean, ignoring assignment