]>
Commit | Line | Data |
---|---|---|
afe3ab58 | 1 | Bugfixes: |
c343be28 | 2 | |
54fcb619 ZJS |
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 | ||
5b326dee ZJS |
7 | * Jun 01 09:43:02 krowka systemd[1]: Unit user@1000.service has alias user@.service. |
8 | Jun 01 09:43:02 krowka systemd[1]: Unit user@6.service has alias user@.service. | |
9 | Jun 01 09:43:02 krowka systemd[1]: Unit user-runtime-dir@6.service has alias user-runtime-dir@.service. | |
10 | ||
f38afcd0 | 11 | External: |
f85857df | 12 | |
f38afcd0 | 13 | * Fedora: add an rpmlint check that verifies that all unit files in the RPM are listed in %systemd_post macros. |
bafb15ba | 14 | |
07eabc2b LB |
15 | * dbus: |
16 | - natively watch for dbus-*.service symlinks (PENDING) | |
17 | - teach dbus to activate all services it finds in /etc/systemd/services/org-*.service | |
18 | ||
19 | * kernel: add device_type = "fb", "fbcon" to class "graphics" | |
20 | ||
21 | * /usr/bin/service should actually show the new command line | |
22 | ||
23 | * 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= | |
24 | ||
25 | * neither pkexec nor sudo initialize environ[] from the PAM environment? | |
26 | ||
27 | * fedora: update policy to declare access mode and ownership of unit files to root:root 0644, and add an rpmlint check for it | |
28 | ||
29 | * register catalog database signature as file magic | |
30 | ||
31 | * zsh shell completion: | |
32 | - <command> <verb> -<TAB> should complete options, but currently does not | |
33 | - systemctl add-wants,add-requires | |
76c068b7 | 34 | - systemctl reboot --boot-loader-entry= |
07eabc2b LB |
35 | |
36 | * systemctl status should know about 'systemd-analyze calendar ... --iterations=' | |
37 | * If timer has just OnInactiveSec=..., it should fire after a specified time | |
38 | after being started. | |
39 | ||
40 | * write blog stories about: | |
41 | - hwdb: what belongs into it, lsusb | |
42 | - enabling dbus services | |
43 | - how to make changes to sysctl and sysfs attributes | |
44 | - remote access | |
45 | - how to pass throw-away units to systemd, or dynamically change properties of existing units | |
46 | - testing with Harald's awesome test kit | |
47 | - auto-restart | |
48 | - how to develop against journal browsing APIs | |
49 | - the journal HTTP iface | |
50 | - non-cgroup resource management | |
51 | - dynamic resource management with cgroups | |
52 | - refreshed, longer missions statement | |
53 | - calendar time events | |
54 | - init=/bin/sh vs. "emergency" mode, vs. "rescue" mode, vs. "multi-user" mode, vs. "graphical" mode, and the debug shell | |
55 | - how to create your own target | |
56 | - instantiated apache, dovecot and so on | |
57 | - hooking a script into various stages of shutdown/rearly booot | |
58 | ||
59 | Regularly: | |
60 | ||
61 | * look for close() vs. close_nointr() vs. close_nointr_nofail() | |
62 | ||
63 | * check for strerror(r) instead of strerror(-r) | |
64 | ||
65 | * pahole | |
66 | ||
67 | * set_put(), hashmap_put() return values check. i.e. == 0 does not free()! | |
68 | ||
69 | * use secure_getenv() instead of getenv() where appropriate | |
70 | ||
71 | * link up selected blog stories from man pages and unit files Documentation= fields | |
72 | ||
5e524b40 LP |
73 | Janitorial Clean-ups: |
74 | ||
5e524b40 LP |
75 | * Rearrange tests so that the various test-xyz.c match a specific src/basic/xyz.c again |
76 | ||
bb527e11 LP |
77 | * rework mount.c and swap.c to follow proper state enumeration/deserialization |
78 | semantics, like we do for device.c now | |
79 | ||
c6407108 LP |
80 | Features: |
81 | ||
2df264e6 LP |
82 | * we should probably drop all use of prefix_roota() and friends, and use |
83 | chase_symlinks() instead | |
84 | ||
d486b26f LP |
85 | * make persistent restarts easier by adding a new setting OpenPersistentFile= |
86 | or so, which allows openeing one or more files that is "persistent" across | |
87 | service restarts, hot reboot, cold reboots (depending on configuration): the | |
88 | files are created empty on first invocation, and on subsequent invocations | |
89 | the files are reboot. The files would be backed by tmpfs, pmem or /var | |
90 | depending on desired level of persistency. | |
91 | ||
92 | * sd-event: add ability to "chain" event sources. Specifically, add a call | |
93 | sd_event_source_chain(x, y), which will automatically enable event source y | |
94 | in oneshit mode once x is triggered. Use case: in src/core/mount.c implement | |
95 | the /proc/self/mountinfo rescan on SIGCHLD with this: whenever a SIGCHLD is | |
96 | seen, trigger the rescan defer event source automatically, and allow it to be | |
97 | dispatched *before* the SIGCHLD is handled (based on priorities). Benefit: | |
98 | dispatch order is strictly controlled by priorities again. (next step: chain | |
99 | event sources to the ratelimit being over) | |
100 | ||
101 | * if we fork of a service with StandardOutput=journal, and it forks off a | |
102 | subprocess that quickly dies, we might not be able to identify the cgroup it | |
103 | comes from, but we can still derive that from the stdin socket its output | |
104 | came from. We apparently don't do that right now. | |
105 | ||
106 | * make systemd-fstab-generator look for a system credential encoding root= or | |
107 | usr= | |
108 | ||
109 | * add ability to set hostname with suffix derived from machine id at boot | |
110 | ||
111 | * ask dracut to generate usr= on the kernel cmdline so that we don't need to | |
112 | read /etc/fstab from the root fs from the initrd and do daemon-reload | |
113 | ||
114 | * document that process resource limits are bullshit | |
115 | ||
116 | * add PR_SET_DUMPABLE service setting | |
117 | ||
027301b4 LP |
118 | * homed/userdb: maybe define a "companion" dir for home directories where apps |
119 | can safely put privileged stuff in. Would not be writable by the user, but | |
120 | still conceptually belong to the user. Would be included in user's quota if | |
121 | possible, even if files are not owned by UID of user. Usecase: container | |
122 | images that owned by arbitrary UIDs, and are owned/managed by the users, but | |
123 | are not directly belonging to the user's UID. Goal: we shouldn't place more | |
124 | privileged dirs inside of unprivileged dirs, and thus containers really | |
125 | should not be placed inside of traditional UNIX home dirs (which are owned by | |
3881fd40 | 126 | users themselves) but somewhere else, that is separate, but still close |
027301b4 LP |
127 | by. Inform user code about path to this companion dir via env var, so that |
128 | container managers find it. the ~/.identity file is also a candidate for a | |
129 | file to move there, since it is managed by privileged code (i.e. homed) and | |
130 | not unprivileged code. | |
131 | ||
132 | * given that /etc/ssh/ssh_config.d/ is a thing now, ship a drop-in for that | |
133 | that hooks up userbdctl ssh-key stuff. | |
134 | ||
135 | * allow embedding a signature blob for PCR hashes into separate section in | |
136 | unified kernel binaries. This section should be picked up by sd-stub, and | |
137 | passed in a file to the booted kernel (via initrd cpio, as usual). Usecase: | |
138 | this way we can implement disk encryption policies that bind to specific | |
139 | kernel PCR state, without breaking things on every kernel update. As long as | |
140 | the kernel includes the PCR signature blob we should be good, as disk | |
141 | encryption can then pass the signature to the TPM to unlock their secrets. | |
142 | Why do this via a separate PE section? That's because the PCR state depends | |
143 | on the measured kernel/initrd of course, thus we cannot put the signature | |
144 | into the kernel/initrd itself, because that would require a time machine. | |
145 | Hence we have to find a separate place. A simple solution is a PE section | |
146 | of its own, because then it is next to the kernel and initrd which after all | |
147 | are stored in PE sections of their own too. Building a unified kernel would | |
148 | thus mean, calculating PCR values for the raw kernel image, and raw initrd | |
149 | image, then signing those PCR values with a vendor key, and then combining | |
150 | sd-stub, raw kernel image, raw initrd, and PCR signature into a unified | |
151 | kernel image. | |
152 | ||
153 | * a new tool "systemd-trust" or so, that can calculate PCR hashes offline, and | |
154 | optionally sign them. for that we should extend our syntax for specifying pcr | |
155 | policies (e.g. the string like "4+7+9") so that it can also include explicit | |
156 | hash values, i.e. | |
157 | 4=sha256:0ef149998289474e4bb31813edda6ad7f3c991b2d8dec6e8fe4db7a1f039f2d1+7=sha256:87428fc522803d31065e7bce3cf03fe475096631e5e07bbd7a0fde60c4cf25c7+9=sha256:0263829989b6fd954f72baaf2fc64bc2e2f01d692d4de72986ea808f6e99813f | |
158 | and file names to calculate hashes from, i.e. | |
159 | 4=file:/boot/vmlinuz+7=file:/boot/initrd/+9=file:/etc/fstab" | |
160 | The systemd-trust tool should then be able to resolve any "underspecifed" | |
161 | form into the form with explicit hash values. | |
162 | ||
3a466def LP |
163 | * maybe add support for binding and connecting AF_UNIX sockets in the file |
164 | system outside of the 108ch limit. When connecting, open O_PATH fd to socket | |
165 | inode first, then connect to /proc/self/fd/XYZ. When binding, create symlink | |
166 | to target dir in /tmp, and bind through it. | |
167 | ||
a8796773 LP |
168 | * tmpfiles: for f/F/w lines, if the argument columns is left unspecified, look |
169 | for a service credential named after the file path to write to, and load | |
170 | contents to write from there. Usecase: provision arbitrary files from | |
171 | credentials. Example use: with a line like "f /root/.ssh/authorized-keys | |
172 | 0644 root root" in a tmpfiles.d/ snippet add | |
173 | LoadCredential=root.ssh.authorized-keys via drop-in to | |
174 | systemd-tmpfiles.service, and then provision an SSH access key through | |
175 | nspawn's --load-credential=, through qemu's fw_cfg, or via systemd-stub's | |
176 | credntial pick-up. The latter is particularly interesting to implement SSH | |
177 | access to an initrd. | |
178 | ||
179 | * systemd-homed: when initializing, look for a credential sysemd.homed.register | |
180 | or so with JSON user records to automatically register if not registered yet. | |
181 | Usecase: deploy a system, and add an account one can directly log into. | |
182 | ||
fd74ed23 LP |
183 | * add a proper concept of a "developer" mode, i.e. where cryptographic |
184 | protections of the root OS are weakened after interactive confirmation, to | |
185 | allow hackers to allow their own stuff. idea: allow entering developer mode | |
186 | only via explicit choice in boot menu: i.e. add explicit boot menu item for | |
187 | it. when developer mode is entered generate a key pair in the TPM2, and add | |
188 | the public part of it automatically to keychain of valid code signature keys | |
189 | on subsequent boots. Then provide a tool to sign code with the key in the | |
190 | TPM2. Ensure that boot menu item is only way to enter developer mode, by | |
191 | binding it to locality/PCRs so that that keys cannot be generated otherwise. | |
192 | ||
193 | * services: add support for cryptographically unlocking per-service directories | |
194 | via TPM2. Specifically, for StateDirectory= (and related dirs) use fscrypt to | |
195 | set up the directory so that it can only be accessed if host and app are in | |
196 | order. | |
197 | ||
98045d12 LP |
198 | * TPM2: add auth policy for signed PCR values to make updates easy. i.e. do |
199 | what tpm2_policyauthorize tool does. To be truly useful scheme needs to be a | |
200 | bit more elaborate though: policy probably must take some nvram based | |
201 | generation counter into account that can only monotonically increase and can | |
202 | be used to invalidate old PCR signatures. Otherwise people could downgrade to | |
203 | old signed PCR sets whenever they want. Usecase: encrypt the rootfs with LUKS | |
204 | with a key that can only be unlocked via a pristine pre-built Fedora | |
205 | kernel+initrd. | |
206 | ||
207 | * update HACKING.md to suggest developing systemd with the ideas from: | |
208 | https://0pointer.net/blog/testing-my-system-code-in-usr-without-modifying-usr.html | |
209 | https://0pointer.net/blog/running-an-container-off-the-host-usr.html | |
210 | ||
211 | * add a clear concept how the initrd can make up credentials on their own to | |
212 | pass to the system when transitioning into the host OS. usecase: things like | |
213 | cloud-init/ignitation and similar can parameterize the host with data they | |
214 | acquire. | |
215 | ||
0b3a64fe LP |
216 | * Add ConditionCredentialExists= or so, that allows conditionalizing services |
217 | depending on whether a specific system credential is set. Usecase: a service | |
218 | similar to the ssh keygen service that installs any SSH host key supplied via | |
219 | system credentials into /etc/ssh. | |
220 | ||
55fc776b LP |
221 | * drop support for kernels that lack ambient capabilities support (i.e. make |
222 | 4.3 new baseline). Then drop support for "!!" modifier for ExecStart= which | |
223 | is only supported for such old kernels | |
224 | ||
8b825133 LP |
225 | * sd-event: compat wd reuse in inotify code: keep a set of removed watch |
226 | descriptors, and clear this set piecemeal when we see the IN_IGNORED event | |
227 | for it, or when read() returns EAGAIN or on IN_Q_OVERFLOW. Then, whenever we | |
228 | see an inotify wd event check against this set, and if it is contained ignore | |
8ac6b05b | 229 | the event. (to be fully correct this would have to count the occurrences, in |
8b825133 LP |
230 | case the same wd is reused multiple times before we start processing |
231 | IN_IGNORED again) | |
232 | ||
9c9c87ac LP |
233 | * sd-stub: set efi var indicating stub features, i.e. whether they pick up |
234 | creds, sysexts and so on. similar to existing variable of sd-boot | |
235 | ||
236 | * sd-stub: set efi vars declaring TPM PCRs we measured creds/cmdline + sysext | |
237 | into (even if we hardcode them) | |
238 | ||
5b06ad51 LP |
239 | * systemd-fstab-generator: support addition mount specifications via kernel |
240 | cmdline. Usecase: invoke a VM, and mount a host homedir into it via | |
241 | virtio-fs. | |
242 | ||
243 | * for vendor-built signed initrds: | |
62471289 | 244 | - make sysext run in the initrd |
5b06ad51 | 245 | - sysext should pick up sysext images from /.extra/ in the initrd, and insist |
62471289 | 246 | on verification if in secureboot mode |
5b06ad51 LP |
247 | - kernel-install should be able to install pre-built unified kernel images in |
248 | type #2 drop-in dir in the ESP. | |
62471289 LP |
249 | - kernel-install should be able install encrypted creds automatically for |
250 | machine id, root pw, rootfs uuid, resume partition uuid, and place next to | |
251 | EFI kernel, for sd-stub to pick them up. These creds should be locked to | |
252 | the TPM, and bind to the right PCR the kernel is measured to. | |
11b957b5 LP |
253 | - kernel-install should be able to pick up initrd sysexts automatically and |
254 | place them next to EFI kernel, for sd-stub to pick them up. | |
5b06ad51 LP |
255 | - systemd-fstab-generator should look for rootfs device to mount in creds |
256 | - pid 1 should look for machine ID in creds | |
62471289 | 257 | - systemd-resume-generator should look for resume partition uuid in creds |
11b957b5 LP |
258 | - sd-stub: automatically pick up microcode from ESP (/loader/microcode/*) |
259 | and synthesize initrd from it, and measure it. Signing is not necessary, as | |
260 | microcode does that on its own. Pass as first initrd to kernel. | |
e4b5226d LP |
261 | - sd-stub should measure the kernel/initrd/… into a separate PCR, so that we |
262 | have one PCR we can bind the encrypted creds to that is not effected by | |
263 | anything else but what we drop in via kernel-install, i.e. by earlier EFI | |
264 | code running (i.e. like PCR 4) | |
5b06ad51 | 265 | |
2df2bb1f LP |
266 | * Add a new service type very similar to Type=notify, that goes one step |
267 | further and extends the protocol to cover reloads. Specifically, SIGHUP will | |
268 | become the official way to reload, and daemon has to respond with sd_notify() | |
269 | to report when it starts reloading, and when it is complete reloading. Care | |
270 | must be taken to remove races from this model. I.e. PID 1 needs to take | |
271 | CLOCK_MONOTONIC, then send SIGHUP, then wait for at least one RELOADING=1 | |
272 | message that comes with a newer timestamp, then wait for a READY=1 message. | |
273 | while we are at it, also maybe extend the logic to require handling of some | |
274 | specific SIGRT signal for setting debug log level, that carries the level via | |
275 | the sigqueue() data parameter. With that we extended with minimal logic the | |
276 | service runtime logic quite substantially. | |
277 | ||
5645b497 LP |
278 | * firstboot: maybe just default to C.UTF-8 locale if nothing is set, so that we |
279 | don't query this unnecessarily in entirely uninitialized | |
280 | containers. (i.e. containers with empty /etc). | |
281 | ||
598e4315 | 282 | * beef up sd_notify() to support AV_VSOCK in $NOTIFY_SOCKET, so that VM |
197be532 LP |
283 | managers can get ready notifications from VMs, just like container managers |
284 | from their payload. Also pick up address from qemu/fw_cfg if set there. | |
285 | (which has benefits, given SecureBoot and kernel cmdline are not necessarily | |
286 | friends.) | |
598e4315 | 287 | |
c0da575a LP |
288 | * mirroring this: maybe support binding to AV_VSOCK in Type=notify services, |
289 | then passing $NOTIFY_SOCKET and $NOTIFY_GUESTCID with PID1's cid (typically | |
290 | fixed to "2", i.e. the official host cid) and the expected guest cid, for the | |
3a258d3a | 291 | two sides of the channel. The latter env var could then be used in an |
c0da575a LP |
292 | appropriate qemu cmdline. That way qemu payloads could talk sd_notify() |
293 | directly to host service manager. | |
294 | ||
598e4315 LP |
295 | * maybe write a tool that binds an AF_VFSOCK socket, then invokes qemu, |
296 | extending the command line to enable vsock on the VM, and using fw_cfg to | |
297 | configure socket address. | |
298 | ||
3e3c49cb LP |
299 | * sd-boot: rework random seed handling following recent kernel changes: always |
300 | pass seed to kernel, but credit only if secure boot is used | |
301 | ||
127927b2 LP |
302 | * sd-boot: also include the hyperv "vm generation id" in the random seed hash, |
303 | to cover nicely for machine clones. It's found in the ACPI tables, which | |
304 | should be easily accessible from UEFI. | |
305 | ||
3e3c49cb LP |
306 | * sd-boot: add menu item for shutdown? or hotkey? |
307 | ||
2cf120f7 LP |
308 | * sd-device has an API to create an sd_device object from a device id, but has |
309 | no api to query the device id | |
310 | ||
311 | * sd-device should return the devnum type (i.e. 'b' or 'c') via some API for an | |
312 | sd_device object, so that data passed into sd_device_new_from_devnum() can | |
313 | also be queried. | |
314 | ||
5b89bff5 LP |
315 | * sd-event: optionally, if per-event source rate limit is hit, downgrade |
316 | priority, but leave enabled, and once ratelimit window is over, upgrade | |
317 | priority again. That way we can combat event source starvation without | |
318 | stopping processing events from one source entirely. | |
319 | ||
e4f92a62 LP |
320 | * sd-event: similar to existing inotify support add fanotify support (given |
321 | that apparently new features in this area are only going to be added to the | |
322 | latter). | |
5b89bff5 LP |
323 | |
324 | * sd-event: add 1st class event source for clock changes | |
325 | ||
326 | * sd-event: add 1st class event source for timezone changes | |
327 | ||
9e83d3e4 LP |
328 | * support uefi/http boots with sd-boot: instead of looking for dropin files in |
329 | /loader/entries/ dir, look for a file /loader/entries/SHA256SUMS and use that | |
330 | as directory manifest. The file would be a standard directory listing as | |
331 | generated by GNU sha256sums. | |
332 | ||
47a9f917 LP |
333 | * sd-boot: maybe add support for embedding the various auxiliary resources we |
334 | look for right in the sd-boot binary. i.e. take inspiration from sd-stub | |
335 | logic: allow combining sd-boot via objcopy with kernels to enumerate, .conf | |
336 | files, drivers, keys to enroll and so on. Then, add whatever we find that way | |
337 | to the menu. Usecase: allow building a single PE image you can boot into via | |
338 | UEFI HTTP boot. | |
339 | ||
340 | * maybe add a new UEFI stub binary "sd-http". It works similar to sd-stub, but | |
341 | all it does is download a file from a http server, and execute it, after | |
342 | optionally checking its hash sum. idea would be: combine this "sd-http" stub | |
343 | binary with some minimal info about an URL + hash sum, plus .osrel data, and | |
344 | drop it into the unified kernel dir in the ESP. And bam you have something | |
345 | that is tiny, feels a lot like a unified kernel, but all it does is chainload | |
346 | the real kernel. benefit: downloading these stubs would be tiny and quick, | |
347 | hence cheap for enumeration. | |
348 | ||
d360eafb LP |
349 | * initialize machine ID from systemd credential picked up from the ESP via |
350 | sd-stub, so that machine ID is stable even on systems where unified kernels | |
351 | are used, and hence kernel cmdline cannot be modified locally | |
352 | ||
353 | * in gpt-auto-generator: check partition uuids against such uuids supplied via | |
354 | sd-stub credentials. That way, we can support parallel OS installations with | |
355 | pre-built kernels. | |
356 | ||
357 | * sysext: measure all activated sysext into a TPM PCR | |
358 | ||
359 | * maybe add a "syscfg" concept, that is almost entirely identical to "sysext", | |
360 | but operates on /etc/ instead of /usr/ and /opt/. Use case would be: trusted, | |
361 | authenticated, atomic, additive configuration management primitive: drop in a | |
362 | configuration bundle, and activate it, so that it is instantly visible, | |
363 | comprehensively. | |
364 | ||
365 | * systemd-dissect: show available versions inside of a disk image, i.e. if | |
366 | multiple versions are around of the same resource, show which ones. (in other | |
367 | words: show partition labels). | |
368 | ||
369 | * systemd-nspawn: make boot assessment do something sensible in a | |
370 | container. i.e send an sd_notify() from payload to container manager once | |
371 | boot-up is completed successfully, and use that in nspawn for dealing with | |
372 | boot counting, implemented in the partition table labels and directory names. | |
373 | ||
374 | * maybe add a generator that reads /proc/cmdline, looks for | |
375 | systemd.pull-raw-portable=, systemd-pull-raw-sysext= and similar switches | |
376 | that take an URL as parameter. It then generates service units for | |
b17a681b | 377 | systemd-pull calls that download these URLs if not installed yet. usecase: |
d360eafb LP |
378 | invoke a VM or nspawn container in a way it automatically deploys/runs these |
379 | images as OS payloads. i.e. have a generic OS image you can point to any | |
380 | payload you like, which is then downloaded, securely verified and run. | |
381 | ||
dca92ca3 LP |
382 | * improve scope units to support creation by pidfd instead of by PID |
383 | ||
e4f92a62 | 384 | * deprecate cgroupsv1 further (print log message at boot) |
f1a147f2 LP |
385 | |
386 | * systemd-dissect: add --cat switch for dumping files such as /etc/os-release | |
387 | ||
f5d0f21c LP |
388 | * per-service sandboxing option: ProtectIds=. If used, will overmount |
389 | /etc/machine-id and /proc/sys/kernel/random/boot_id with synthetic files, to | |
390 | make it harder for the service to identify the host. Depending on the user | |
391 | setting it should be fully randomized at invocation time, or a hash of the | |
392 | real thing, keyed by the unit name or so. Of course, there are other ways to | |
393 | get these IDs (e.g. journal) or similar ids (e.g. MAC addresses, DMI ids, CPU | |
394 | ids), so this knob would only be useful in combination with other lockdown | |
395 | options. Particularly useful for portable services, and anything else that | |
396 | uses RootDirectory= or RootImage=. (Might also over-mount | |
397 | /sys/class/dmi/id/*{uuid,serial} with /dev/null). | |
398 | ||
63a185dc LP |
399 | * journalctl/timesyncd: whenever timesyncd acquires a synchronization from NTP, |
400 | create a structured log entry that contains boot ID, monotonic clock and | |
401 | realtime clock (I mean, this requires no special work, as these three fields | |
402 | are implicit). Then in journalctl when attempting to display the realtime | |
403 | timestamp of a log entry, first search for the closest later log entry | |
404 | of this kinda that has a matching boot id, and convert the monotonic clock | |
405 | timestamp of the entry to the realtime clock using this info. This way we can | |
406 | retroactively correct the wallclock timestamps, in particular for systems | |
407 | without RTC, i.e. where initially wallclock timestamps carry rubbish, until | |
408 | an NTP sync is acquired. | |
409 | ||
45cab6e3 LP |
410 | * kernel-install: |
411 | - add --all switch for rerunning kernel-install for all installed kernels | |
412 | - maybe add env var that shortcuts kernel-install for installers that want to | |
413 | call it at the end only | |
414 | ||
a57d72ce LP |
415 | * doc: prep a document explaining resolved's internal objects, i.e. Query |
416 | vs. Question vs. Transaction vs. Stream and so on. | |
417 | ||
a5a316e7 LP |
418 | * doc: prep a document explaining PID 1's internal logic, i.e. transactions, |
419 | jobs, units | |
420 | ||
4cddc18d LP |
421 | * bootspec: remove tries counter from boot entry ids |
422 | ||
a57d72ce LP |
423 | * bootspec: bring UEFI and userspace enumeration of bootspec entries back into |
424 | sync, i.e. parse out tries in both | |
425 | ||
594f64f4 LP |
426 | * automatically ignore threaded cgroups in cg_xyz(). |
427 | ||
e6f48be8 LP |
428 | * add linker script that implicitly adds symbol for build ID and new coredump |
429 | json package metadata, and use that when logging | |
430 | ||
f08e143c LP |
431 | * systemd-dissect: show GPT disk UUID in output |
432 | ||
433 | * Enable RestricFileSystems= for all our long-running services (similar: | |
434 | RestrictNetworkInterfaces=) | |
435 | ||
436 | * Add systemd-analyze security checks for RestrictFileSystems= and | |
437 | RestrictNetworkInterfaces= | |
438 | ||
1ccfb792 LP |
439 | * cryptsetup/homed: implement TOTP authentication backed by TPM2 and its |
440 | internal clock. | |
441 | ||
1ccfb792 | 442 | * nspawn: optionally set up nftables/iptables routes that forward UDP/TCP |
f08e143c | 443 | traffic on port 53 to resolved stub 127.0.0.54 |
e0c311b1 | 444 | |
f5ba8115 LP |
445 | * man: rework os-release(5), and clearly separate our extension-release.d/ and |
446 | initrd-release parts, i.e. list explicitly which fields are about what. | |
447 | ||
448 | * sysext: before applying a sysext, do a superficial validation run so that | |
449 | things are not rearranged to wildy. I.e. protect against accidental fuckups, | |
450 | such as masking out /usr/lib/ or so. We should probably refuse if existing | |
451 | inodes are replaced by other types of inodes or so. | |
452 | ||
453 | * sysext: ensure one can build a sysext that can safely apply to *any* system | |
454 | (because it contains only static go binaries in /opt/ or so) | |
455 | ||
af11e0ef LP |
456 | * userdb: when synthesizing NSS records, pick "best" password from defined |
457 | passwords, not just the first. i.e. if there are multiple defined, prefer | |
458 | unlocked over locked and prefer non-empty over empty. | |
459 | ||
460 | * maybe add a tool inspired by the GPT auto discovery spec that runs in the | |
461 | initrd and rearranges the rootfs hierarchy via bind mounts, if | |
462 | enabled. Specifically in some top-level dir /@auto/ it will look for | |
463 | dirs/symlinks/subvolumes that are named after their purpose, and optionally | |
464 | encode a version as well as assessment counters, and then mount them into the | |
465 | file system tree to boot into, similar to how we do that for the gpt auto | |
466 | logic. Maybe then bind mount the original root into /.superior or something | |
467 | like that (so that update tools can look there). Further discussion in this | |
468 | thread: | |
469 | https://lists.freedesktop.org/archives/systemd-devel/2021-November/047059.html | |
470 | The GPT dissection logic should automatically enable this tool whenever we | |
471 | detect a specially marked root fs (i.e introduce a new generic root gpt type | |
472 | for this, that is arch independent). The also implement this in the image | |
473 | dissection logic, so that nspawn/RootImage= and so on grok it. Maybe make | |
474 | generic enough so that it can also work for ostrees arrangements. | |
475 | ||
476 | * if a path ending in ".auto.d/" is set for RootDirectory=/RootImage= then do a | |
477 | strverscmp() of everything inside that dir and use that. i.e. implement very | |
478 | simple version control. Also use this in systemd-nspawn --image= and so on. | |
479 | ||
480 | * homed: while a home dir is not activated generate slightly different NSS | |
481 | records for it, that reports the home dir as "/" and the shell as some binary | |
482 | provided by us. Then, when an SSH login happens and SSH permits it our binary | |
483 | is invoked. This binary can then talk to homed and activate the homedir if | |
484 | it's not around yet, prompting the user for a password. Once that succeeded | |
485 | we'll switch to the real user record, i.e. home dir and shell, and our tool | |
486 | exec()s the latter. Net effect: ssh'ing into a homed account will just work: | |
487 | we'll neatly prompt for the homedir's password if its needed. –– Building on | |
488 | this we could take this even further: since this tool will potentially have | |
489 | access to the client's ssh-agent (if ssh-agent forwarding is enabled) we | |
490 | could implement SSH unlocking of a homedir with that: when enrolling a new | |
491 | ssh pubkey in a user record we'd ask the ssh-agent to sign some random value | |
492 | with the privkey, then use that as luks key to unlock the home dir. Will not | |
493 | work for ECDSA keys since their signatures contain a random component, but | |
494 | will work for RSA and Ed25519 keys. | |
495 | ||
bb5464ad LP |
496 | * add tiny service that decrypts encrypted user records passed via initrd |
497 | credential logic and drops them into /run where nss-systemd can pick them up, | |
498 | similar to /run/host/userdb/. Usecase: drop a root user JSON record there, | |
499 | and use it in the initrd to log in as root with locally selected password, | |
11b957b5 LP |
500 | for debugging purposes. Other usecase: boot into qemu with regular user |
501 | mounted from host. maybe put this in systemd-user-sessions.service? | |
bb5464ad LP |
502 | |
503 | * drop dependency on libcap, replace by direct syscalls based on | |
504 | CapabilityQuintet we already have. (This likely allows us drop drop libcap | |
505 | dep in the base OS image) | |
506 | ||
507 | * sysext: automatically activate sysext images dropped in via new sd-stub | |
508 | sysext pickup logic. | |
509 | ||
510 | * add concept for "exitrd" as inverse of "initrd", that we can transition to at | |
511 | shutdown, and has similar security semantics. This should then take the place | |
512 | of dracut's shutdown logic. Should probably support sysexts too. Care needs | |
513 | to be taken that the resulting logic ends up in RAM, i.e. is copied out of | |
514 | on-disk storage. | |
515 | ||
a07ab1dd LP |
516 | * userdbd: implement an additional varlink service socket that provides the |
517 | host user db in restricted form, then allow this to be bind mounted into | |
518 | sandboxed environments that want the host database in minimal form. All | |
519 | records would be stripped of all meta info, except the basic UID/name | |
520 | info. Then use this in portabled environments that do not use PrivateUsers=1. | |
521 | ||
a5bf435e LP |
522 | * logind introduce two types of sessions: "heavy" and "light". The former would |
523 | be our current sessions. But the latter would be a new type of session that | |
524 | is mostly the same but does not pull in user@.service or wait for it. Then, | |
525 | allow configuration which type of session is desired via pam_systemd | |
526 | parameters, and then make user@.service's session one of these "light" ones. | |
527 | People could then choose to make FTP sessions and suchlike "light" if they | |
528 | don't want the service manager to be started for that. | |
529 | ||
da3ab57c LP |
530 | * /etc/veritytab: allow that the roothash column can be specified as fs path |
531 | including a path to an AF_UNIX path, similar to how we do things with the | |
532 | keys of /etc/crypttab. That way people can store/provide the roothash | |
533 | externally and provide to us on demand only. | |
534 | ||
636c8a1f LP |
535 | * add high-level lockdown level for GPT dissection logic: e.g. an enum that can |
536 | be ANY (to mount anything), TRUSTED (to require that /usr is on signed | |
537 | verity, but rest doesn't matter), LOCKEDDOWN (to require that everything is | |
538 | on signed verity, except for ESP), SUPERLOCKDOWN (like LOCKEDDOWN but ESP not | |
539 | allowed). And then maybe some flavours of that that declare what is expected | |
540 | from home/srv/var… Then, add a new cmdline flag to all tools that parse such | |
541 | images, to configure this. Also, add a kernel cmdline option for this, to be | |
542 | honoured by the gpt auto generator. | |
543 | ||
544 | * nspawn: maybe optionally insert .nspawn file as GPT partition into images, so | |
545 | that such container images are entirely stand-alone and can be updated as | |
546 | one. | |
547 | ||
3fc0688d LP |
548 | * we probably should extend the root verity hash of the root fs into some PCR |
549 | on boot. (i.e. maybe add a crypttab option tpm2-measure=8 or so to measure it | |
550 | into PCR 8) | |
551 | ||
552 | * add a "policy" to the dissection logic. i.e. a bit mask what is OK to mount, | |
553 | what must be read-only, what requires encryption, and what requires | |
554 | authentication. | |
555 | ||
f4529c4d LP |
556 | * in uefi stub: query firmware regarding which PCRs are being used, store that |
557 | in EFI var. then use this when enrolling TPM2 in cryptsetup to verify that | |
558 | the selected PCRs actually are used by firmware. | |
559 | ||
322b3b38 LP |
560 | * rework recursive read-only remount to use new mount API |
561 | ||
63a185dc | 562 | * PAM: pick up authentication token from credentials |
9a6549f6 | 563 | |
9a6549f6 LP |
564 | * when mounting disk images: if IMAGE_ID/IMAGE_VERSION is set in os-release |
565 | data in the image, make sure the image filename actually matches this, so | |
566 | that images cannot be misused. | |
567 | ||
aca8ecc3 | 568 | * New udev block device symlink names: |
178d3ff2 | 569 | /dev/disk/by-parttypelabel/<pttype>-<ptlabel>. Use case: if pt label is used |
aca8ecc3 LP |
570 | as partition image version string, this is a safe way to reference a specific |
571 | version of a specific partition type, in particular where related partitions | |
572 | are processed (e.g. verity + rootfs both named "LennartOS_0.7"). | |
573 | ||
178d3ff2 LP |
574 | * sysupdate: |
575 | - add fuzzing to the pattern parser | |
576 | - support casync as download mechanism | |
577 | - direct TPM2 PCR change handling, possible renrolling LUKS2 media if needed. | |
578 | - "systemd-sysupdate update --all" support, that iterates through all components | |
579 | defined on the host, plus all images installed into /var/lib/machines/, | |
580 | /var/lib/portable/ and so on. | |
581 | - figure out what to do about system extensions (i.e. they need to imply an | |
582 | update component, since otherwise system extenion' sysupdate.d/ files would | |
583 | override the host's update files.) | |
584 | - Allow invocation with a single transfer definition, i.e. with | |
585 | --definitions= pointing to a file rather than a dir. | |
586 | - add ability to disable implicit decompression of downloaded artifacts, | |
587 | i.e. a Compress=no option in the transfer definitions | |
588 | ||
cf2ab2e7 LP |
589 | * in sd-id128: also parse UUIDs in RFC4122 URN syntax (i.e. chop off urn:uuid: prefix) |
590 | ||
31892e8d LP |
591 | * DynamicUser= + StateDirectory= → use uid mapping mounts, too, in order to |
592 | make dirs appear under right UID. | |
593 | ||
f3e58b55 LP |
594 | * systemd-sysext: optionally, run it in initrd already, before transitioning |
595 | into host, to open up possibility for services shipped like that. | |
596 | ||
49bd547b LP |
597 | * maybe add a tool that displays most recent journal logs as QR code to scan |
598 | off screen and run it automatically on boot failures, emergency logs and | |
599 | such. Use DRM APIs directly, see | |
600 | https://github.com/dvdhrm/docs/blob/master/drm-howto/modeset.c for an example | |
601 | for doing that. | |
602 | ||
24063ba1 | 603 | * introduce /dev/disk/root/* symlinks that allow referencing partitions on the |
f3e58b55 LP |
604 | disk the rootfs is on in a reasonably secure way. (or maybe: add |
605 | /dev/gpt-auto-{home,srv,boot,…} similar in style to /dev/gpt-auto-root as we | |
606 | already have it. | |
24063ba1 | 607 | |
7ed72cfa LP |
608 | * whenever we receive fds via SCM_RIGHTS make sure none got dropped due to the |
609 | reception limit the kernel silently enforces. | |
610 | ||
66e52d22 LP |
611 | * add an Open= setting to service unit files that can open arbitrary file |
612 | system paths at service startup time and pass them to the service process via | |
613 | our usual socket activation protocol. If passed path refers to AF_UNIX | |
614 | socket: connect() to it. | |
615 | ||
256cfb71 LP |
616 | * Similar, ConnectStream= which takes IP addresses and connects to them. |
617 | ||
618 | * Similar, Load= which takes literal data in text or base64 format, and puts it | |
619 | into a memfd, and passes that. This enables some fun stuff, such as embedding | |
620 | bash scripts in unit files, by combining Load= with ExecStart=/bin/bash | |
621 | /proc/self/fd/3 | |
622 | ||
66e52d22 LP |
623 | * add a ConnectSocket= setting to service unit files, that may reference a |
624 | socket unit, and which will connect to the socket defined therein, and pass | |
625 | the resulting fd to the service program via socket activation proto. | |
626 | ||
627 | * Add a concept of ListenStream=anonymous to socket units: listen on a socket | |
628 | that is deleted in the fs. Usecase would be with ConnectSocket= above. | |
629 | ||
14f7d087 LP |
630 | * importd: support image signature verification with PKCS#7 + OpenBSD signify |
631 | logic, as alternative to crummy gpg | |
632 | ||
9786ba13 LP |
633 | * add "systemd-analyze debug" + AttachDebugger= in unit files: The former |
634 | specifies a command to execute; the latter specifies that an already running | |
635 | "systemd-analyze debug" instance shall be contacted and execution paused | |
636 | until it gives an OK. That way, tools like gdb or strace can be safely be | |
637 | invoked on processes forked off PID 1. | |
638 | ||
38abd1bf LP |
639 | * expose MS_NOSYMFOLLOW in various places |
640 | ||
199b097d | 641 | * credentials system: |
199b097d LP |
642 | - acquire from EFI variable? |
643 | - acquire via via ask-password? | |
644 | - acquire creds via keyring? | |
645 | - pass creds via keyring? | |
646 | - pass creds via memfd? | |
647 | - acquire + decrypt creds from pkcs11? | |
648 | - make systemd-cryptsetup acquire pw via creds logic | |
649 | - make PAMName= acquire pw via creds logic | |
650 | - make macsec/wireguard code in networkd read key via creds logic | |
651 | - make gatwayd/remote read key via creds logic | |
652 | - add sd_notify() command for flushing out creds not needed anymore | |
1f066ce2 LB |
653 | - make user manager instances create and use a user-specific key (the one in |
654 | /var/lib is root-only) and add --user switch to systemd-creds to use it | |
199b097d | 655 | |
199b097d LP |
656 | * add tpm.target or so which is delayed until TPM2 device showed up in case |
657 | firmware indicates there is one. | |
07eabc2b | 658 | |
80670e74 LP |
659 | * Add concept for upgrading TPM2 enrollments, maybe a new switch |
660 | --pcrs=4:<hash> or so, i.e. select a PCR to include in the hash, and then | |
661 | override its hash | |
662 | ||
07eabc2b LB |
663 | * TPM2: auto-reenroll in cryptsetup, as fallback for hosed firmware upgrades |
664 | and such | |
665 | ||
666 | * introduce a new group to own TPM devices | |
80670e74 | 667 | |
80670e74 LP |
668 | * cryptsetup: if only recovery keys are registered and no regular passphrases, |
669 | ask user for "recovery key", not "passphrase" | |
670 | ||
671 | * cyptsetup: add option for automatically removing empty password slot on boot | |
672 | ||
7d7c75f1 | 673 | * cryptsetup: optionally, when run during boot-up and password is never |
80670e74 LP |
674 | entered, and we are on battery power (or so), power off machine again |
675 | ||
80670e74 LP |
676 | * cryptsetup: when waiting for FIDO2/PKCS#11 token, tell plymouth that, and |
677 | allow plymouth to abort the waiting and enter pw instead | |
7d7c75f1 | 678 | |
07eabc2b LB |
679 | * make cryptsetup lower --iter-time |
680 | ||
681 | * cryptsetup: allow encoding key directly in /etc/crypttab, maybe with a | |
682 | "base64:" prefix. Useful in particular for pkcs11 mode. | |
683 | ||
684 | * cryptsetup: reimplement the mkswap/mke2fs in cryptsetup-generator to use | |
685 | systemd-makefs.service instead. | |
686 | ||
687 | * cryptsetup: | |
688 | - cryptsetup-generator: allow specification of passwords in crypttab itself | |
689 | - support rd.luks.allow-discards= kernel cmdline params in cryptsetup generator | |
690 | ||
7d7c75f1 LP |
691 | * when configuring loopback netif, and it fails due to EPERM, eat up error if |
692 | it happens to be set up alright already. | |
693 | ||
694 | * at boot: check if battery above some threshold, if not power off again after explanation | |
695 | ||
696 | * userdb: add field for ambient caps, so that a user can have CAP_WAKE_ALARM | |
697 | for example. And add code that resets ambient caps for all services by | |
698 | default. | |
699 | ||
ff640bd2 LP |
700 | * sd-bus: when connecting to some dbus server socker, set originating AF_UNIX |
701 | socket name in abstract namespace to include "description" string, and pick | |
702 | it up from there in sd_bus_creds logic. i.e. we can use the socket peer | |
703 | address as conduit for some minimal connection metainfo, and use it to | |
704 | restore the "description" logic that kdbus used to have. | |
705 | ||
08d33656 | 706 | * systemd-analyze netif that explains predictable interface (or networkctl) |
d775d8e6 | 707 | |
816d460a LP |
708 | * Add service setting to run a service within the specified VRF. i.e. do the |
709 | equivalent of "ip vrf exec". | |
710 | ||
e59d030f LP |
711 | * change SwitchRoot() implementation in PID 1 to use pivot_root(".", "."), as |
712 | documented in the pivot_root(2) man page, so that we can drop the /oldroot | |
713 | temporary dir. | |
714 | ||
715 | * special case some calls of chase_symlinks() to use openat2() internally, so | |
716 | that the kernel does what we otherwise do. | |
717 | ||
6dbfbc46 LB |
718 | * add a new flag to chase_symlinks() that stops chasing once the first missing |
719 | component is found and then allows the caller to create the rest. | |
720 | ||
fc733bed LP |
721 | * make use of new glibc 2.32 APIs sigabbrev_np() and strerrorname_np(). |
722 | ||
effefa30 LP |
723 | * if /usr/bin/swapoff fails due to OOM, log a friendly explanatory message about it |
724 | ||
42ba8d25 LP |
725 | * Remove any support for booting without /usr pre-mounted in the initrd entirely. |
726 | Update INITRD_INTERFACE.md accordingly. | |
727 | ||
f1eb0ccd LP |
728 | * pid1: Move to tracking of main pid/control pid of units per pidfd |
729 | ||
730 | * pid1: support new clone3() fork-into-cgroup feature | |
731 | ||
7cc8fb3e LP |
732 | * pid1: also remove PID files of a service when the service starts, not just |
733 | when it exits | |
734 | ||
cdfd8537 LP |
735 | * make us use dynamically fewer deps for containers in general purpose distros: |
736 | o turn into dlopen() deps: | |
cdfd8537 LP |
737 | - p11-kit-trust (always) |
738 | - kmod-libs (only when called from PID 1) | |
a52dc0b6 | 739 | - libblkid (only in RootImage= handling in PID 1, but not elsewhere) |
cdfd8537 LP |
740 | - libpam (only when called from PID 1) |
741 | - bzip2, xz, lz4 (always — gzip and zstd should probably stay static deps the way they are, | |
742 | since they are so basic and our defaults) | |
743 | o move into separate libsystemd-shared-iptables.so .so | |
744 | - iptables-libs (only used by nspawn + networkd) | |
745 | ||
77169ed0 LP |
746 | * seccomp: maybe use seccomp_merge() to merge our filters per-arch if we can. |
747 | Apparently kernel performance is much better with fewer larger seccomp | |
748 | filters than with more smaller seccomp filters. | |
749 | ||
bfafec25 LP |
750 | * systemd-path: add ESP and XBOOTLDR path. Add "private" runtime/state/cache dir enum, |
751 | mapping to $RUNTIME_DIRECTORY, $STATE_DIRECTORY and such | |
752 | ||
a6e1018d | 753 | * All tools that support --root= should also learn --image= so that they can |
612d3a68 LP |
754 | operate on disk images directly. Specifically: bootctl, systemctl, |
755 | coredumpctl. (Already done: systemd-nspawn, systemd-firstboot, | |
756 | systemd-repart, systemd-tmpfiles, systemd-sysusers, journalctl) | |
a6e1018d | 757 | |
112bed84 LP |
758 | * seccomp: by default mask x32 ABI system wide on x86-64. it's on its way out |
759 | ||
760 | * seccomp: don't install filters for ABIs that are masked anyway for the | |
761 | specific service | |
762 | ||
cbe952fe LP |
763 | * busctl: maybe expose a verb "ping" for pinging a dbus service to see if it |
764 | exists and responds. | |
765 | ||
b47261e5 LP |
766 | * Maybe add a separate GPT partition type to the discoverable partition spec |
767 | for "hibernate" partitions, that are exactly like swap partitions but only | |
768 | activated right before hibernation and thus never used for regular swapping. | |
769 | ||
91fc013f | 770 | * socket units: allow creating a udev monitor socket with ListenDevices= or so, |
86b52a39 | 771 | with matches, then activate app through that passing socket over |
91fc013f | 772 | |
7e8facb3 | 773 | * unify on openssl: |
070f7370 LB |
774 | - kill gnutls support in resolved |
775 | - figure out what to do about libmicrohttpd, which has a hard dependency on | |
776 | gnutls | |
777 | - port fsprg over to a dlopen lib, then switch it to openssl | |
492f91d8 | 778 | |
492f91d8 LP |
779 | * add growvol and makevol options for /etc/crypttab, similar to |
780 | x-systemd.growfs and x-systemd-makefs. | |
781 | ||
77b19caf LP |
782 | * userdb: allow username prefix searches in varlink API, allow realname and |
783 | realname substr searches in varlink API | |
006c44c1 | 784 | |
76410e98 LP |
785 | * userdb: allow uid/gid range checks |
786 | ||
2a4be3c5 | 787 | * userdb: allow existence checks |
006c44c1 | 788 | |
f1eb0ccd | 789 | * pid1: activation by journal search expression |
006c44c1 | 790 | |
492f91d8 LP |
791 | * when switching root from initrd to host, set the machine_id env var so that |
792 | if the host has no machine ID set yet we continue to use the random one the | |
793 | initrd had set. | |
794 | ||
173c7873 | 795 | * sd-event: add native support for P_ALL waitid() watching, then move PID 1 to |
84a1ff94 | 796 | it for reaping assigned but unknown children. This needs to some special care |
173c7873 LP |
797 | to operate somewhat sensibly in light of priorities: P_ALL will return |
798 | arbitrary processes, regardless of the priority we want to watch them with, | |
799 | hence on each event loop iteration check all processes which we shall watch | |
800 | with higher prio explicitly, and then watch the entire rest with P_ALL. | |
801 | ||
802 | * tweak sd-event's child watching: keep a prioq of children to watch and use | |
803 | waitid() only on the children with the highest priority until one is waitable | |
804 | and ignore all lower-prio ones from that point on | |
805 | ||
173c7873 LP |
806 | * maybe introduce xattrs that can be set on the root dir of the root fs |
807 | partition that declare the volatility mode to use the image in. Previously I | |
808 | thought marking this via GPT partition flags but that's not ideal since | |
809 | that's outside of the LUKS encryption/verity verification, and we probably | |
810 | shouldn't operate in a volatile mode unless we got told so from a trusted | |
811 | source. | |
812 | ||
97e5cc88 LP |
813 | * coredump: maybe when coredumping read a new xattr from /proc/$PID/exe that |
814 | may be used to mark a whole binary as non-coredumpable. Would fix: | |
815 | https://bugs.freedesktop.org/show_bug.cgi?id=69447 | |
816 | ||
f3e361c1 LP |
817 | * teach parse_timestamp() timezones like the calendar spec already knows it |
818 | ||
9ef3376b LP |
819 | * beef up hibernation to optionally do swapon/swapoff immediately before/after |
820 | the hibernation | |
821 | ||
c846b233 LP |
822 | * beef up s2h to implement a battery watch loop: instead of entering |
823 | hibernation unconditionally after coming back from resume make a decision | |
824 | based on the battery load level: if battery level is above a specific | |
825 | threshold, go to suspend again, only hibernate if below it. This means we'd | |
826 | stick to suspend usually, but fall back to hibernation only when battery runs | |
827 | empty (well, subject to our sampling interval). Related to this, check if we | |
828 | can make ACPI _BTP (i.e. /sys/class/power_supply/*/alarm) work for us too, | |
829 | i.e. see if it can wake up machines from suspend, so that we could resume | |
830 | automatically when the system is low on power and move automatically to | |
831 | hibernation mode. (see | |
832 | https://uefi.org/sites/default/files/resources/ACPI%206_2_A_Sept29.pdf | |
833 | section 10.2.2.8 and | |
834 | https://docs.microsoft.com/en-us/windows-hardware/design/device-experiences/modern-standby-wake-sources | |
835 | at the end). | |
836 | ||
6fe23ff3 LB |
837 | * We should probably replace /etc/rc.d/README with a symlink to doc |
838 | content. After all it is constant vendor data. | |
9c230b8f | 839 | |
c6526b8d | 840 | * maybe add kernel cmdline params: to force random seed crediting |
312dc153 | 841 | |
ac5dca64 LP |
842 | * introduce a new per-process uuid, similar to the boot id, the machine id, the |
843 | invocation id, that is derived from process creds, specifically a hashed | |
844 | combination of AT_RANDOM + getpid() + the starttime from | |
845 | /proc/self/status. Then add these ids implicitly when logging. Deriving this | |
846 | uuid from these three things has the benefit that it can be derived easily | |
847 | from /proc/$PID/ in a stable, and unique way that changes on both fork() and | |
848 | exec(). | |
849 | ||
850 | * let's not GC a unit while its ratelimits are still pending | |
851 | ||
5daeeecf LP |
852 | * when killing due to service watchdog timeout maybe detect whether target |
853 | process is under ptracing and then log loudly and continue instead. | |
854 | ||
5802d977 LP |
855 | * make rfkill uaccess controllable by default, i.e. steal rule from |
856 | gnome-bluetooth and friends | |
857 | ||
03ccc7f0 LP |
858 | * make MAINPID= message reception checks even stricter: if service uses User=, |
859 | then check sending UID and ignore message if it doesn't match the user or | |
860 | root. | |
861 | ||
862 | * maybe trigger a uevent "change" on a device if "systemctl reload xyz.device" | |
863 | is issued. | |
864 | ||
db3cea22 LP |
865 | * when importing an fs tree with machined, optionally apply userns-rec-chown |
866 | ||
867 | * when importing an fs tree with machined, complain if image is not an OS | |
868 | ||
707b3fbd LP |
869 | * Maybe introduce a helper safe_exec() or so, which is to execve() which |
870 | safe_fork() is to fork(). And then make revert the RLIMIT_NOFILE soft limit | |
871 | to 1K implicitly, unless explicitly opted-out. | |
d7b659ef | 872 | |
d238709c | 873 | * rework seccomp/nnp logic that even if User= is used in combination with |
d96c081a LP |
874 | a seccomp option we don't have to set NNP. For that, change uid first whil |
875 | keeping CAP_SYS_ADMIN, then apply seccomp, the drop cap. | |
876 | ||
8a7cf157 LP |
877 | * when no locale is configured, default to UEFI's PlatformLang variable |
878 | ||
d96c081a LP |
879 | * add a new syscall group "@esoteric" for more esoteric stuff such as bpf() and |
880 | usefaultd() and make systemd-analyze check for it. | |
881 | ||
06898123 | 882 | * paranoia: whenever we process passwords, call mlock() on the memory |
309c6b19 | 883 | first. i.e. look for all places we use free_and_erasep() and |
9ae4ef49 | 884 | augment them with mlock(). Also use MADV_DONTDUMP. |
8b213bf1 | 885 | Alternatively (preferably?) use memfd_secret(). |
06898123 | 886 | |
32875617 LP |
887 | * Move RestrictAddressFamily= to the new cgroup create socket |
888 | ||
76853293 LP |
889 | * maybe implicitly attach monotonic+realtime timestamps to outgoing messages in |
890 | log.c and sd-journal-send | |
891 | ||
06898123 LP |
892 | * optionally: turn on cgroup delegation for per-session scope units |
893 | ||
d3aeddb8 LP |
894 | * introduce per-unit (i.e. per-slice, per-service) journal log size limits. |
895 | ||
42e18088 LP |
896 | * sd-boot: optionally, show boot menu when previous default boot item has |
897 | non-zero "tries done" count | |
898 | ||
8f2eb730 LP |
899 | * augment CODE_FILE=, CODE_LINE= with something like CODE_BASE= or so which |
900 | contains some identifier for the project, which allows us to include | |
901 | clickable links to source files generating these log messages. The identifier | |
902 | could be some abberviated URL prefix or so (taking inspiration from Go | |
903 | imports). For example, for systemd we could use | |
904 | CODE_BASE=github.com/systemd/systemd/blob/98b0b1123cc or so which is | |
905 | sufficient to build a link by prefixing "http://" and suffixing the | |
906 | CODE_FILE. | |
907 | ||
8f2eb730 LP |
908 | * Augment MESSAGE_ID with MESSAGE_BASE, in a similar fashion so that we can |
909 | make clickable links from log messages carrying a MESSAGE_ID, that lead to | |
910 | some explanatory text online. | |
911 | ||
7bd4bcf7 LP |
912 | * maybe extend .path units to expose fanotify() per-mount change events |
913 | ||
bb527e11 LP |
914 | * When reloading configuration PID 1 should reset all its properties to the |
915 | original defaults before calling parse_config() | |
916 | ||
c633b0a6 LP |
917 | * hibernate/s2h: make this robust and safe to enable in Fedora by default. |
918 | Specifically: | |
919 | ||
920 | 1. add resume_offset support to the resume code (i.e. support swap files | |
921 | properly) | |
e83419d0 | 922 | 2. check if swap is on weird storage and refuse if so |
46c41478 | 923 | 3. add auto-detection of hibernation images |
c633b0a6 | 924 | |
070d0ac9 LP |
925 | * cgroups: use inotify to get notified when somebody else modifies cgroups |
926 | owned by us, then log a friendly warning. | |
927 | ||
e44924f5 LP |
928 | * beef up log.c with support for stripping ANSI sequences from strings, so that |
929 | it is OK to include them in log strings. This would be particularly useful so | |
930 | that our log messages could contain clickable links for example for unit | |
931 | files and suchlike we operate on. | |
070d0ac9 | 932 | |
07eabc2b LB |
933 | * importd: add ability download images for portabled + sysext |
934 | ||
8c4c2dfc LP |
935 | * add support for "portablectl attach http://foobar.com/waaa.raw (i.e. importd integration) |
936 | ||
8c4c2dfc LP |
937 | * sync dynamic uids/gids between host+portable srvice (i.e. if DynamicUser=1 is set for a service, make sure that the |
938 | selected user is resolvable in the service even if it ships its own /etc/passwd) | |
939 | ||
5da19043 | 940 | * Fix DECIMAL_STR_MAX or DECIMAL_STR_WIDTH. One includes a trailing NUL, the |
63a185dc | 941 | other doesn't. What a disaster. Probably to exclude it. |
5da19043 | 942 | |
16270697 LP |
943 | * Check that users of inotify's IN_DELETE_SELF flag are using it properly, as |
944 | usually IN_ATTRIB is the right way to watch deleted files, as the former only | |
945 | fires when a file is actually removed from disk, i.e. the link count drops to | |
946 | zero and is not open anymore, while the latter happens when a file is | |
947 | unlinked from any dir. | |
948 | ||
bd1b3f75 | 949 | * port systemctl, busctl, … over to format-table.[ch]'s table formatters |
5da19043 | 950 | |
7bc756ff LP |
951 | * pid1: lock image configured with RootDirectory=/RootImage= using the usual nspawn semantics while the unit is up |
952 | ||
953 | * add --vacuum-xyz options to coredumpctl, matching those journalctl already has. | |
954 | ||
53c70a27 LP |
955 | * introduce Ephemeral= unit file switch, that creates an ephemeral copy of all |
956 | files and directories that are left writable for a unit, and which are | |
957 | removed after the unit goes down again. A bit like --ephemeral for | |
958 | systemd-nspawn but for system services. If used together with RootImage= this | |
959 | should reflink the image file itself. | |
960 | ||
961 | Related: add Ephemeral=<path1> <path2> … which would allow marking | |
962 | specific paths only like this. | |
963 | ||
53c70a27 | 964 | * add CopyFile= or so as unit file setting that may be used to copy files or |
5238e957 | 965 | directory trees from the host to the services RootImage= and RootDirectory= |
53c70a27 LP |
966 | environment. Which we can use for /etc/machine-id and in particular |
967 | /etc/resolv.conf. Should be smart and do something useful on read-only | |
2aed63f4 | 968 | images, for example fall back to read-only bind mounting the file instead. |
53c70a27 | 969 | |
d9b50610 LP |
970 | * show invocation ID in systemd-run output |
971 | ||
972 | * bypass SIGTERM state in unit files if KillSignal is SIGKILL | |
973 | ||
9711b1ad LP |
974 | * add proper dbus APIs for the various sd_notify() commands, such as MAINPID=1 |
975 | and so on, which would mean we could report errors and such. | |
976 | ||
6b7b0f39 LP |
977 | * introduce DefaultSlice= or so in system.conf that allows changing where we |
978 | place our units by default, i.e. change system.slice to something | |
979 | else. Similar, ManagerSlice= should exist so that PID1's own scope unit could | |
980 | be moved somewhere else too. Finally machined and logind should get similar | |
981 | options so that it is possible to move user session scopes and machines to a | |
982 | different slice too by default. Usecase: people who want to put resources on | |
983 | the entire system, with the exception of one specific service. See: | |
984 | https://lists.freedesktop.org/archives/systemd-devel/2018-February/040369.html | |
985 | ||
aa79f932 LP |
986 | * maybe rework get_user_creds() to query the user database if $SHELL is used |
987 | for root, but only then. | |
586a8e93 | 988 | |
46099c9e LP |
989 | * be stricter with fds we receive for the fdstore: close them asynchronously |
990 | ||
46099c9e LP |
991 | * calenderspec: add support for week numbers and day numbers within a |
992 | year. This would allow us to define "bi-weekly" triggers safely. | |
993 | ||
c3cd7cc9 LP |
994 | * sd-bus: add vtable flag, that may be used to request client creds implicitly |
995 | and asynchronously before dispatching the operation | |
996 | ||
d0e5db44 ZJS |
997 | * sd-bus: parse addresses given in sd_bus_set_addresses immediately and not |
998 | only when used. Add unit tests. | |
999 | ||
06345858 LP |
1000 | * make use of ethtool veth peer info in machined, for automatically finding out |
1001 | host-side interface pointing to the container. | |
1002 | ||
48f1b5e5 LP |
1003 | * add some special mode to LogsDirectory=/StateDirectory=… that allows |
1004 | declaring these directories without necessarily pulling in deps for them, or | |
1005 | creating them when starting up. That way, we could declare that | |
1006 | systemd-journald writes to /var/log/journal, which could be useful when we | |
1007 | doing disk usage calculations and so on. | |
1008 | ||
899feb72 | 1009 | * deprecate RootDirectoryStartOnly= in favour of a new ExecStart= prefix char |
c6009ff0 | 1010 | |
3d80d454 LP |
1011 | * add a new RuntimeDirectoryPreserve= mode that defines a similar lifecycle for |
1012 | the runtime dir as we maintain for the fdstore: i.e. keep it around as long | |
1013 | as the unit is running or has a job queued. | |
1014 | ||
5f7ecd61 | 1015 | * support projid-based quota in machinectl for containers |
5962e9db | 1016 | |
17b6f896 LP |
1017 | * add a way to lock down cgroup migration: a boolean, which when set for a unit |
1018 | makes sure the processes in it can never migrate out of it | |
1019 | ||
17b6f896 LP |
1020 | * blog about fd store and restartable services |
1021 | ||
1022 | * document Environment=SYSTEMD_LOG_LEVEL=debug drop-in in debugging document | |
1023 | ||
370f9c21 LP |
1024 | * rework ExecOutput and ExecInput enums so that EXEC_OUTPUT_NULL loses its |
1025 | magic meaning and is no longer upgraded to something else if set explicitly. | |
1026 | ||
17b6f896 LP |
1027 | * in the long run: permit a system with /etc/machine-id linked to /dev/null, to |
1028 | make it lose its identity, i.e. be anonymous. For this we'd have to patch | |
1029 | through the whole tree to make all code deal with the case where no machine | |
1030 | ID is available. | |
1031 | ||
1032 | * optionally, collect cgroup resource data, and store it in per-unit RRD files, | |
1033 | suitable for processing with rrdtool. Add bus API to access this data, and | |
1034 | possibly implement a CPULoad property based on it. | |
1035 | ||
e7e4a258 LP |
1036 | * beef up pam_systemd to take unit file settings such as cgroups properties as |
1037 | parameters | |
1038 | ||
63a185dc | 1039 | * maybe hook up xfs/ext4 quotactl() with services? i.e. automatically manage |
d51c4fca | 1040 | the quota of the user indicated in User= via unit file settings, like the |
5962e9db LP |
1041 | other resource management concepts. Would mix nicely with DynamicUser=1. Or |
1042 | alternatively, do this with projids, so that we can also cover services | |
1043 | running as root. Quota should probably cover all the special dirs such as | |
1044 | StateDirectory=, LogsDirectory=, CacheDirectory=, as well as RootDirectory= if it | |
1045 | is set, plus the whole disk space any image configured with RootImage=. | |
6fc373ee | 1046 | |
9a92a97a LP |
1047 | * In DynamicUser= mode: before selecting a UID, use disk quota APIs on relevant |
1048 | disks to see if the UID is already in use. | |
1049 | ||
fa991fb7 LP |
1050 | * expose IO accounting data on the bus, show it in systemd-run --wait and log |
1051 | about it in the resource log message | |
1052 | ||
d73b607d LP |
1053 | * Add AddUser= setting to unit files, similar to DynamicUser=1 which however |
1054 | creates a static, persistent user rather than a dynamic, transient user. We | |
1055 | can leverage code from sysusers.d for this. | |
1056 | ||
fd63f36c LP |
1057 | * add some optional flag to ReadWritePaths= and friends, that has the effect |
1058 | that we create the dir in question when the service is started. Example: | |
1059 | ||
1060 | ReadWritePaths=:/var/lib/foobar | |
1061 | ||
f59d1da8 LP |
1062 | * Add ExecMonitor= setting. May be used multiple times. Forks off a process in |
1063 | the service cgroup, which is supposed to monitor the service, and when it | |
1064 | exits the service is considered failed by its monitor. | |
1065 | ||
33bac67b LP |
1066 | * track the per-service PAM process properly (i.e. as an additional control |
1067 | process), so that it may be queried on the bus and everything. | |
1068 | ||
1069 | * add a new "debug" job mode, that is propagated to unit_start() and for | |
1070 | services results in two things: we raise SIGSTOP right before invoking | |
1071 | execve() and turn off watchdog support. Then, use that to implement | |
1072 | "systemd-gdb" for attaching to the start-up of any system service in its | |
1073 | natural habitat. | |
1074 | ||
8eb7383b LP |
1075 | * 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 |
1076 | ||
b5bdbcd5 LP |
1077 | * add a percentage syntax for TimeoutStopSec=, e.g. TimeoutStopSec=150%, and |
1078 | then use that for the setting used in user@.service. It should be understood | |
1079 | relative to the configured default value. | |
1080 | ||
d21494ea LP |
1081 | * enable LockMLOCK to take a percentage value relative to physical memory |
1082 | ||
04397464 | 1083 | * Permit masking specific netlink APIs with RestrictAddressFamily= |
d82047be | 1084 | |
d82047be LP |
1085 | * define gpt header bits to select volatility mode |
1086 | ||
42d61ded LP |
1087 | * ProtectClock= (drops CAP_SYS_TIMES, adds seecomp filters for settimeofday, adjtimex), sets DeviceAllow o /dev/rtc |
1088 | ||
04397464 | 1089 | * ProtectTracing= (drops CAP_SYS_PTRACE, blocks ptrace syscall, makes /sys/kernel/tracing go away) |
42d61ded | 1090 | |
04397464 | 1091 | * ProtectMount= (drop mount/umount/pivot_root from seccomp, disallow fuse via DeviceAllow, imply Mountflags=slave) |
563a69f4 | 1092 | |
04397464 | 1093 | * ProtectKeyRing= to take keyring calls away |
2c5f2958 | 1094 | |
8ce9b83a | 1095 | * RemoveKeyRing= to remove all keyring entries of the specified user |
a46eac1b LP |
1096 | |
1097 | * ProtectReboot= that masks reboot() and kexec_load() syscalls, prohibits kill | |
1098 | on PID 1 with the relevant signals, and makes relevant files in /sys and | |
1099 | /proc (such as the sysrq stuff) unavailable | |
1100 | ||
88511a37 LB |
1101 | * Support ReadWritePaths/ReadOnlyPaths/InaccessiblePaths in systemd --user instances |
1102 | via the new unprivileged Landlock LSM (https://landlock.io) | |
1103 | ||
e40a326c | 1104 | * make sure the ratelimit object can deal with USEC_INFINITY as way to turn off things |
89f193fa | 1105 | |
8ce9b83a LP |
1106 | * in nss-systemd, if we run inside of RootDirectory= with PrivateUsers= set, |
1107 | find a way to map the User=/Group= of the service to the right name. This way | |
1108 | a user/group for a service only has to exist on the host for the right | |
1109 | mapping to work. | |
1110 | ||
2c5f2958 LP |
1111 | * add bus API for creating unit files in /etc, reusing the code for transient units |
1112 | ||
1113 | * add bus API to remove unit files from /etc | |
1114 | ||
1115 | * add bus API to retrieve current unit file contents (i.e. implement "systemctl cat" on the bus only) | |
1116 | ||
b8c7afdf LP |
1117 | * rework fopen_temporary() to make use of open_tmpfile_linkable() (problem: the |
1118 | kernel doesn't support linkat() that replaces existing files, currently) | |
1119 | ||
1e555cb5 LP |
1120 | * transient units: don't bother with actually setting unit properties, we |
1121 | reload the unit file anyway | |
1122 | ||
c8048350 LP |
1123 | * optionally, also require WATCHDOG=1 notifications during service start-up and shutdown |
1124 | ||
3d39e6e5 LP |
1125 | * cache sd_event_now() result from before the first iteration... |
1126 | ||
bd098bce LP |
1127 | * PID1: find a way how we can reload unit file configuration for |
1128 | specific units only, without reloading the whole of systemd | |
1129 | ||
f9bf1b8f | 1130 | * add an explicit parser for LimitRTPRIO= that verifies |
de7399eb | 1131 | the specified range and generates sane error messages for incorrect |
f9bf1b8f | 1132 | specifications. |
de7399eb | 1133 | |
3efc8c72 LP |
1134 | * when we detect that there are waiting jobs but no running jobs, do something |
1135 | ||
e6a26d8c LP |
1136 | * PID 1 should send out sd_notify("WATCHDOG=1") messages (for usage in the --user mode, and when run via nspawn) |
1137 | ||
a2088fd0 | 1138 | * there's probably something wrong with having user mounts below /sys, |
5238e957 | 1139 | as we have for debugfs. for example, src/core/mount.c handles mounts |
a2088fd0 | 1140 | prefixed with /sys generally special. |
41d6f3bf | 1141 | https://lists.freedesktop.org/archives/systemd-devel/2015-June/032962.html |
a2088fd0 | 1142 | |
8aa20381 LP |
1143 | * fstab-generator: default to tmpfs-as-root if only usr= is specified on the kernel cmdline |
1144 | ||
a7972611 ZJS |
1145 | * initrd-parse-etc.service: can we skip daemon-reload if /sysroot/etc/fstab is missing? |
1146 | Note that we start initrd-fs.target and initrd-cleanup.target there, so a straightforward | |
1147 | ConditionPathExists= is not enough. | |
1148 | ||
931bc195 | 1149 | * docs: bring https://www.freedesktop.org/wiki/Software/systemd/MyServiceCantGetRealtime up to date |
b18d23d7 | 1150 | |
60d17b74 LP |
1151 | * add a job mode that will fail if a transaction would mean stopping |
1152 | running units. Use this in timedated to manage the NTP service | |
1153 | state. | |
41d6f3bf | 1154 | https://lists.freedesktop.org/archives/systemd-devel/2015-April/030229.html |
60d17b74 | 1155 | |
477e75ef LP |
1156 | * The udev blkid built-in should expose a property that reflects |
1157 | whether media was sensed in USB CF/SD card readers. This should then | |
1158 | be used to control SYSTEMD_READY=1/0 so that USB card readers aren't | |
1159 | picked up by systemd unless they contain a medium. This would mirror | |
1160 | the behaviour we already have for CD drives. | |
1161 | ||
c3a0d00d LP |
1162 | * hostnamectl: show root image uuid |
1163 | ||
d2f81fb0 | 1164 | * Find a solution for SMACK capabilities stuff: |
41d6f3bf | 1165 | https://lists.freedesktop.org/archives/systemd-devel/2014-December/026188.html |
98cd2651 | 1166 | |
0a86c1a9 | 1167 | * synchronize console access with BSD locks: |
41d6f3bf | 1168 | https://lists.freedesktop.org/archives/systemd-devel/2014-October/024582.html |
0a86c1a9 | 1169 | |
e031c227 | 1170 | * as soon as we have sender timestamps, revisit coalescing multiple parallel daemon reloads: |
41d6f3bf | 1171 | https://lists.freedesktop.org/archives/systemd-devel/2014-December/025862.html |
0a86c1a9 | 1172 | |
8514b677 LP |
1173 | * figure out when we can use the coarse timers |
1174 | ||
8483d73f LP |
1175 | * maybe allow timer units with an empty Units= setting, so that they |
1176 | can be used for resuming the system but nothing else. | |
1177 | ||
25e773ee | 1178 | * what to do about udev db binary stability for apps? (raw access is not an option) |
b857e042 | 1179 | |
720652b3 | 1180 | * exponential backoff in timesyncd when we cannot reach a server |
42aeb14a | 1181 | |
720652b3 | 1182 | * timesyncd: add ugly bus calls to set NTP servers per-interface, for usage by NM |
e25b5a8d | 1183 | |
563b1bdc LP |
1184 | * merge ~/.local/share and ~/.local/lib into one similar /usr/lib and /usr/share.... |
1185 | ||
6bd7941e TG |
1186 | * add systemd.abort_on_kill or some other such flag to send SIGABRT instead of SIGKILL |
1187 | (throughout the codebase, not only PID1) | |
1188 | ||
07eabc2b LB |
1189 | * drop nss-myhostname in favour of nss-resolve? |
1190 | ||
9d6db739 | 1191 | * resolved: |
9d6db739 | 1192 | - mDNS/DNS-SD |
ccc3e8a1 LP |
1193 | - service registration |
1194 | - service/domain/types browsing | |
0f47ed0a | 1195 | - avahi compat |
9d6db739 | 1196 | - DNS-SD service registration from socket units |
e25b5a8d DH |
1197 | - resolved should optionally register additional per-interface LLMNR |
1198 | names, so that for the container case we can establish the same name | |
1199 | (maybe "host") for referencing the server, everywhere. | |
720652b3 | 1200 | - allow clients to request DNSSEC for a single lookup even if DNSSEC is off (?) |
3efb871a | 1201 | - hook up resolved with machined-based address resolution |
3f77a1b1 | 1202 | |
e25b5a8d | 1203 | * refcounting in sd-resolve is borked |
e2a69298 | 1204 | |
a940778f LP |
1205 | * add new gpt type for btrfs volumes |
1206 | ||
3de03738 LP |
1207 | * generator that automatically discovers btrfs subvolumes, identifies their purpose based on some xattr on them. |
1208 | ||
37efac5d LP |
1209 | * a way for container managers to turn off getty starting via $container_headless= or so... |
1210 | ||
7348b3ad LP |
1211 | * figure out a nice way how we can let the admin know what child/sibling unit causes cgroup membership for a specific unit |
1212 | ||
81429136 KS |
1213 | * For timer units: add some mechanisms so that timer units that trigger immediately on boot do not have the services |
1214 | they run added to the initial transaction and thus confuse Type=idle. | |
e107ed18 | 1215 | |
edb2935c LP |
1216 | * add bus api to query unit file's X fields. |
1217 | ||
6a3f892a | 1218 | * gpt-auto-generator: |
2a781fc9 | 1219 | - Define new partition type for encrypted swap? Support probed LUKS for encrypted swap? |
6a3f892a | 1220 | - Make /home automount rather than mount? |
6a3f892a | 1221 | |
65026403 LP |
1222 | * add generator that pulls in systemd-network from containers when |
1223 | CAP_NET_ADMIN is set, more than the loopback device is defined, even | |
1224 | when it is otherwise off | |
f8901862 | 1225 | |
f9bf1b8f | 1226 | * MessageQueueMessageSize= (and suchlike) should use parse_iec_size(). |
eda8f067 | 1227 | |
af1082b0 LP |
1228 | * implement Distribute= in socket units to allow running multiple |
1229 | service instances processing the listening socket, and open this up | |
1230 | for ReusePort= | |
1231 | ||
f38afcd0 | 1232 | * cgroups: |
f38afcd0 | 1233 | - implement per-slice CPUFairScheduling=1 switch |
f38afcd0 LP |
1234 | - introduce high-level settings for RT budget, swappiness |
1235 | - how to reset dynamically changed unit cgroup attributes sanely? | |
1236 | - when reloading configuration, apply new cgroup configuration | |
1237 | - when recursively showing the cgroup hierarchy, optionally also show | |
1238 | the hierarchies of child processes | |
b8df7f86 ZJS |
1239 | - add settings for cgroup.max.descendants and cgroup.max.depth, |
1240 | maybe use them for user@.service | |
0bee65f0 | 1241 | |
f38afcd0 | 1242 | * transient units: |
f38afcd0 | 1243 | - add field to transient units that indicate whether systemd or somebody else saves/restores its settings, for integration with libvirt |
ebcf1f97 | 1244 | |
718db961 LP |
1245 | * when we detect low battery and no AC on boot, show pretty splash and refuse boot |
1246 | ||
718db961 LP |
1247 | * libsystemd-journal, libsystemd-login, libudev: add calls to easily attach these objects to sd-event event loops |
1248 | ||
966204e0 LP |
1249 | * be more careful what we export on the bus as (usec_t) 0 and (usec_t) -1 |
1250 | ||
41644622 LP |
1251 | * 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 |
1252 | ||
7f79cd71 | 1253 | * After coming back from hibernation reset hibernation swap partition using the /dev/snapshot ioctl APIs |
0aafd43d | 1254 | |
19aadacf JE |
1255 | * If we try to find a unit via a dangling symlink, generate a clean |
1256 | error. Currently, we just ignore it and read the unit from the search | |
df5f6971 LP |
1257 | path anyway. |
1258 | ||
04397464 | 1259 | * refuse boot if /usr/lib/os-release is missing or /etc/machine-id cannot be set up |
fcba531e | 1260 | |
bdeeb6b5 LP |
1261 | * man: the documentation of Restart= currently is very misleading and suggests the tools from ExecStartPre= might get restarted. |
1262 | ||
1263 | * load .d/*.conf dropins for device units | |
1264 | ||
07eabc2b | 1265 | * There's currently no way to cancel fsck (used to be possible via C-c or c on the console) |
f38afcd0 | 1266 | |
07eabc2b | 1267 | * add option to sockets to avoid activation. Instead just drop packets/connections, see http://cyberelk.net/tim/2012/02/15/portreserve-systemd-solution/ |
eb01ba5d | 1268 | |
07eabc2b LB |
1269 | * make sure systemd-ask-password-wall does not shutdown systemd-ask-password-console too early |
1270 | ||
1271 | * verify that the AF_UNIX sockets of a service in the fs still exist | |
1272 | when we start a service in order to avoid confusion when a user | |
1273 | assumes starting a service is enough to make it accessible | |
1274 | ||
1275 | * Make it possible to set the keymap independently from the font on | |
1276 | the kernel cmdline. Right now setting one resets also the other. | |
1277 | ||
1278 | * and a dbus call to generate target from current state | |
1279 | ||
1280 | * investigate whether the gnome pty helper should be moved into systemd, to provide cgroup support. | |
1281 | ||
1282 | * dot output for --test showing the 'initial transaction' | |
1283 | ||
1284 | * be able to specify a forced restart of service A where service B depends on, in case B | |
1285 | needs to be auto-respawned? | |
1286 | ||
1287 | * pid1: | |
1288 | - When logging about multiple units (stopping BoundTo units, conflicts, etc.), | |
1289 | log both units as UNIT=, so that journalctl -u triggers on both. | |
1290 | - generate better errors when people try to set transient properties | |
1291 | that are not supported... | |
41d6f3bf | 1292 | https://lists.freedesktop.org/archives/systemd-devel/2015-February/028076.html |
07eabc2b | 1293 | - maybe introduce WantsMountsFor=? Usecase: |
41d6f3bf | 1294 | https://lists.freedesktop.org/archives/systemd-devel/2015-January/027729.html |
07eabc2b LB |
1295 | - recreate systemd's D-Bus private socket file on SIGUSR2 |
1296 | - move PAM code into its own binary | |
1297 | - when we automatically restart a service, ensure we restart its rdeps, too. | |
1298 | - hide PAM options in fragment parser when compile time disabled | |
1299 | - Support --test based on current system state | |
1300 | - 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(). | |
1301 | - after deserializing sockets in socket.c we should reapply sockopts and things | |
1302 | - drop PID 1 reloading, only do reexecing (difficult: Reload() | |
1303 | currently is properly synchronous, Reexec() is weird, because we | |
1304 | cannot delay the response properly until we are back, so instead of | |
1305 | being properly synchronous we just keep open the fd and close it | |
1306 | when done. That means clients do not get a successful method reply, | |
1307 | but much rather a disconnect on success. | |
1308 | - when breaking cycles drop sysv services first, then services from /run, then from /etc, then from /usr | |
1309 | - when a bus name of a service disappears from the bus make sure to queue further activation requests | |
1310 | - maybe introduce CoreScheduling=yes/no to optionally set a PR_SCHED_CORE cookie, so that all | |
1311 | processes in a service's cgroup share the same cookie and are guaranteed not to share SMT cores | |
1312 | with other units https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/admin-guide/hw-vuln/core-scheduling.rst | |
1313 | ||
1314 | * unit files: | |
1315 | - allow port=0 in .socket units | |
1316 | - maybe introduce ExecRestartPre= | |
1317 | - add ReloadSignal= for configuring a reload signal to use | |
1318 | - implement Register= switch in .socket units to enable registration | |
1319 | in Avahi, RPC and other socket registration services. | |
1320 | - allow Type=simple with PIDFile= | |
1321 | https://bugzilla.redhat.com/show_bug.cgi?id=723942 | |
1322 | - allow writing multiple conditions in unit files on one line | |
1323 | - introduce Type=pid-file | |
1324 | - add a concept of RemainAfterExit= to scope units | |
1325 | - Allow multiple ExecStart= for all Type= settings, so that we can cover rescue.service nicely | |
1326 | - add verification of [Install] section to systemd-analyze verify | |
1327 | ||
1328 | * timer units: | |
63a185dc | 1329 | - timer units should get the ability to trigger when DST changes |
07eabc2b LB |
1330 | - Modulate timer frequency based on battery state |
1331 | ||
1332 | * add libsystemd-password or so to query passwords during boot using the password agent logic | |
1333 | ||
1334 | * clean up date formatting and parsing so that all absolute/relative timestamps we format can also be parsed | |
1335 | ||
1336 | * on shutdown: move utmp, wall, audit logic all into PID 1 (or logind?), get rid of systemd-update-utmp-runlevel | |
1337 | ||
1338 | * make repeated alt-ctrl-del presses printing a dump | |
1339 | ||
07eabc2b | 1340 | * currently x-systemd.timeout is lost in the initrd, since crypttab is copied into dracut, but fstab is not |
461bd8e4 | 1341 | |
ab8e074c LP |
1342 | * add a pam module that passes the hdd passphrase into the PAM stack and then expires it, for usage by gdm auto-login. |
1343 | ||
1344 | * add a pam module that on password changes updates any LUKS slot where the password matches | |
1345 | ||
fff87a35 | 1346 | * test/: |
20d52ab6 | 1347 | - add unit tests for config_parse_device_allow() |
b8b4d3dd | 1348 | |
b5c03638 | 1349 | * seems that when we follow symlinks to units we prefer the symlink |
d28315e4 | 1350 | destination path over /etc and /usr. We should not do that. Instead |
b5c03638 LP |
1351 | /etc should always override /run+/usr and also any symlink |
1352 | destination. | |
1353 | ||
eece8c6f LP |
1354 | * when isolating, try to figure out a way how we implicitly can order |
1355 | all units we stop before the isolating unit... | |
1356 | ||
356ce991 LP |
1357 | * teach ConditionKernelCommandLine= globs or regexes (in order to match foobar={no,0,off}) |
1358 | ||
6daebf9e ZJS |
1359 | * Add ConditionDirectoryNotEmpty= handle non-absoute paths as a search path or add |
1360 | ConditionConfigSearchPathNotEmpty= or different syntax? See the discussion starting at | |
1361 | https://github.com/systemd/systemd/pull/15109#issuecomment-607740136. | |
1362 | ||
1ccfb792 LP |
1363 | * BootLoaderSpec: Define a way how an installer can figure out whether a BLS |
1364 | compliant boot loader is installed. | |
795607b2 | 1365 | |
0be8342c LP |
1366 | * think about requeuing jobs when daemon-reload is issued? usecase: |
1367 | the initrd issues a reload after fstab from the host is accessible | |
1368 | and we might want to requeue the mounts local-fs acquired through | |
1369 | that automatically. | |
1370 | ||
e5ec62c5 | 1371 | * systemd-inhibit: make taking delay locks useful: support sending SIGINT or SIGTERM on PrepareForSleep() |
54c31a79 | 1372 | |
ccddd104 | 1373 | * remove any syslog support from log.c — we probably cannot do this before split-off udev is gone for good |
826872b6 | 1374 | |
3679d112 LP |
1375 | * shutdown logging: store to EFI var, and store to USB stick? |
1376 | ||
356ce991 | 1377 | * merge unit_kill_common() and unit_kill_context() |
490b7e47 | 1378 | |
07eabc2b LB |
1379 | * add a dependency on standard-conf.xml and other included files to man pages |
1380 | ||
1381 | * MountFlags=shared acts as MountFlags=slave right now. | |
1382 | ||
1383 | * properly handle loop back mounts via fstab, especially regards to fsck/passno | |
1384 | ||
1385 | * initialize the hostname from the fs label of /, if /etc/hostname does not exist? | |
1386 | ||
1387 | * sd-bus: | |
1388 | - EBADSLT handling | |
1389 | - GetAllProperties() on a non-existing object does not result in a failure currently | |
1390 | - port to sd-resolve for connecting to TCP dbus servers | |
1391 | - 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 | |
1392 | - see if we can drop more message validation on the sending side | |
1393 | - add API to clone sd_bus_message objects | |
1394 | - longer term: priority inheritance | |
1395 | - dbus spec updates: | |
1396 | - NameLost/NameAcquired obsolete | |
07eabc2b LB |
1397 | - path escaping |
1398 | - update systemd.special(7) to mention that dbus.socket is only about the compatibility socket now | |
1399 | ||
1400 | * sd-event | |
1401 | - allow multiple signal handlers per signal? | |
1402 | - document chaining of signal handler for SIGCHLD and child handlers | |
1403 | - define more intervals where we will shift wakeup intervals around in, 1h, 6h, 24h, ... | |
1404 | - maybe support iouring as backend, so that we allow hooking read and write | |
1405 | operations instead of IO ready events into event loops. See considerations | |
1406 | here: | |
1407 | http://blog.vmsplice.net/2020/07/rethinking-event-loop-integration-for.html | |
1408 | ||
1409 | * dbus: when a unit failed to load (i.e. is in UNIT_ERROR state), we | |
1410 | should be able to safely try another attempt when the bus call LoadUnit() is invoked. | |
1411 | ||
1412 | * maybe do not install getty@tty1.service symlink in /etc but in /usr? | |
1413 | ||
1414 | * print a nicer explanation if people use variable/specifier expansion in ExecStart= for the first word | |
1415 | ||
1416 | * mount: turn dependency information from /proc/self/mountinfo into dependency information between systemd units. | |
1417 | ||
1418 | * firstboot: allow provisioning of /etc/hosts entries, so that we can via the | |
1419 | credentials logic insert host name to resolve into containers/hosts. Usecase: | |
1420 | fork a container, and make it ping some specific address which is defined by | |
1421 | the host on invocation | |
1422 | ||
1423 | * systemd-firstboot: make sure to always use chase_symlinks() before | |
1424 | reading/writing files | |
1425 | ||
1426 | * 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 | |
1427 | ||
1428 | * sd-boot: define a drop-in dir in the ESP that may contain X.509 | |
1429 | certificates. If the firmware is detected to be in setup mode, automatically | |
1430 | enroll them as PK/KEK/db, turn off setup mode and proceed. Optionally, | |
1431 | instead of auto-enrolling them add them to the sd-boot menu, giving the user | |
1432 | the option to manually enroll them, after selecting the menu entry. This way, | |
1433 | installer images can just drop the certfiicates in the ESP, and on first boot | |
1434 | can easily enroll the keys without ever booting up. | |
1435 | ||
1436 | * efi stub: optionally, load initrd from disk as a separate file, HMAC check it | |
1437 | with key from TPM, bound to PCR, refusing if failing. This would then allow | |
1438 | traditional distros that generate initrds locally to secure them with TPM: | |
1439 | after generating the initrd, do the HMAC calculation, put result in initrd | |
1440 | filename, done. This would then bind the validity of the initrd to the local | |
1441 | host, and used kernel, and means people cannot change initrd or kernel | |
1442 | without booting the kernel + initrd. | |
1443 | ||
b44be3ec | 1444 | * EFI: |
b44be3ec LP |
1445 | - honor language efi variables for default language selection (if there are any?) |
1446 | - honor timezone efi variables for default timezone selection (if there are any?) | |
466784c8 | 1447 | - change bootctl to be backed by systemd-bootd to control temporary and persistent default boot goal plus efi variables |
631427d6 | 1448 | * bootctl |
631427d6 | 1449 | - recognize the case when not booted on EFI |
e4181484 | 1450 | |
07eabc2b | 1451 | * bootctl,sd-boot: actually honour the "architecture" key |
e9fd44b7 | 1452 | |
07eabc2b | 1453 | * bootctl: |
483091b0 | 1454 | - show whether UEFI audit mode is available |
07eabc2b LB |
1455 | - teach it to prepare an ESP wholesale, i.e. with mkfs.vfat invocation |
1456 | - teach it to copy in unified kernel images and maybe type #1 boot loader spec entries from host | |
1457 | - make it operate on loopback files, dissecting enough to find ESP to operate on | |
f620a368 LP |
1458 | - bootspec: properly support boot attempt counters when parsing entry file names |
1459 | ||
1460 | * kernel-install: | |
1461 | - optionally, support generating type #2 entries instead of type #1, including signing them | |
e6c6e7af | 1462 | |
b44be3ec LP |
1463 | * logind: |
1464 | - logind: optionally, ignore idle-hint logic for autosuspend, block suspend as long as a session is around | |
b44be3ec LP |
1465 | - logind: wakelock/opportunistic suspend support |
1466 | - Add pretty name for seats in logind | |
1467 | - logind: allow showing logout dialog from system? | |
f38afcd0 LP |
1468 | - add Suspend() bus calls which take timestamps to fix double suspend issues when somebody hits suspend and closes laptop quickly. |
1469 | - if pam_systemd is invoked by su from a process that is outside of a | |
1470 | any session we should probably just become a NOP, since that's | |
1471 | usually not a real user session but just some system code that just | |
1472 | needs setuid(). | |
279f0366 LP |
1473 | - logind: make the Suspend()/Hibernate() bus calls wait for the for |
1474 | the job to be completed. before returning, so that clients can wait | |
1475 | for "systemctl suspend" to finish to know when the suspending is | |
1476 | complete. | |
1477 | - logind: when the power button is pressed short, just popup a | |
1478 | logout dialog. If it is pressed for 1s, do the usual | |
1479 | shutdown. Inspiration are Macs here. | |
28423d9a | 1480 | - expose "Locked" property on logind session objects |
e25b5a8d DH |
1481 | - maybe allow configuration of the StopTimeout for session scopes |
1482 | - rename session scope so that it includes the UID. THat way | |
1483 | the session scope can be arranged freely in slices and we don't have | |
1484 | make assumptions about their slice anymore. | |
1485 | - follow PropertiesChanged state more closely, to deal with quick logouts and | |
1486 | relogins | |
77b19caf | 1487 | - (optionally?) spawn seat-manager@$SEAT.service whenever a seat shows up that as CanGraphical set |
4a5f779f ZJS |
1488 | - expose details of boot entries on the bus. In particular, it should be possible |
1489 | to query the list of boot entry titles that bootctl / sd-boot would show. | |
1490 | Currently we only expose their identifiers. | |
07eabc2b LB |
1491 | |
1492 | * move multiseat vid/pid matches from logind udev rule to hwdb | |
1493 | ||
1494 | * logind: rework pam_logind to also do a bus call in case of invocation from | |
1495 | user@.service, which returns the XDG_RUNTIME_DIR value, and make this | |
1496 | behaviour selectable via pam module option. | |
1497 | ||
1498 | * delay activation of logind until somebody logs in, or when /dev/tty0 pulls it | |
1499 | in or lingering is on (so that containers don't bother with it until PAM is used). also exit-on-idle | |
1500 | ||
b44be3ec | 1501 | * journal: |
57f2a947 | 1502 | - consider introducing implicit _TTY= + _PPID= + _EUID= + _EGID= + _FSUID= + _FSGID= fields |
b44be3ec LP |
1503 | - journald: also get thread ID from client, plus thread name |
1504 | - journal: when waiting for journal additions in the client always sleep at least 1s or so, in order to minimize wakeups | |
1505 | - add API to close/reopen/get fd for journal client fd in libsystemd-journal. | |
2aed63f4 | 1506 | - fall back to /dev/log based logging in libsystemd-journal, if we cannot log natively? |
b44be3ec | 1507 | - declare the local journal protocol stable in the wiki interface chart |
b44be3ec LP |
1508 | - sd-journal: speed up sd_journal_get_data() with transparent hash table in bg |
1509 | - journald: when dropping msgs due to ratelimit make sure to write | |
1510 | "dropped %u messages" not only when we are about to print the next | |
5238e957 | 1511 | message that works, but already after a short timeout |
b44be3ec LP |
1512 | - check if we can make journalctl by default use --follow mode inside of less if called without args? |
1513 | - maybe add API to send pairs of iovecs via sd_journal_send | |
f47ec8eb | 1514 | - journal: add a setgid "systemd-journal" utility to invoke from libsystemd-journal, which passes fds via STDOUT and does PK access |
b44be3ec LP |
1515 | - journactl: support negative filtering, i.e. FOOBAR!="waldo", |
1516 | and !FOOBAR for events without FOOBAR. | |
06847d0f | 1517 | - journal: store timestamp of journal_file_set_offline() in the header, |
038cf334 | 1518 | so it is possible to display when the file was last synced. |
b44be3ec LP |
1519 | - 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. |
1520 | - journal: find a way to allow dropping history early, based on priority, other rules | |
1521 | - journal: When used on NFS, check payload hashes | |
b44be3ec LP |
1522 | - journald: add kernel cmdline option to disable ratelimiting for debug purposes |
1523 | - refuse taking lower-case variable names in sd_journal_send() and friends. | |
1524 | - journald: we currently rotate only after MaxUse+MaxFilesize has been reached. | |
1525 | - journal: deal nicely with byte-by-byte copied files, especially regards header | |
b44be3ec | 1526 | - journal: sanely deal with entries which are larger than the individual file size, but where the components would fit |
601d9d6f | 1527 | - Replace utmp, wtmp, btmp, and lastlog completely with journal |
f38afcd0 | 1528 | - journalctl: instead --after-cursor= maybe have a --cursor=XYZ+1 syntax? |
f38afcd0 LP |
1529 | - when a kernel driver logs in a tight loop, we should ratelimit that too. |
1530 | - journald: optionally, log debug messages to /run but everything else to /var | |
1531 | - journald: when we drop syslog messages because the syslog socket is | |
1532 | full, make sure to write how many messages are lost as first thing | |
1533 | to syslog when it works again. | |
279f0366 LP |
1534 | - journald: allow per-priority and per-service retention times when rotating/vacuuming |
1535 | - journald: make use of uid-range.h to managed uid ranges to split | |
1536 | journals in. | |
1537 | - journalctl: add the ability to look for the most recent process of a binary. journalctl /usr/bin/X11 --pid=-1 or so... | |
1538 | - improve journalctl performance by loading journal files | |
1539 | lazily. Encode just enough information in the file name, so that we | |
1540 | do not have to open it to know that it is not interesting for us, for | |
1541 | the most common operations. | |
e25b5a8d | 1542 | - man: document that corrupted journal files is nothing to act on |
e25b5a8d DH |
1543 | - rework journald sigbus stuff to use mutex |
1544 | - Set RLIMIT_NPROC for systemd-journal-xyz, and all other of our | |
1545 | services that run under their own user ids, and use User= (but only | |
1546 | in a world where userns is ubiquitous since otherwise we cannot | |
1547 | invoke those daemons on the host AND in a container anymore). Also, | |
1548 | if LimitNPROC= is used without User= we should warn and refuse | |
1549 | operation. | |
1550 | - journalctl --verify: don't show files that are currently being | |
1551 | written to as FAIL, but instead show that their are being written to. | |
1552 | - add journalctl -H that talks via ssh to a remote peer and passes through | |
1553 | binary logs data | |
e25b5a8d | 1554 | - add a version of --merge which also merges /var/log/journal/remote |
e25b5a8d DH |
1555 | - journalctl: -m should access container journals directly by enumerating |
1556 | them via machined, and also watch containers coming and going. | |
1557 | Benefit: nspawn --ephemeral would start working nicely with the journal. | |
1558 | - assign MESSAGE_ID to log messages about failed services | |
06847d0f | 1559 | - check if loop in decompress_blob_xz() is necessary |
b44be3ec | 1560 | |
07eabc2b LB |
1561 | * journald: support RFC3164 fully for the incoming syslog transport, see |
1562 | https://github.com/systemd/systemd/issues/19251#issuecomment-816601955 | |
1563 | ||
1564 | * Hook up journald's FSS logic with TPM2: seal the verification disk by | |
1565 | time-based policy, so that the verification key can remain on host and ve | |
1566 | validated via TPM. | |
1567 | ||
1568 | * build short web pages out of each catalog entry, build them along with man | |
1569 | pages, and include hyperlinks to them in the journal output | |
1570 | ||
1571 | * journald: do journal file writing out-of-process, with one writer process per | |
1572 | client UID, so that synthetic hash table collisions can slow down a specific | |
1573 | user's journal stream down but not the others. | |
1574 | ||
1575 | * tweak journald context caching. In addition to caching per-process attributes | |
1576 | keyed by PID, cache per-cgroup attributes (i.e. the various xattrs we read) | |
1577 | keyed by cgroup path, and guarded by ctime changes. This should provide us | |
1578 | with a nice speed-up on services that have many processes running in the same | |
1579 | cgroup. | |
1580 | ||
1581 | * maybe add call sd_journal_set_block_timeout() or so to set SO_SNDTIMEO for | |
1582 | the sd-journal logging socket, and, if the timeout is set to 0, sets | |
1583 | O_NONBLOCK on it. That way people can control if and when to block for | |
1584 | logging. | |
1585 | ||
1586 | * journalctl: make sure -f ends when the container indicated by -M terminates | |
1587 | ||
1588 | * journald: sigbus API via a signal-handler safe function that people may call | |
1589 | from the SIGBUS handler | |
1590 | ||
590171d1 ZJS |
1591 | * add a test if all entries in the catalog are properly formatted. |
1592 | (Adding dashes in a catalog entry currently results in the catalog entry | |
1593 | being silently skipped. journalctl --update-catalog must warn about this, | |
1594 | and we should also have a unit test to check that all our message are OK.) | |
1595 | ||
07eabc2b LB |
1596 | * homed: |
1597 | - when user tries to log into record signed by unrecognized key, automatically add key to our chain after polkit auth | |
1598 | - rollback when resize fails mid-operation | |
1599 | - GNOME's side for forget key on suspend (requires rework so that lock screen runs outside of uid) | |
07eabc2b LB |
1600 | - update LUKS password on login if we find there's a password that unlocks the JSON record but not the LUKS device. |
1601 | - create on activate? | |
1602 | - properties: icon url?, preferred session type?, administrator bool (which translates to 'wheel' membership)?, address?, telephone?, vcard?, samba stuff?, parental controls? | |
1603 | - communicate clearly when usb stick is safe to remove. probably involves | |
1604 | beefing up logind to make pam session close hook synchronous and wait until | |
1605 | systemd --user is shut down. | |
1606 | - logind: maybe keep a "busy fd" as long as there's a non-released session around or the user@.service | |
1607 | - maybe make automatic, read-only, time-based reflink-copies of LUKS disk | |
1608 | images (and btrfs snapshots of subvolumes) (think: time machine) | |
1609 | - distinguish destroy / remove (i.e. currently we can unregister a user, unregister+remove their home directory, but not just remove their home directory) | |
1610 | - in systemd's PAMName= logic: query passwords with ssh-askpassword, so that we can make "loginctl set-linger" mode work | |
1611 | - fingerprint authentication, pattern authentication, … | |
1612 | - make sure "classic" user records can also be managed by homed | |
1613 | - make size of $XDG_RUNTIME_DIR configurable in user record | |
1614 | - query password from kernel keyring first | |
1615 | - update even if record is "absent" | |
07eabc2b LB |
1616 | - move acct mgmt stuff from pam_systemd_home to pam_systemd? |
1617 | - when "homectl --pkcs11-token-uri=" is used, synthesize ssh-authorized-keys records for all keys we have private keys on the stick for | |
1618 | - make slice for users configurable (requires logind rework) | |
1619 | - logind: populate auto-login list bus property from PKCS#11 token | |
1620 | - when determining state of a LUKS home directory, check DM suspended sysfs file | |
a11e7c0b LB |
1621 | - when homed is in use, maybe start the user session manager in a mount namespace with MS_SLAVE, |
1622 | so that mounts propagate down but not up - eg, user A setting up a backup volume | |
1623 | doesn't mean user B sees it | |
9c53de8b | 1624 | - use credentials logic/TPM2 logic to store homed signing key |
9c53de8b LP |
1625 | - permit multiple user record signing keys to be used locally, and pick |
1626 | the right one for signing records automatically depending on a pre-existing | |
1627 | signature | |
1628 | - add a way to "adopt" a home directory, i.e. strip foreign signatures | |
1629 | and insert a local signature instead. | |
1630 | - as an extension to the directory+subvolume backend: if located on | |
1631 | especially marked fs, then sync down password into LUKS header of that fs, | |
1632 | and always verify passwords against it too. Bootstrapping is a problem | |
1633 | though: if no one is logged in (or no other user even exists yet), how do you | |
1634 | unlock the volume in order to create the first user and add the first pw. | |
1635 | - support new FS_IOC_ADD_ENCRYPTION_KEY ioctl for setting up fscrypt | |
1636 | - maybe pre-create ~/.cache as subvol so that it can have separate quota | |
1637 | easily? | |
9c53de8b LP |
1638 | - add a switch to homectl (maybe called --first-boot) where it will check if |
1639 | any non-system users exist, and if not prompts interactively for basic user | |
1640 | info, mimicking systemd-firstboot. Then, place this in a service that runs | |
1641 | after systemd-homed, but before gdm and friends, as a simple, barebones | |
1642 | fallback logic to get a regular user created on uninitialized systems. | |
1643 | - store PKCS#11 + FIDO2 token info in LUKS2 header, compatible with | |
1644 | systemd-cryptsetup, so that it can unlock homed volumes | |
6d975fe7 LP |
1645 | - maybe make all *.home files owned by `systemd-home` user or so, so that we |
1646 | can easily set overall quota for all users | |
1647 | - on login, if we can't fallocate initially, but rebalance is on, then allow | |
1648 | login in discard mode, then immediately rebalance, then turn off discard | |
f1a147f2 | 1649 | - extend user records with optional "bulk" data. Specifically, a user |
288bd406 | 1650 | avatar/photo or so. This data should be stored along with the user record, |
f1a147f2 LP |
1651 | but probably shouldn't be part of the record itself, since it might be |
1652 | large. | |
07eabc2b | 1653 | |
07eabc2b LB |
1654 | * add a new switch --auto-definitions=yes/no or so to systemd-repart. If |
1655 | specified, synthesize a definition automatically if we can: enlarge last | |
1656 | partition on disk, but only if it is marked for growing and not read-only. | |
1657 | ||
2df2bb1f | 1658 | * systemd-repart: read LUKS encryption key from $CREDENTIALS_DIRECTORY |
07eabc2b LB |
1659 | |
1660 | * systemd-repart: add a switch to factory reset the partition table without | |
1661 | immediately applying the new configuration again. i.e. --factory-reset=leave | |
1662 | or so. (this is useful to factory reset an image, then putting it into | |
1663 | another machine, ensuring that luks key is generated on new machine, not old) | |
1664 | ||
1665 | * systemd-repart: support setting up dm-integrity with HMAC | |
1666 | ||
1667 | * systemd-repart: maybe remove half-initialized image on failure. It fails | |
1668 | if the output file exists, so a repeated invocation will usually fail if | |
1669 | something goes wrong on the way. | |
1670 | ||
1671 | * systemd-repart: drop pager mode on normal operation? | |
1672 | ||
1673 | * systemd-repart: by default generate minimized partition tables (i.e. tables | |
1674 | that only cover the space actually used, excluding any free space at the | |
1675 | end), in order to maximize dd'ability. Requires libfdisk work, see | |
1676 | https://github.com/karelzak/util-linux/issues/907 | |
1677 | ||
1678 | * systemd-repart: MBR partition table support. Care needs to be taken regarding | |
1679 | Type=, so that partition definitions can sanely apply to both the GPT and the | |
1680 | MBR case. Idea: accept syntax "Type=gpt:home mbr:0x83" for setting the types | |
1681 | for the two partition types explicitly. And provide an internal mapping so | |
1682 | that "Type=linux-generic" maps to the right types for both partition tables | |
1683 | automatically. | |
1684 | ||
1685 | * systemd-repart: allow sizing partitions as factor of available RAM, so that | |
1686 | we can reasonably size swap partitions for hibernation. | |
1687 | ||
1688 | * systemd-repart: allow boolean option that ensures that if existing partition | |
1689 | doesn't exist within the configured size bounds the whole command fails. This | |
1690 | is useful to implement ESP vs. XBOOTLDR schemes in installers: have one set | |
1691 | of repart files for the case where ESP is large enough and one where it isn't | |
1692 | and XBOOTLDR is added in instead. Then apply the former first, and if it | |
1693 | fails to apply use the latter. | |
1694 | ||
1695 | * systemd-repart: add per-partition option to never reuse existing partition | |
1696 | and always create anew even if matching partition already exists. | |
1697 | ||
1698 | * systemd-repart: add per-partition option to fail if partition already exist, | |
1699 | i.e. is not added new. Similar, add option to fail if partition does not exist yet. | |
1700 | ||
1701 | * systemd-repart: allow disabling growing of specific partitions, or making | |
1702 | them (think ESP: we don't ever want to grow it, since we cannot resize vfat) | |
1703 | ||
1704 | * systemd-repart: make it a static checker during early boot for existence and | |
1705 | absence of other partitions for trusted boot environments | |
1706 | ||
b44be3ec | 1707 | * document: |
8b8f2591 | 1708 | - document that deps in [Unit] sections ignore Alias= fields in |
b44be3ec LP |
1709 | [Install] units of other units, unless those units are disabled |
1710 | - man: clarify that time-sync.target is not only sysv compat but also useful otherwise. Same for similar targets | |
b44be3ec | 1711 | - document that service reload may be implemented as service reexec |
f38afcd0 LP |
1712 | - add a man page containing packaging guidelines and recommending usage of things like Documentation=, PrivateTmp=, PrivateNetwork= and ReadOnlyDirectories=/etc /usr. |
1713 | - document systemd-journal-flush.service properly | |
f38afcd0 LP |
1714 | - documentation: recommend to connect the timer units of a service to the service via Also= in [Install] |
1715 | - man: document the very specific env the shutdown drop-in tools live in | |
5cf821ac ZJS |
1716 | - man: add more examples to man pages, |
1717 | - in particular an example how to do the equivalent of switching runlevels | |
f38afcd0 | 1718 | - man: maybe sort directives in man pages, and take sections from --help and apply them to man too |
17ec531f | 1719 | - document root=gpt-auto properly |
b44be3ec LP |
1720 | |
1721 | * systemctl: | |
b44be3ec LP |
1722 | - add systemctl switch to dump transaction without executing it |
1723 | - Add a verbose mode to "systemctl start" and friends that explains what is being done or not done | |
1724 | - "systemctl disable" on a static unit prints no message and does | |
1725 | nothing. "systemctl enable" does nothing, and gives a bad message | |
1726 | about it. Should fix both to print nice actionable messages. | |
1727 | - print nice message from systemctl --failed if there are no entries shown, and hook that into ExecStartPre of rescue.service/emergency.service | |
1728 | - add new command to systemctl: "systemctl system-reexec" which reexecs as many daemons as virtually possible | |
d28315e4 | 1729 | - systemctl enable: fail if target to alias into does not exist? maybe show how many units are enabled afterwards? |
b44be3ec | 1730 | - systemctl: "Journal has been rotated since unit was started." message is misleading |
61233823 | 1731 | - systemctl status output should include list of triggering units and their status |
f38afcd0 | 1732 | |
07eabc2b LB |
1733 | * introduce an option (or replacement) for "systemctl show" that outputs all |
1734 | properties as JSON, similar to busctl's new JSON output. In contrast to that | |
1735 | it should skip the variant type string though. | |
8b04b925 | 1736 | |
07eabc2b LB |
1737 | * add an explicit "vertical" mode to format-table, so that "systemctl |
1738 | status"-like outputs (i.e. with a series of field names left and values | |
1739 | right) become genuine first class citizens, and we gain automatic, sane JSON | |
1740 | output for them. | |
d2e83c23 | 1741 | |
07eabc2b LB |
1742 | * Add a "systemctl list-units --by-slice" mode or so, which rearranges the |
1743 | output of "systemctl list-units" slightly by showing the tree structure of | |
1744 | the slices, and the units attached to them. | |
a19554ed | 1745 | |
07eabc2b LB |
1746 | * add "systemctl wait" or so, which does what "systemd-run --wait" does, but |
1747 | for all units. It should be both a way to pin units into memory as well as a | |
1748 | wait to retrieve their exit data. | |
a7a3f28b | 1749 | |
07eabc2b LB |
1750 | * show whether a service has out-of-date configuration in "systemctl status" by |
1751 | using mtime data of ConfigurationDirectory=. | |
08f95888 | 1752 | |
07eabc2b LB |
1753 | * "systemctl preset-all" should probably order the unit files it |
1754 | operates on lexicographically before starting to work, in order to | |
1755 | ensure deterministic behaviour if two unit files conflict (like DMs | |
1756 | do, for example) | |
dcfc4b2e | 1757 | |
07eabc2b LB |
1758 | * add "systemctl start -v foobar.service" that shows logs of a service |
1759 | while the start command runs. This is non-trivial to do without | |
1760 | races though, since we should flush out all journal messages before | |
1761 | returning from the "systemctl stop". | |
71ef24d0 | 1762 | |
07eabc2b LB |
1763 | * systemctl: if some operation fails, show log output? |
1764 | ||
1765 | * Add a new verb "systemctl top" | |
1766 | ||
1767 | * unit install: | |
1768 | - "systemctl mask" should find all names by which a unit is accessible | |
1769 | (i.e. by scanning for symlinks to it) and link them all to /dev/null | |
1b89884b | 1770 | |
b44be3ec | 1771 | * nspawn: |
e25b5a8d DH |
1772 | - emulate /dev/kmsg using CUSE and turn off the syslog syscall |
1773 | with seccomp. That should provide us with a useful log buffer that | |
1774 | systemd can log to during early boot, and disconnect container logs | |
1775 | from the kernel's logs. | |
1776 | - as soon as networkd has a bus interface, hook up --network-interface=, | |
1777 | --network-bridge= with networkd, to trigger netdev creation should an | |
1778 | interface be missing | |
e25b5a8d DH |
1779 | - a nice way to boot up without machine id set, so that it is set at boot |
1780 | automatically for supporting --ephemeral. Maybe hash the host machine id | |
1781 | together with the machine name to generate the machine id for the container | |
1782 | - fix logic always print a final newline on output. | |
1783 | https://github.com/systemd/systemd/pull/272#issuecomment-113153176 | |
1784 | - should optionally support receiving WATCHDOG=1 messages from its payload | |
1785 | PID 1... | |
e25b5a8d DH |
1786 | - optionally automatically add FORWARD rules to iptables whenever nspawn is |
1787 | running, remove them when shut down. | |
e25b5a8d | 1788 | |
07eabc2b LB |
1789 | * nspawn: add support for sysext extensions, too. i.e. a new --extension= |
1790 | switch that takes one or more arguments, and applies the extensions already | |
1791 | during startup. | |
1792 | ||
1793 | * when main nspawn supervisor process gets suspended due to SIGSTOP/SIGTTOU or | |
1794 | so, freeze the payload too. | |
1795 | ||
1796 | * machined: add API to acquire UID range. add API to mount/dissect loopback | |
1797 | file. Both protected by PK. Then make nspawn use these APIs to run | |
1798 | unprivileged containers. i.e. push the truly privileged bits into machined, | |
1799 | so that the client side can remain entirely unprivileged, with SUID or | |
1800 | anything like that. | |
1801 | ||
1802 | * nspawn: support time namespaces | |
1803 | ||
1804 | * nspawn: on cgroupsv1 issue cgroup empty handler process based on host events, | |
1805 | so that we make cgroup agent logic safe | |
1806 | ||
1807 | * nspawn/machined: add API to invoke binary in container, then use that as | |
1808 | fallback in "machinectl shell" | |
1809 | ||
1810 | * nspawn: make nspawn suitable for shell pipelines: instead of triggering a | |
1811 | hangup when input is finished, send ^D, which synthesizes an EOF. Then wait | |
1812 | for hangup or ^D before passing on the EOF. | |
1813 | ||
1814 | * nspawn: greater control over selinux label? | |
1815 | ||
1816 | * nspawn: support that /proc, /sys/, /dev are pre-mounted | |
1817 | ||
e25b5a8d | 1818 | * machined: |
e25b5a8d DH |
1819 | - add an API so that libvirt-lxc can inform us about network interfaces being |
1820 | removed or added to an existing machine | |
1821 | - "machinectl migrate" or similar to copy a container from or to a | |
1822 | difference host, via ssh | |
e25b5a8d DH |
1823 | - introduce systemd-nspawn-ephemeral@.service, and hook it into |
1824 | "machinectl start" with a new --ephemeral switch | |
1825 | - "machinectl status" should also show internal logs of the container in | |
1826 | question | |
e25b5a8d DH |
1827 | - "machinectl history" |
1828 | - "machinectl diff" | |
1829 | - "machinectl commit" that takes a writable snapshot of a tree, invokes a | |
1830 | shell in it, and marks it read-only after use | |
1831 | ||
abd55b16 | 1832 | * udev: |
abd55b16 | 1833 | - move to LGPL |
abd55b16 KS |
1834 | - kill scsi_id |
1835 | - add trigger --subsystem-match=usb/usb_device device | |
e8d842a0 | 1836 | - reimport udev db after MOVE events for devices without dev_t |
b8217b7b | 1837 | |
e25b5a8d DH |
1838 | * coredump: |
1839 | - save coredump in Windows/Mozilla minidump format | |
73a99163 | 1840 | - when truncating coredumps, also log the full size that the process had, and make a metadata field so we can report truncated coredumps |
87a8baa3 LP |
1841 | |
1842 | * support crash reporting operation modes (https://live.gnome.org/GnomeOS/Design/Whiteboards/ProblemReporting) | |
1843 | ||
f38afcd0 | 1844 | * tmpfiles: |
f38afcd0 | 1845 | - apply "x" on "D" too (see patch from William Douglas) |
614cc34f | 1846 | - instead of ignoring unknown fields, reject them. |
e25b5a8d DH |
1847 | - creating new directories/subvolumes/fifos/device nodes |
1848 | should not follow symlinks. None of the other adjustment or creation | |
1849 | calls follow symlinks. | |
36f57e02 | 1850 | - add --test mode |
ba405b22 ZJS |
1851 | - teach tmpfiles.d q/Q logic something sensible in the context of XFS/ext4 |
1852 | project quota | |
1258097c | 1853 | |
af6f0d42 TG |
1854 | * udev-link-config: |
1855 | - Make sure ID_PATH is always exported and complete for | |
1856 | network devices where possible, so we can safely rely | |
1857 | on Path= matching | |
af6f0d42 | 1858 | |
88e4d1d7 | 1859 | * sd-rtnl: |
88e4d1d7 | 1860 | - add support for more attribute types |
c589a0e6 | 1861 | - inbuilt piping support (essentially degenerate async)? see loopback-setup.c and other places |
88e4d1d7 | 1862 | |
0a4b9a07 | 1863 | * networkd: |
c74ecd71 TG |
1864 | - add more keys to [Route] and [Address] sections |
1865 | - add support for more DHCPv4 options (and, longer term, other kinds of dynamic config) | |
e8d842a0 | 1866 | - add reduced [Link] support to .network files |
798e174a | 1867 | - properly handle routerless dhcp leases |
a8eaaee7 | 1868 | - work with non-Ethernet devices |
e25b5a8d | 1869 | - dhcp: do we allow configuring dhcp routes on interfaces that are not the one we got the dhcp info from? |
e25b5a8d DH |
1870 | - the DHCP lease data (such as NTP/DNS) is still made available when |
1871 | a carrier is lost on a link. It should be removed instantly. | |
1872 | - expose in the API the following bits: | |
e2da6491 | 1873 | - option 15, domain name |
38b38500 | 1874 | - option 12, hostname and/or option 81, fqdn |
e25b5a8d DH |
1875 | - option 123, 144, geolocation |
1876 | - option 252, configure http proxy (PAC/wpad) | |
1877 | - provide a way to define a per-network interface default metric value | |
1878 | for all routes to it. possibly a second default for DHCP routes. | |
1879 | - allow Name= to be specified repeatedly in the [Match] section. Maybe also | |
1880 | support Name=foo*|bar*|baz ? | |
e25b5a8d | 1881 | - whenever uplink info changes, make DHCP server send out FORCERENEW |
155e8b9a | 1882 | |
07eabc2b LB |
1883 | * in networkd, when matching device types, fix up DEVTYPE rubbish the kernel passes to us |
1884 | ||
d5e172d2 ZJS |
1885 | * Figure out how to do unittests of networkd's state serialization |
1886 | ||
ac976532 | 1887 | * dhcp: |
424a8732 | 1888 | - figure out how much we can increase Maximum Message Size |
ac976532 | 1889 | |
37d8b536 PF |
1890 | * dhcp6: |
1891 | - add functions to set previously stored IPv6 addresses on startup and get | |
1892 | them at shutdown; store them in client->ia_na | |
1893 | - write more test cases | |
37d8b536 | 1894 | - implement reconfigure support, see 5.3., 15.11. and 22.20. |
37d8b536 | 1895 | - implement support for temporary adressess (IA_TA) |
37d8b536 PF |
1896 | - implement dhcpv6 authentication |
1897 | - investigate the usefulness of Confirm messages; i.e. are there any | |
1898 | situations where the link changes without any loss in carrier detection | |
1899 | or interface down | |
1900 | - some servers don't do rapid commit without a filled in IA_NA, verify | |
1901 | this behavior | |
4a77c53d | 1902 | - RouteTable= ? |