]>
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 | ||
ceaf24d4 ZJS |
7 | * userdbctl: "Password OK: yes" is shown even when there are no passwords |
8 | or the password is locked. | |
9 | ||
5b326dee ZJS |
10 | * Jun 01 09:43:02 krowka systemd[1]: Unit user@1000.service has alias user@.service. |
11 | Jun 01 09:43:02 krowka systemd[1]: Unit user@6.service has alias user@.service. | |
12 | Jun 01 09:43:02 krowka systemd[1]: Unit user-runtime-dir@6.service has alias user-runtime-dir@.service. | |
13 | ||
f38afcd0 | 14 | External: |
f85857df | 15 | |
f38afcd0 | 16 | * Fedora: add an rpmlint check that verifies that all unit files in the RPM are listed in %systemd_post macros. |
bafb15ba | 17 | |
07eabc2b LB |
18 | * dbus: |
19 | - natively watch for dbus-*.service symlinks (PENDING) | |
20 | - teach dbus to activate all services it finds in /etc/systemd/services/org-*.service | |
21 | ||
22 | * kernel: add device_type = "fb", "fbcon" to class "graphics" | |
23 | ||
24 | * /usr/bin/service should actually show the new command line | |
25 | ||
26 | * 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= | |
27 | ||
28 | * neither pkexec nor sudo initialize environ[] from the PAM environment? | |
29 | ||
30 | * fedora: update policy to declare access mode and ownership of unit files to root:root 0644, and add an rpmlint check for it | |
31 | ||
32 | * register catalog database signature as file magic | |
33 | ||
34 | * zsh shell completion: | |
35 | - <command> <verb> -<TAB> should complete options, but currently does not | |
36 | - systemctl add-wants,add-requires | |
37 | ||
38 | * systemctl status should know about 'systemd-analyze calendar ... --iterations=' | |
39 | * If timer has just OnInactiveSec=..., it should fire after a specified time | |
40 | after being started. | |
41 | ||
42 | * write blog stories about: | |
43 | - hwdb: what belongs into it, lsusb | |
44 | - enabling dbus services | |
45 | - how to make changes to sysctl and sysfs attributes | |
46 | - remote access | |
47 | - how to pass throw-away units to systemd, or dynamically change properties of existing units | |
48 | - testing with Harald's awesome test kit | |
49 | - auto-restart | |
50 | - how to develop against journal browsing APIs | |
51 | - the journal HTTP iface | |
52 | - non-cgroup resource management | |
53 | - dynamic resource management with cgroups | |
54 | - refreshed, longer missions statement | |
55 | - calendar time events | |
56 | - init=/bin/sh vs. "emergency" mode, vs. "rescue" mode, vs. "multi-user" mode, vs. "graphical" mode, and the debug shell | |
57 | - how to create your own target | |
58 | - instantiated apache, dovecot and so on | |
59 | - hooking a script into various stages of shutdown/rearly booot | |
60 | ||
61 | Regularly: | |
62 | ||
63 | * look for close() vs. close_nointr() vs. close_nointr_nofail() | |
64 | ||
65 | * check for strerror(r) instead of strerror(-r) | |
66 | ||
67 | * pahole | |
68 | ||
69 | * set_put(), hashmap_put() return values check. i.e. == 0 does not free()! | |
70 | ||
71 | * use secure_getenv() instead of getenv() where appropriate | |
72 | ||
73 | * link up selected blog stories from man pages and unit files Documentation= fields | |
74 | ||
5e524b40 LP |
75 | Janitorial Clean-ups: |
76 | ||
5e524b40 LP |
77 | * Rearrange tests so that the various test-xyz.c match a specific src/basic/xyz.c again |
78 | ||
bb527e11 LP |
79 | * rework mount.c and swap.c to follow proper state enumeration/deserialization |
80 | semantics, like we do for device.c now | |
81 | ||
c6407108 LP |
82 | Features: |
83 | ||
e6f48be8 LP |
84 | * add linker script that implicitly adds symbol for build ID and new coredump |
85 | json package metadata, and use that when logging | |
86 | ||
f08e143c LP |
87 | * systemd-dissect: show GPT disk UUID in output |
88 | ||
89 | * Enable RestricFileSystems= for all our long-running services (similar: | |
90 | RestrictNetworkInterfaces=) | |
91 | ||
92 | * Add systemd-analyze security checks for RestrictFileSystems= and | |
93 | RestrictNetworkInterfaces= | |
94 | ||
1ccfb792 LP |
95 | * cryptsetup/homed: implement TOTP authentication backed by TPM2 and its |
96 | internal clock. | |
97 | ||
1ccfb792 | 98 | * nspawn: optionally set up nftables/iptables routes that forward UDP/TCP |
f08e143c | 99 | traffic on port 53 to resolved stub 127.0.0.54 |
e0c311b1 | 100 | |
f5ba8115 LP |
101 | * man: rework os-release(5), and clearly separate our extension-release.d/ and |
102 | initrd-release parts, i.e. list explicitly which fields are about what. | |
103 | ||
104 | * sysext: before applying a sysext, do a superficial validation run so that | |
105 | things are not rearranged to wildy. I.e. protect against accidental fuckups, | |
106 | such as masking out /usr/lib/ or so. We should probably refuse if existing | |
107 | inodes are replaced by other types of inodes or so. | |
108 | ||
109 | * sysext: ensure one can build a sysext that can safely apply to *any* system | |
110 | (because it contains only static go binaries in /opt/ or so) | |
111 | ||
af11e0ef LP |
112 | * userdb: when synthesizing NSS records, pick "best" password from defined |
113 | passwords, not just the first. i.e. if there are multiple defined, prefer | |
114 | unlocked over locked and prefer non-empty over empty. | |
115 | ||
116 | * maybe add a tool inspired by the GPT auto discovery spec that runs in the | |
117 | initrd and rearranges the rootfs hierarchy via bind mounts, if | |
118 | enabled. Specifically in some top-level dir /@auto/ it will look for | |
119 | dirs/symlinks/subvolumes that are named after their purpose, and optionally | |
120 | encode a version as well as assessment counters, and then mount them into the | |
121 | file system tree to boot into, similar to how we do that for the gpt auto | |
122 | logic. Maybe then bind mount the original root into /.superior or something | |
123 | like that (so that update tools can look there). Further discussion in this | |
124 | thread: | |
125 | https://lists.freedesktop.org/archives/systemd-devel/2021-November/047059.html | |
126 | The GPT dissection logic should automatically enable this tool whenever we | |
127 | detect a specially marked root fs (i.e introduce a new generic root gpt type | |
128 | for this, that is arch independent). The also implement this in the image | |
129 | dissection logic, so that nspawn/RootImage= and so on grok it. Maybe make | |
130 | generic enough so that it can also work for ostrees arrangements. | |
131 | ||
132 | * if a path ending in ".auto.d/" is set for RootDirectory=/RootImage= then do a | |
133 | strverscmp() of everything inside that dir and use that. i.e. implement very | |
134 | simple version control. Also use this in systemd-nspawn --image= and so on. | |
135 | ||
136 | * homed: while a home dir is not activated generate slightly different NSS | |
137 | records for it, that reports the home dir as "/" and the shell as some binary | |
138 | provided by us. Then, when an SSH login happens and SSH permits it our binary | |
139 | is invoked. This binary can then talk to homed and activate the homedir if | |
140 | it's not around yet, prompting the user for a password. Once that succeeded | |
141 | we'll switch to the real user record, i.e. home dir and shell, and our tool | |
142 | exec()s the latter. Net effect: ssh'ing into a homed account will just work: | |
143 | we'll neatly prompt for the homedir's password if its needed. –– Building on | |
144 | this we could take this even further: since this tool will potentially have | |
145 | access to the client's ssh-agent (if ssh-agent forwarding is enabled) we | |
146 | could implement SSH unlocking of a homedir with that: when enrolling a new | |
147 | ssh pubkey in a user record we'd ask the ssh-agent to sign some random value | |
148 | with the privkey, then use that as luks key to unlock the home dir. Will not | |
149 | work for ECDSA keys since their signatures contain a random component, but | |
150 | will work for RSA and Ed25519 keys. | |
151 | ||
bb5464ad LP |
152 | * add tiny service that decrypts encrypted user records passed via initrd |
153 | credential logic and drops them into /run where nss-systemd can pick them up, | |
154 | similar to /run/host/userdb/. Usecase: drop a root user JSON record there, | |
155 | and use it in the initrd to log in as root with locally selected password, | |
156 | for debugging purposes. | |
157 | ||
158 | * drop dependency on libcap, replace by direct syscalls based on | |
159 | CapabilityQuintet we already have. (This likely allows us drop drop libcap | |
160 | dep in the base OS image) | |
161 | ||
162 | * sysext: automatically activate sysext images dropped in via new sd-stub | |
163 | sysext pickup logic. | |
164 | ||
165 | * add concept for "exitrd" as inverse of "initrd", that we can transition to at | |
166 | shutdown, and has similar security semantics. This should then take the place | |
167 | of dracut's shutdown logic. Should probably support sysexts too. Care needs | |
168 | to be taken that the resulting logic ends up in RAM, i.e. is copied out of | |
169 | on-disk storage. | |
170 | ||
171 | * sd-stub: automatically pick up microcode from ESP and synthesize initrd from | |
172 | it, and measure it. Signing is not necessary, as microcode does that on its | |
173 | own. Pass as first initrd to kernel. | |
174 | ||
a07ab1dd LP |
175 | * userdbd: implement an additional varlink service socket that provides the |
176 | host user db in restricted form, then allow this to be bind mounted into | |
177 | sandboxed environments that want the host database in minimal form. All | |
178 | records would be stripped of all meta info, except the basic UID/name | |
179 | info. Then use this in portabled environments that do not use PrivateUsers=1. | |
180 | ||
a5bf435e LP |
181 | * logind introduce two types of sessions: "heavy" and "light". The former would |
182 | be our current sessions. But the latter would be a new type of session that | |
183 | is mostly the same but does not pull in user@.service or wait for it. Then, | |
184 | allow configuration which type of session is desired via pam_systemd | |
185 | parameters, and then make user@.service's session one of these "light" ones. | |
186 | People could then choose to make FTP sessions and suchlike "light" if they | |
187 | don't want the service manager to be started for that. | |
188 | ||
da3ab57c LP |
189 | * /etc/veritytab: allow that the roothash column can be specified as fs path |
190 | including a path to an AF_UNIX path, similar to how we do things with the | |
191 | keys of /etc/crypttab. That way people can store/provide the roothash | |
192 | externally and provide to us on demand only. | |
193 | ||
636c8a1f LP |
194 | * add high-level lockdown level for GPT dissection logic: e.g. an enum that can |
195 | be ANY (to mount anything), TRUSTED (to require that /usr is on signed | |
196 | verity, but rest doesn't matter), LOCKEDDOWN (to require that everything is | |
197 | on signed verity, except for ESP), SUPERLOCKDOWN (like LOCKEDDOWN but ESP not | |
198 | allowed). And then maybe some flavours of that that declare what is expected | |
199 | from home/srv/var… Then, add a new cmdline flag to all tools that parse such | |
200 | images, to configure this. Also, add a kernel cmdline option for this, to be | |
201 | honoured by the gpt auto generator. | |
202 | ||
203 | * nspawn: maybe optionally insert .nspawn file as GPT partition into images, so | |
204 | that such container images are entirely stand-alone and can be updated as | |
205 | one. | |
206 | ||
3fc0688d LP |
207 | * we probably should extend the root verity hash of the root fs into some PCR |
208 | on boot. (i.e. maybe add a crypttab option tpm2-measure=8 or so to measure it | |
209 | into PCR 8) | |
210 | ||
211 | * add a "policy" to the dissection logic. i.e. a bit mask what is OK to mount, | |
212 | what must be read-only, what requires encryption, and what requires | |
213 | authentication. | |
214 | ||
f4529c4d LP |
215 | * in uefi stub: query firmware regarding which PCRs are being used, store that |
216 | in EFI var. then use this when enrolling TPM2 in cryptsetup to verify that | |
217 | the selected PCRs actually are used by firmware. | |
218 | ||
322b3b38 LP |
219 | * rework recursive read-only remount to use new mount API |
220 | ||
9c53de8b LP |
221 | * PAM: pick auf one authentication token from credentials |
222 | ||
9a6549f6 LP |
223 | * tpm2: figure out if we need to do anything for TPM2 parameter encryption? And |
224 | if so, what precisely? | |
225 | ||
9a6549f6 LP |
226 | * when mounting disk images: if IMAGE_ID/IMAGE_VERSION is set in os-release |
227 | data in the image, make sure the image filename actually matches this, so | |
228 | that images cannot be misused. | |
229 | ||
aca8ecc3 LP |
230 | * New udev block device symlink names: |
231 | /dev/disk/by-parttypelabel/<pttype>/<ptlabel>. Use case: if pt label is used | |
232 | as partition image version string, this is a safe way to reference a specific | |
233 | version of a specific partition type, in particular where related partitions | |
234 | are processed (e.g. verity + rootfs both named "LennartOS_0.7"). | |
235 | ||
cf2ab2e7 LP |
236 | * in sd-id128: also parse UUIDs in RFC4122 URN syntax (i.e. chop off urn:uuid: prefix) |
237 | ||
31892e8d LP |
238 | * DynamicUser= + StateDirectory= → use uid mapping mounts, too, in order to |
239 | make dirs appear under right UID. | |
240 | ||
f3e58b55 LP |
241 | * systemd-sysext: optionally, run it in initrd already, before transitioning |
242 | into host, to open up possibility for services shipped like that. | |
243 | ||
49bd547b LP |
244 | * maybe add a tool that displays most recent journal logs as QR code to scan |
245 | off screen and run it automatically on boot failures, emergency logs and | |
246 | such. Use DRM APIs directly, see | |
247 | https://github.com/dvdhrm/docs/blob/master/drm-howto/modeset.c for an example | |
248 | for doing that. | |
249 | ||
250 | * pass systemd-detect-virt result to generators as env var. Modifying behaviour | |
251 | based on whether we are virtualized or not is a pretty common thing, hence | |
252 | maybe just pass that info along for free in an env var. We cache the result | |
253 | anyway, so it's basically free. | |
254 | ||
24063ba1 | 255 | * introduce /dev/disk/root/* symlinks that allow referencing partitions on the |
f3e58b55 LP |
256 | disk the rootfs is on in a reasonably secure way. (or maybe: add |
257 | /dev/gpt-auto-{home,srv,boot,…} similar in style to /dev/gpt-auto-root as we | |
258 | already have it. | |
24063ba1 | 259 | |
7ed72cfa LP |
260 | * whenever we receive fds via SCM_RIGHTS make sure none got dropped due to the |
261 | reception limit the kernel silently enforces. | |
262 | ||
66e52d22 LP |
263 | * add an Open= setting to service unit files that can open arbitrary file |
264 | system paths at service startup time and pass them to the service process via | |
265 | our usual socket activation protocol. If passed path refers to AF_UNIX | |
266 | socket: connect() to it. | |
267 | ||
268 | * add a ConnectSocket= setting to service unit files, that may reference a | |
269 | socket unit, and which will connect to the socket defined therein, and pass | |
270 | the resulting fd to the service program via socket activation proto. | |
271 | ||
272 | * Add a concept of ListenStream=anonymous to socket units: listen on a socket | |
273 | that is deleted in the fs. Usecase would be with ConnectSocket= above. | |
274 | ||
14f7d087 LP |
275 | * importd: support image signature verification with PKCS#7 + OpenBSD signify |
276 | logic, as alternative to crummy gpg | |
277 | ||
dfbbb4f7 LP |
278 | * sysext: optionally, if the merged trees allow it use bind mounts instead of |
279 | overlayfs | |
280 | ||
9786ba13 LP |
281 | * add "systemd-analyze debug" + AttachDebugger= in unit files: The former |
282 | specifies a command to execute; the latter specifies that an already running | |
283 | "systemd-analyze debug" instance shall be contacted and execution paused | |
284 | until it gives an OK. That way, tools like gdb or strace can be safely be | |
285 | invoked on processes forked off PID 1. | |
286 | ||
38abd1bf LP |
287 | * expose MS_NOSYMFOLLOW in various places |
288 | ||
199b097d LP |
289 | * make LoadCredential= automatically find credentials in /etc/creds, |
290 | /run/creds, … and so on, if path component is unqualified | |
291 | ||
292 | * teach LoadCredential=/LoadCredentialEncrypted= to load credentials from | |
293 | kernel cmdline, maybe: LoadCredentialEncrypted=foobar:proc-cmdline:foobar | |
294 | ||
295 | * credentials system: | |
296 | - acquire from kernel command line | |
297 | - acquire from EFI variable? | |
298 | - acquire via via ask-password? | |
299 | - acquire creds via keyring? | |
300 | - pass creds via keyring? | |
301 | - pass creds via memfd? | |
302 | - acquire + decrypt creds from pkcs11? | |
303 | - make systemd-cryptsetup acquire pw via creds logic | |
304 | - make PAMName= acquire pw via creds logic | |
305 | - make macsec/wireguard code in networkd read key via creds logic | |
306 | - make gatwayd/remote read key via creds logic | |
307 | - add sd_notify() command for flushing out creds not needed anymore | |
308 | ||
309 | * teach LoadCredential= the ability to load all files from a specified dir as | |
310 | individual creds | |
311 | ||
312 | * add tpm.target or so which is delayed until TPM2 device showed up in case | |
313 | firmware indicates there is one. | |
07eabc2b | 314 | |
9aea1e58 LP |
315 | * tpm2: support a PIN policy, i.e. allowing windows-style short authentication |
316 | passwords by using the TPM2 to enforce ratelimiting and such, use for | |
317 | cryptsetup and homed | |
318 | ||
80670e74 LP |
319 | * Add concept for upgrading TPM2 enrollments, maybe a new switch |
320 | --pcrs=4:<hash> or so, i.e. select a PCR to include in the hash, and then | |
321 | override its hash | |
322 | ||
07eabc2b LB |
323 | * TPM2: auto-reenroll in cryptsetup, as fallback for hosed firmware upgrades |
324 | and such | |
325 | ||
326 | * introduce a new group to own TPM devices | |
80670e74 | 327 | |
80670e74 LP |
328 | * cryptsetup: if only recovery keys are registered and no regular passphrases, |
329 | ask user for "recovery key", not "passphrase" | |
330 | ||
331 | * cyptsetup: add option for automatically removing empty password slot on boot | |
332 | ||
7d7c75f1 | 333 | * cryptsetup: optionally, when run during boot-up and password is never |
80670e74 LP |
334 | entered, and we are on battery power (or so), power off machine again |
335 | ||
80670e74 LP |
336 | * cryptsetup: when waiting for FIDO2/PKCS#11 token, tell plymouth that, and |
337 | allow plymouth to abort the waiting and enter pw instead | |
7d7c75f1 | 338 | |
07eabc2b LB |
339 | * make cryptsetup lower --iter-time |
340 | ||
341 | * cryptsetup: allow encoding key directly in /etc/crypttab, maybe with a | |
342 | "base64:" prefix. Useful in particular for pkcs11 mode. | |
343 | ||
344 | * cryptsetup: reimplement the mkswap/mke2fs in cryptsetup-generator to use | |
345 | systemd-makefs.service instead. | |
346 | ||
347 | * cryptsetup: | |
348 | - cryptsetup-generator: allow specification of passwords in crypttab itself | |
349 | - support rd.luks.allow-discards= kernel cmdline params in cryptsetup generator | |
350 | ||
7d7c75f1 LP |
351 | * when configuring loopback netif, and it fails due to EPERM, eat up error if |
352 | it happens to be set up alright already. | |
353 | ||
354 | * at boot: check if battery above some threshold, if not power off again after explanation | |
355 | ||
356 | * userdb: add field for ambient caps, so that a user can have CAP_WAKE_ALARM | |
357 | for example. And add code that resets ambient caps for all services by | |
358 | default. | |
359 | ||
ff640bd2 LP |
360 | * sd-bus: when connecting to some dbus server socker, set originating AF_UNIX |
361 | socket name in abstract namespace to include "description" string, and pick | |
362 | it up from there in sd_bus_creds logic. i.e. we can use the socket peer | |
363 | address as conduit for some minimal connection metainfo, and use it to | |
364 | restore the "description" logic that kdbus used to have. | |
365 | ||
08d33656 | 366 | * systemd-analyze netif that explains predictable interface (or networkctl) |
d775d8e6 | 367 | |
816d460a LP |
368 | * Add service setting to run a service within the specified VRF. i.e. do the |
369 | equivalent of "ip vrf exec". | |
370 | ||
e59d030f LP |
371 | * change SwitchRoot() implementation in PID 1 to use pivot_root(".", "."), as |
372 | documented in the pivot_root(2) man page, so that we can drop the /oldroot | |
373 | temporary dir. | |
374 | ||
375 | * special case some calls of chase_symlinks() to use openat2() internally, so | |
376 | that the kernel does what we otherwise do. | |
377 | ||
6dbfbc46 LB |
378 | * add a new flag to chase_symlinks() that stops chasing once the first missing |
379 | component is found and then allows the caller to create the rest. | |
380 | ||
fc733bed LP |
381 | * make use of new glibc 2.32 APIs sigabbrev_np() and strerrorname_np(). |
382 | ||
effefa30 LP |
383 | * if /usr/bin/swapoff fails due to OOM, log a friendly explanatory message about it |
384 | ||
42ba8d25 LP |
385 | * Remove any support for booting without /usr pre-mounted in the initrd entirely. |
386 | Update INITRD_INTERFACE.md accordingly. | |
387 | ||
f1eb0ccd LP |
388 | * pid1: Move to tracking of main pid/control pid of units per pidfd |
389 | ||
390 | * pid1: support new clone3() fork-into-cgroup feature | |
391 | ||
7cc8fb3e LP |
392 | * pid1: also remove PID files of a service when the service starts, not just |
393 | when it exits | |
394 | ||
cdfd8537 LP |
395 | * make us use dynamically fewer deps for containers in general purpose distros: |
396 | o turn into dlopen() deps: | |
cdfd8537 LP |
397 | - elfutils (always) |
398 | - p11-kit-trust (always) | |
399 | - kmod-libs (only when called from PID 1) | |
a52dc0b6 | 400 | - libblkid (only in RootImage= handling in PID 1, but not elsewhere) |
cdfd8537 LP |
401 | - libpam (only when called from PID 1) |
402 | - bzip2, xz, lz4 (always — gzip and zstd should probably stay static deps the way they are, | |
403 | since they are so basic and our defaults) | |
404 | o move into separate libsystemd-shared-iptables.so .so | |
405 | - iptables-libs (only used by nspawn + networkd) | |
406 | ||
77169ed0 LP |
407 | * seccomp: maybe use seccomp_merge() to merge our filters per-arch if we can. |
408 | Apparently kernel performance is much better with fewer larger seccomp | |
409 | filters than with more smaller seccomp filters. | |
410 | ||
bfafec25 LP |
411 | * systemd-path: add ESP and XBOOTLDR path. Add "private" runtime/state/cache dir enum, |
412 | mapping to $RUNTIME_DIRECTORY, $STATE_DIRECTORY and such | |
413 | ||
a6e1018d | 414 | * All tools that support --root= should also learn --image= so that they can |
612d3a68 LP |
415 | operate on disk images directly. Specifically: bootctl, systemctl, |
416 | coredumpctl. (Already done: systemd-nspawn, systemd-firstboot, | |
417 | systemd-repart, systemd-tmpfiles, systemd-sysusers, journalctl) | |
a6e1018d | 418 | |
112bed84 LP |
419 | * seccomp: by default mask x32 ABI system wide on x86-64. it's on its way out |
420 | ||
421 | * seccomp: don't install filters for ABIs that are masked anyway for the | |
422 | specific service | |
423 | ||
424 | * seccomp: maybe merge all filters we install into one with that libseccomp API that allows merging. | |
425 | ||
cbe952fe LP |
426 | * busctl: maybe expose a verb "ping" for pinging a dbus service to see if it |
427 | exists and responds. | |
428 | ||
b47261e5 LP |
429 | * Maybe add a separate GPT partition type to the discoverable partition spec |
430 | for "hibernate" partitions, that are exactly like swap partitions but only | |
431 | activated right before hibernation and thus never used for regular swapping. | |
432 | ||
91fc013f | 433 | * socket units: allow creating a udev monitor socket with ListenDevices= or so, |
86b52a39 | 434 | with matches, then activate app through that passing socket over |
91fc013f | 435 | |
cdfd8537 LP |
436 | * unify on openssl (as soon as OpenSSL 3.0 is out, and the Debian license |
437 | confusion is gone) | |
492f91d8 LP |
438 | - port resolved over from libgcrypt (DNSSEC code) |
439 | - port journald + fsprg over from libgcrypt | |
492f91d8 LP |
440 | - when that's done: kill gnutls support in resolved |
441 | ||
492f91d8 LP |
442 | * add growvol and makevol options for /etc/crypttab, similar to |
443 | x-systemd.growfs and x-systemd-makefs. | |
444 | ||
77b19caf LP |
445 | * userdb: allow username prefix searches in varlink API, allow realname and |
446 | realname substr searches in varlink API | |
006c44c1 | 447 | |
76410e98 LP |
448 | * userdb: allow uid/gid range checks |
449 | ||
2a4be3c5 | 450 | * userdb: allow existence checks |
006c44c1 | 451 | |
f1eb0ccd | 452 | * pid1: activation by journal search expression |
006c44c1 | 453 | |
492f91d8 LP |
454 | * when switching root from initrd to host, set the machine_id env var so that |
455 | if the host has no machine ID set yet we continue to use the random one the | |
456 | initrd had set. | |
457 | ||
173c7873 | 458 | * sd-event: add native support for P_ALL waitid() watching, then move PID 1 to |
84a1ff94 | 459 | it for reaping assigned but unknown children. This needs to some special care |
173c7873 LP |
460 | to operate somewhat sensibly in light of priorities: P_ALL will return |
461 | arbitrary processes, regardless of the priority we want to watch them with, | |
462 | hence on each event loop iteration check all processes which we shall watch | |
463 | with higher prio explicitly, and then watch the entire rest with P_ALL. | |
464 | ||
465 | * tweak sd-event's child watching: keep a prioq of children to watch and use | |
466 | waitid() only on the children with the highest priority until one is waitable | |
467 | and ignore all lower-prio ones from that point on | |
468 | ||
173c7873 LP |
469 | * maybe introduce xattrs that can be set on the root dir of the root fs |
470 | partition that declare the volatility mode to use the image in. Previously I | |
471 | thought marking this via GPT partition flags but that's not ideal since | |
472 | that's outside of the LUKS encryption/verity verification, and we probably | |
473 | shouldn't operate in a volatile mode unless we got told so from a trusted | |
474 | source. | |
475 | ||
97e5cc88 LP |
476 | * coredump: maybe when coredumping read a new xattr from /proc/$PID/exe that |
477 | may be used to mark a whole binary as non-coredumpable. Would fix: | |
478 | https://bugs.freedesktop.org/show_bug.cgi?id=69447 | |
479 | ||
f3e361c1 LP |
480 | * teach parse_timestamp() timezones like the calendar spec already knows it |
481 | ||
9ef3376b LP |
482 | * beef up hibernation to optionally do swapon/swapoff immediately before/after |
483 | the hibernation | |
484 | ||
c846b233 LP |
485 | * beef up s2h to implement a battery watch loop: instead of entering |
486 | hibernation unconditionally after coming back from resume make a decision | |
487 | based on the battery load level: if battery level is above a specific | |
488 | threshold, go to suspend again, only hibernate if below it. This means we'd | |
489 | stick to suspend usually, but fall back to hibernation only when battery runs | |
490 | empty (well, subject to our sampling interval). Related to this, check if we | |
491 | can make ACPI _BTP (i.e. /sys/class/power_supply/*/alarm) work for us too, | |
492 | i.e. see if it can wake up machines from suspend, so that we could resume | |
493 | automatically when the system is low on power and move automatically to | |
494 | hibernation mode. (see | |
495 | https://uefi.org/sites/default/files/resources/ACPI%206_2_A_Sept29.pdf | |
496 | section 10.2.2.8 and | |
497 | https://docs.microsoft.com/en-us/windows-hardware/design/device-experiences/modern-standby-wake-sources | |
498 | at the end). | |
499 | ||
6fe23ff3 LB |
500 | * We should probably replace /etc/rc.d/README with a symlink to doc |
501 | content. After all it is constant vendor data. | |
9c230b8f | 502 | |
c6526b8d | 503 | * maybe add kernel cmdline params: to force random seed crediting |
312dc153 | 504 | |
ac5dca64 LP |
505 | * introduce a new per-process uuid, similar to the boot id, the machine id, the |
506 | invocation id, that is derived from process creds, specifically a hashed | |
507 | combination of AT_RANDOM + getpid() + the starttime from | |
508 | /proc/self/status. Then add these ids implicitly when logging. Deriving this | |
509 | uuid from these three things has the benefit that it can be derived easily | |
510 | from /proc/$PID/ in a stable, and unique way that changes on both fork() and | |
511 | exec(). | |
512 | ||
513 | * let's not GC a unit while its ratelimits are still pending | |
514 | ||
5daeeecf LP |
515 | * when killing due to service watchdog timeout maybe detect whether target |
516 | process is under ptracing and then log loudly and continue instead. | |
517 | ||
5802d977 LP |
518 | * make rfkill uaccess controllable by default, i.e. steal rule from |
519 | gnome-bluetooth and friends | |
520 | ||
03ccc7f0 LP |
521 | * make MAINPID= message reception checks even stricter: if service uses User=, |
522 | then check sending UID and ignore message if it doesn't match the user or | |
523 | root. | |
524 | ||
525 | * maybe trigger a uevent "change" on a device if "systemctl reload xyz.device" | |
526 | is issued. | |
527 | ||
db3cea22 LP |
528 | * when importing an fs tree with machined, optionally apply userns-rec-chown |
529 | ||
530 | * when importing an fs tree with machined, complain if image is not an OS | |
531 | ||
707b3fbd LP |
532 | * Maybe introduce a helper safe_exec() or so, which is to execve() which |
533 | safe_fork() is to fork(). And then make revert the RLIMIT_NOFILE soft limit | |
534 | to 1K implicitly, unless explicitly opted-out. | |
d7b659ef | 535 | |
d238709c | 536 | * rework seccomp/nnp logic that even if User= is used in combination with |
d96c081a LP |
537 | a seccomp option we don't have to set NNP. For that, change uid first whil |
538 | keeping CAP_SYS_ADMIN, then apply seccomp, the drop cap. | |
539 | ||
8a7cf157 LP |
540 | * when no locale is configured, default to UEFI's PlatformLang variable |
541 | ||
d96c081a LP |
542 | * add a new syscall group "@esoteric" for more esoteric stuff such as bpf() and |
543 | usefaultd() and make systemd-analyze check for it. | |
544 | ||
06898123 | 545 | * paranoia: whenever we process passwords, call mlock() on the memory |
309c6b19 | 546 | first. i.e. look for all places we use free_and_erasep() and |
9ae4ef49 | 547 | augment them with mlock(). Also use MADV_DONTDUMP. |
8b213bf1 | 548 | Alternatively (preferably?) use memfd_secret(). |
06898123 | 549 | |
32875617 LP |
550 | * Move RestrictAddressFamily= to the new cgroup create socket |
551 | ||
552 | * support the bind/connect/sendmsg cgroup stuff for sandboxing, and possibly | |
553 | patching around | |
554 | ||
76853293 LP |
555 | * maybe implicitly attach monotonic+realtime timestamps to outgoing messages in |
556 | log.c and sd-journal-send | |
557 | ||
06898123 LP |
558 | * optionally: turn on cgroup delegation for per-session scope units |
559 | ||
d3aeddb8 LP |
560 | * introduce per-unit (i.e. per-slice, per-service) journal log size limits. |
561 | ||
42e18088 LP |
562 | * sd-boot: optionally, show boot menu when previous default boot item has |
563 | non-zero "tries done" count | |
564 | ||
8f2eb730 LP |
565 | * augment CODE_FILE=, CODE_LINE= with something like CODE_BASE= or so which |
566 | contains some identifier for the project, which allows us to include | |
567 | clickable links to source files generating these log messages. The identifier | |
568 | could be some abberviated URL prefix or so (taking inspiration from Go | |
569 | imports). For example, for systemd we could use | |
570 | CODE_BASE=github.com/systemd/systemd/blob/98b0b1123cc or so which is | |
571 | sufficient to build a link by prefixing "http://" and suffixing the | |
572 | CODE_FILE. | |
573 | ||
8f2eb730 LP |
574 | * Augment MESSAGE_ID with MESSAGE_BASE, in a similar fashion so that we can |
575 | make clickable links from log messages carrying a MESSAGE_ID, that lead to | |
576 | some explanatory text online. | |
577 | ||
7bd4bcf7 LP |
578 | * maybe extend .path units to expose fanotify() per-mount change events |
579 | ||
bb527e11 LP |
580 | * When reloading configuration PID 1 should reset all its properties to the |
581 | original defaults before calling parse_config() | |
582 | ||
c633b0a6 LP |
583 | * hibernate/s2h: make this robust and safe to enable in Fedora by default. |
584 | Specifically: | |
585 | ||
586 | 1. add resume_offset support to the resume code (i.e. support swap files | |
587 | properly) | |
e83419d0 | 588 | 2. check if swap is on weird storage and refuse if so |
46c41478 | 589 | 3. add auto-detection of hibernation images |
c633b0a6 | 590 | |
070d0ac9 LP |
591 | * cgroups: use inotify to get notified when somebody else modifies cgroups |
592 | owned by us, then log a friendly warning. | |
593 | ||
e44924f5 LP |
594 | * beef up log.c with support for stripping ANSI sequences from strings, so that |
595 | it is OK to include them in log strings. This would be particularly useful so | |
596 | that our log messages could contain clickable links for example for unit | |
597 | files and suchlike we operate on. | |
070d0ac9 | 598 | |
07eabc2b LB |
599 | * importd: add ability download images for portabled + sysext |
600 | ||
8c4c2dfc LP |
601 | * add support for "portablectl attach http://foobar.com/waaa.raw (i.e. importd integration) |
602 | ||
8c4c2dfc LP |
603 | * sync dynamic uids/gids between host+portable srvice (i.e. if DynamicUser=1 is set for a service, make sure that the |
604 | selected user is resolvable in the service even if it ships its own /etc/passwd) | |
605 | ||
5da19043 | 606 | * Fix DECIMAL_STR_MAX or DECIMAL_STR_WIDTH. One includes a trailing NUL, the |
5238e957 | 607 | other doesn't. What a disaster. Probably to exclude it. Also |
5da19043 LP |
608 | DECIMAL_STR_WIDTH should probably add an extra "-" into account for negative |
609 | numbers. | |
610 | ||
16270697 LP |
611 | * Check that users of inotify's IN_DELETE_SELF flag are using it properly, as |
612 | usually IN_ATTRIB is the right way to watch deleted files, as the former only | |
613 | fires when a file is actually removed from disk, i.e. the link count drops to | |
614 | zero and is not open anymore, while the latter happens when a file is | |
615 | unlinked from any dir. | |
616 | ||
bd1b3f75 | 617 | * port systemctl, busctl, … over to format-table.[ch]'s table formatters |
5da19043 | 618 | |
7bc756ff LP |
619 | * pid1: lock image configured with RootDirectory=/RootImage= using the usual nspawn semantics while the unit is up |
620 | ||
621 | * add --vacuum-xyz options to coredumpctl, matching those journalctl already has. | |
622 | ||
53c70a27 LP |
623 | * introduce Ephemeral= unit file switch, that creates an ephemeral copy of all |
624 | files and directories that are left writable for a unit, and which are | |
625 | removed after the unit goes down again. A bit like --ephemeral for | |
626 | systemd-nspawn but for system services. If used together with RootImage= this | |
627 | should reflink the image file itself. | |
628 | ||
629 | Related: add Ephemeral=<path1> <path2> … which would allow marking | |
630 | specific paths only like this. | |
631 | ||
53c70a27 | 632 | * add CopyFile= or so as unit file setting that may be used to copy files or |
5238e957 | 633 | directory trees from the host to the services RootImage= and RootDirectory= |
53c70a27 LP |
634 | environment. Which we can use for /etc/machine-id and in particular |
635 | /etc/resolv.conf. Should be smart and do something useful on read-only | |
2aed63f4 | 636 | images, for example fall back to read-only bind mounting the file instead. |
53c70a27 | 637 | |
d9b50610 LP |
638 | * show invocation ID in systemd-run output |
639 | ||
640 | * bypass SIGTERM state in unit files if KillSignal is SIGKILL | |
641 | ||
9711b1ad LP |
642 | * add proper dbus APIs for the various sd_notify() commands, such as MAINPID=1 |
643 | and so on, which would mean we could report errors and such. | |
644 | ||
6b7b0f39 LP |
645 | * introduce DefaultSlice= or so in system.conf that allows changing where we |
646 | place our units by default, i.e. change system.slice to something | |
647 | else. Similar, ManagerSlice= should exist so that PID1's own scope unit could | |
648 | be moved somewhere else too. Finally machined and logind should get similar | |
649 | options so that it is possible to move user session scopes and machines to a | |
650 | different slice too by default. Usecase: people who want to put resources on | |
651 | the entire system, with the exception of one specific service. See: | |
652 | https://lists.freedesktop.org/archives/systemd-devel/2018-February/040369.html | |
653 | ||
aa79f932 LP |
654 | * maybe rework get_user_creds() to query the user database if $SHELL is used |
655 | for root, but only then. | |
586a8e93 | 656 | |
46099c9e LP |
657 | * be stricter with fds we receive for the fdstore: close them asynchronously |
658 | ||
46099c9e LP |
659 | * calenderspec: add support for week numbers and day numbers within a |
660 | year. This would allow us to define "bi-weekly" triggers safely. | |
661 | ||
c3cd7cc9 LP |
662 | * sd-bus: add vtable flag, that may be used to request client creds implicitly |
663 | and asynchronously before dispatching the operation | |
664 | ||
d0e5db44 ZJS |
665 | * sd-bus: parse addresses given in sd_bus_set_addresses immediately and not |
666 | only when used. Add unit tests. | |
667 | ||
06345858 LP |
668 | * make use of ethtool veth peer info in machined, for automatically finding out |
669 | host-side interface pointing to the container. | |
670 | ||
48f1b5e5 LP |
671 | * add some special mode to LogsDirectory=/StateDirectory=… that allows |
672 | declaring these directories without necessarily pulling in deps for them, or | |
673 | creating them when starting up. That way, we could declare that | |
674 | systemd-journald writes to /var/log/journal, which could be useful when we | |
675 | doing disk usage calculations and so on. | |
676 | ||
cb77e122 | 677 | * taint systemd if there are fewer than 65536 users assigned (userns) to the system. |
c6009ff0 | 678 | |
899feb72 | 679 | * deprecate RootDirectoryStartOnly= in favour of a new ExecStart= prefix char |
c6009ff0 | 680 | |
3d80d454 LP |
681 | * add a new RuntimeDirectoryPreserve= mode that defines a similar lifecycle for |
682 | the runtime dir as we maintain for the fdstore: i.e. keep it around as long | |
683 | as the unit is running or has a job queued. | |
684 | ||
5f7ecd61 | 685 | * support projid-based quota in machinectl for containers |
5962e9db | 686 | |
17b6f896 LP |
687 | * add a way to lock down cgroup migration: a boolean, which when set for a unit |
688 | makes sure the processes in it can never migrate out of it | |
689 | ||
17b6f896 LP |
690 | * blog about fd store and restartable services |
691 | ||
692 | * document Environment=SYSTEMD_LOG_LEVEL=debug drop-in in debugging document | |
693 | ||
370f9c21 LP |
694 | * rework ExecOutput and ExecInput enums so that EXEC_OUTPUT_NULL loses its |
695 | magic meaning and is no longer upgraded to something else if set explicitly. | |
696 | ||
17b6f896 LP |
697 | * in the long run: permit a system with /etc/machine-id linked to /dev/null, to |
698 | make it lose its identity, i.e. be anonymous. For this we'd have to patch | |
699 | through the whole tree to make all code deal with the case where no machine | |
700 | ID is available. | |
701 | ||
702 | * optionally, collect cgroup resource data, and store it in per-unit RRD files, | |
703 | suitable for processing with rrdtool. Add bus API to access this data, and | |
704 | possibly implement a CPULoad property based on it. | |
705 | ||
e7e4a258 LP |
706 | * beef up pam_systemd to take unit file settings such as cgroups properties as |
707 | parameters | |
708 | ||
6fc373ee | 709 | * maybe hook of xfs/ext4 quotactl() with services? i.e. automatically manage |
d51c4fca | 710 | the quota of the user indicated in User= via unit file settings, like the |
5962e9db LP |
711 | other resource management concepts. Would mix nicely with DynamicUser=1. Or |
712 | alternatively, do this with projids, so that we can also cover services | |
713 | running as root. Quota should probably cover all the special dirs such as | |
714 | StateDirectory=, LogsDirectory=, CacheDirectory=, as well as RootDirectory= if it | |
715 | is set, plus the whole disk space any image configured with RootImage=. | |
6fc373ee | 716 | |
9a92a97a LP |
717 | * In DynamicUser= mode: before selecting a UID, use disk quota APIs on relevant |
718 | disks to see if the UID is already in use. | |
719 | ||
fa991fb7 LP |
720 | * expose IO accounting data on the bus, show it in systemd-run --wait and log |
721 | about it in the resource log message | |
722 | ||
d73b607d LP |
723 | * Add AddUser= setting to unit files, similar to DynamicUser=1 which however |
724 | creates a static, persistent user rather than a dynamic, transient user. We | |
725 | can leverage code from sysusers.d for this. | |
726 | ||
fd63f36c LP |
727 | * add some optional flag to ReadWritePaths= and friends, that has the effect |
728 | that we create the dir in question when the service is started. Example: | |
729 | ||
730 | ReadWritePaths=:/var/lib/foobar | |
731 | ||
fe6a0235 LP |
732 | * hostnamed: populate form factor data from a new hwdb database, so that old |
733 | yogas can be recognized as "convertible" too, even if they predate the DMI | |
734 | "convertible" form factor | |
735 | ||
f59d1da8 LP |
736 | * Add ExecMonitor= setting. May be used multiple times. Forks off a process in |
737 | the service cgroup, which is supposed to monitor the service, and when it | |
738 | exits the service is considered failed by its monitor. | |
739 | ||
33bac67b LP |
740 | * track the per-service PAM process properly (i.e. as an additional control |
741 | process), so that it may be queried on the bus and everything. | |
742 | ||
743 | * add a new "debug" job mode, that is propagated to unit_start() and for | |
744 | services results in two things: we raise SIGSTOP right before invoking | |
745 | execve() and turn off watchdog support. Then, use that to implement | |
746 | "systemd-gdb" for attaching to the start-up of any system service in its | |
747 | natural habitat. | |
748 | ||
8eb7383b LP |
749 | * 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 |
750 | ||
b5bdbcd5 LP |
751 | * add a percentage syntax for TimeoutStopSec=, e.g. TimeoutStopSec=150%, and |
752 | then use that for the setting used in user@.service. It should be understood | |
753 | relative to the configured default value. | |
754 | ||
d21494ea LP |
755 | * enable LockMLOCK to take a percentage value relative to physical memory |
756 | ||
04397464 | 757 | * Permit masking specific netlink APIs with RestrictAddressFamily= |
d82047be | 758 | |
d82047be LP |
759 | * define gpt header bits to select volatility mode |
760 | ||
42d61ded LP |
761 | * ProtectClock= (drops CAP_SYS_TIMES, adds seecomp filters for settimeofday, adjtimex), sets DeviceAllow o /dev/rtc |
762 | ||
04397464 | 763 | * ProtectTracing= (drops CAP_SYS_PTRACE, blocks ptrace syscall, makes /sys/kernel/tracing go away) |
42d61ded | 764 | |
04397464 | 765 | * ProtectMount= (drop mount/umount/pivot_root from seccomp, disallow fuse via DeviceAllow, imply Mountflags=slave) |
563a69f4 | 766 | |
04397464 | 767 | * ProtectKeyRing= to take keyring calls away |
2c5f2958 | 768 | |
8ce9b83a | 769 | * RemoveKeyRing= to remove all keyring entries of the specified user |
a46eac1b LP |
770 | |
771 | * ProtectReboot= that masks reboot() and kexec_load() syscalls, prohibits kill | |
772 | on PID 1 with the relevant signals, and makes relevant files in /sys and | |
773 | /proc (such as the sysrq stuff) unavailable | |
774 | ||
88511a37 LB |
775 | * Support ReadWritePaths/ReadOnlyPaths/InaccessiblePaths in systemd --user instances |
776 | via the new unprivileged Landlock LSM (https://landlock.io) | |
777 | ||
e40a326c | 778 | * make sure the ratelimit object can deal with USEC_INFINITY as way to turn off things |
89f193fa | 779 | |
8ce9b83a LP |
780 | * in nss-systemd, if we run inside of RootDirectory= with PrivateUsers= set, |
781 | find a way to map the User=/Group= of the service to the right name. This way | |
782 | a user/group for a service only has to exist on the host for the right | |
783 | mapping to work. | |
784 | ||
2c5f2958 LP |
785 | * add bus API for creating unit files in /etc, reusing the code for transient units |
786 | ||
787 | * add bus API to remove unit files from /etc | |
788 | ||
789 | * add bus API to retrieve current unit file contents (i.e. implement "systemctl cat" on the bus only) | |
790 | ||
b8c7afdf LP |
791 | * rework fopen_temporary() to make use of open_tmpfile_linkable() (problem: the |
792 | kernel doesn't support linkat() that replaces existing files, currently) | |
793 | ||
1e555cb5 LP |
794 | * transient units: don't bother with actually setting unit properties, we |
795 | reload the unit file anyway | |
796 | ||
c8048350 LP |
797 | * optionally, also require WATCHDOG=1 notifications during service start-up and shutdown |
798 | ||
3d39e6e5 LP |
799 | * cache sd_event_now() result from before the first iteration... |
800 | ||
bd098bce LP |
801 | * PID1: find a way how we can reload unit file configuration for |
802 | specific units only, without reloading the whole of systemd | |
803 | ||
f9bf1b8f | 804 | * add an explicit parser for LimitRTPRIO= that verifies |
de7399eb | 805 | the specified range and generates sane error messages for incorrect |
f9bf1b8f | 806 | specifications. |
de7399eb | 807 | |
3efc8c72 LP |
808 | * when we detect that there are waiting jobs but no running jobs, do something |
809 | ||
5ebbb45b | 810 | * push CPUAffinity= also into the "cpuset" cgroup controller |
03364e47 | 811 | |
e6a26d8c LP |
812 | * PID 1 should send out sd_notify("WATCHDOG=1") messages (for usage in the --user mode, and when run via nspawn) |
813 | ||
a2088fd0 | 814 | * there's probably something wrong with having user mounts below /sys, |
5238e957 | 815 | as we have for debugfs. for example, src/core/mount.c handles mounts |
a2088fd0 LP |
816 | prefixed with /sys generally special. |
817 | http://lists.freedesktop.org/archives/systemd-devel/2015-June/032962.html | |
818 | ||
8aa20381 LP |
819 | * fstab-generator: default to tmpfs-as-root if only usr= is specified on the kernel cmdline |
820 | ||
a7972611 ZJS |
821 | * initrd-parse-etc.service: can we skip daemon-reload if /sysroot/etc/fstab is missing? |
822 | Note that we start initrd-fs.target and initrd-cleanup.target there, so a straightforward | |
823 | ConditionPathExists= is not enough. | |
824 | ||
b18d23d7 LP |
825 | * docs: bring http://www.freedesktop.org/wiki/Software/systemd/MyServiceCantGetRealtime up to date |
826 | ||
60d17b74 LP |
827 | * add a job mode that will fail if a transaction would mean stopping |
828 | running units. Use this in timedated to manage the NTP service | |
829 | state. | |
830 | http://lists.freedesktop.org/archives/systemd-devel/2015-April/030229.html | |
831 | ||
477e75ef LP |
832 | * The udev blkid built-in should expose a property that reflects |
833 | whether media was sensed in USB CF/SD card readers. This should then | |
834 | be used to control SYSTEMD_READY=1/0 so that USB card readers aren't | |
835 | picked up by systemd unless they contain a medium. This would mirror | |
836 | the behaviour we already have for CD drives. | |
837 | ||
c3a0d00d LP |
838 | * hostnamectl: show root image uuid |
839 | ||
d2f81fb0 LP |
840 | * Find a solution for SMACK capabilities stuff: |
841 | http://lists.freedesktop.org/archives/systemd-devel/2014-December/026188.html | |
98cd2651 | 842 | |
0a86c1a9 LP |
843 | * synchronize console access with BSD locks: |
844 | http://lists.freedesktop.org/archives/systemd-devel/2014-October/024582.html | |
845 | ||
e031c227 | 846 | * as soon as we have sender timestamps, revisit coalescing multiple parallel daemon reloads: |
0a86c1a9 LP |
847 | http://lists.freedesktop.org/archives/systemd-devel/2014-December/025862.html |
848 | ||
8514b677 LP |
849 | * figure out when we can use the coarse timers |
850 | ||
8483d73f LP |
851 | * maybe allow timer units with an empty Units= setting, so that they |
852 | can be used for resuming the system but nothing else. | |
853 | ||
25e773ee | 854 | * what to do about udev db binary stability for apps? (raw access is not an option) |
b857e042 | 855 | |
720652b3 | 856 | * exponential backoff in timesyncd when we cannot reach a server |
42aeb14a | 857 | |
720652b3 | 858 | * timesyncd: add ugly bus calls to set NTP servers per-interface, for usage by NM |
e25b5a8d | 859 | |
563b1bdc LP |
860 | * merge ~/.local/share and ~/.local/lib into one similar /usr/lib and /usr/share.... |
861 | ||
6bd7941e TG |
862 | * add systemd.abort_on_kill or some other such flag to send SIGABRT instead of SIGKILL |
863 | (throughout the codebase, not only PID1) | |
864 | ||
07eabc2b LB |
865 | * drop nss-myhostname in favour of nss-resolve? |
866 | ||
9d6db739 | 867 | * resolved: |
9d6db739 | 868 | - mDNS/DNS-SD |
ccc3e8a1 LP |
869 | - service registration |
870 | - service/domain/types browsing | |
0f47ed0a | 871 | - avahi compat |
9d6db739 | 872 | - DNS-SD service registration from socket units |
e25b5a8d DH |
873 | - resolved should optionally register additional per-interface LLMNR |
874 | names, so that for the container case we can establish the same name | |
875 | (maybe "host") for referencing the server, everywhere. | |
720652b3 | 876 | - allow clients to request DNSSEC for a single lookup even if DNSSEC is off (?) |
3efb871a | 877 | - hook up resolved with machined-based address resolution |
3f77a1b1 | 878 | |
e25b5a8d | 879 | * refcounting in sd-resolve is borked |
e2a69298 | 880 | |
a940778f LP |
881 | * add new gpt type for btrfs volumes |
882 | ||
3de03738 LP |
883 | * generator that automatically discovers btrfs subvolumes, identifies their purpose based on some xattr on them. |
884 | ||
37efac5d LP |
885 | * a way for container managers to turn off getty starting via $container_headless= or so... |
886 | ||
7348b3ad LP |
887 | * figure out a nice way how we can let the admin know what child/sibling unit causes cgroup membership for a specific unit |
888 | ||
81429136 KS |
889 | * For timer units: add some mechanisms so that timer units that trigger immediately on boot do not have the services |
890 | they run added to the initial transaction and thus confuse Type=idle. | |
e107ed18 | 891 | |
edb2935c LP |
892 | * add bus api to query unit file's X fields. |
893 | ||
6a3f892a | 894 | * gpt-auto-generator: |
2a781fc9 | 895 | - Define new partition type for encrypted swap? Support probed LUKS for encrypted swap? |
6a3f892a | 896 | - Make /home automount rather than mount? |
6a3f892a | 897 | |
65026403 LP |
898 | * add generator that pulls in systemd-network from containers when |
899 | CAP_NET_ADMIN is set, more than the loopback device is defined, even | |
900 | when it is otherwise off | |
f8901862 | 901 | |
f9bf1b8f | 902 | * MessageQueueMessageSize= (and suchlike) should use parse_iec_size(). |
eda8f067 | 903 | |
af1082b0 LP |
904 | * implement Distribute= in socket units to allow running multiple |
905 | service instances processing the listening socket, and open this up | |
906 | for ReusePort= | |
907 | ||
f38afcd0 | 908 | * cgroups: |
f38afcd0 | 909 | - implement per-slice CPUFairScheduling=1 switch |
f38afcd0 LP |
910 | - introduce high-level settings for RT budget, swappiness |
911 | - how to reset dynamically changed unit cgroup attributes sanely? | |
912 | - when reloading configuration, apply new cgroup configuration | |
913 | - when recursively showing the cgroup hierarchy, optionally also show | |
914 | the hierarchies of child processes | |
0bee65f0 | 915 | |
f38afcd0 | 916 | * transient units: |
f38afcd0 | 917 | - add field to transient units that indicate whether systemd or somebody else saves/restores its settings, for integration with libvirt |
ebcf1f97 | 918 | |
718db961 LP |
919 | * when we detect low battery and no AC on boot, show pretty splash and refuse boot |
920 | ||
718db961 LP |
921 | * libsystemd-journal, libsystemd-login, libudev: add calls to easily attach these objects to sd-event event loops |
922 | ||
966204e0 LP |
923 | * be more careful what we export on the bus as (usec_t) 0 and (usec_t) -1 |
924 | ||
41644622 LP |
925 | * 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 |
926 | ||
7f79cd71 | 927 | * After coming back from hibernation reset hibernation swap partition using the /dev/snapshot ioctl APIs |
0aafd43d | 928 | |
19aadacf JE |
929 | * If we try to find a unit via a dangling symlink, generate a clean |
930 | error. Currently, we just ignore it and read the unit from the search | |
df5f6971 LP |
931 | path anyway. |
932 | ||
04397464 | 933 | * refuse boot if /usr/lib/os-release is missing or /etc/machine-id cannot be set up |
fcba531e | 934 | |
bdeeb6b5 LP |
935 | * man: the documentation of Restart= currently is very misleading and suggests the tools from ExecStartPre= might get restarted. |
936 | ||
937 | * load .d/*.conf dropins for device units | |
938 | ||
07eabc2b | 939 | * There's currently no way to cancel fsck (used to be possible via C-c or c on the console) |
f38afcd0 | 940 | |
07eabc2b | 941 | * add option to sockets to avoid activation. Instead just drop packets/connections, see http://cyberelk.net/tim/2012/02/15/portreserve-systemd-solution/ |
eb01ba5d | 942 | |
07eabc2b LB |
943 | * make sure systemd-ask-password-wall does not shutdown systemd-ask-password-console too early |
944 | ||
945 | * verify that the AF_UNIX sockets of a service in the fs still exist | |
946 | when we start a service in order to avoid confusion when a user | |
947 | assumes starting a service is enough to make it accessible | |
948 | ||
949 | * Make it possible to set the keymap independently from the font on | |
950 | the kernel cmdline. Right now setting one resets also the other. | |
951 | ||
952 | * and a dbus call to generate target from current state | |
953 | ||
954 | * investigate whether the gnome pty helper should be moved into systemd, to provide cgroup support. | |
955 | ||
956 | * dot output for --test showing the 'initial transaction' | |
957 | ||
958 | * be able to specify a forced restart of service A where service B depends on, in case B | |
959 | needs to be auto-respawned? | |
960 | ||
961 | * pid1: | |
962 | - When logging about multiple units (stopping BoundTo units, conflicts, etc.), | |
963 | log both units as UNIT=, so that journalctl -u triggers on both. | |
964 | - generate better errors when people try to set transient properties | |
965 | that are not supported... | |
966 | http://lists.freedesktop.org/archives/systemd-devel/2015-February/028076.html | |
967 | - maybe introduce WantsMountsFor=? Usecase: | |
968 | http://lists.freedesktop.org/archives/systemd-devel/2015-January/027729.html | |
969 | - recreate systemd's D-Bus private socket file on SIGUSR2 | |
970 | - move PAM code into its own binary | |
971 | - when we automatically restart a service, ensure we restart its rdeps, too. | |
972 | - hide PAM options in fragment parser when compile time disabled | |
973 | - Support --test based on current system state | |
974 | - 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(). | |
975 | - after deserializing sockets in socket.c we should reapply sockopts and things | |
976 | - drop PID 1 reloading, only do reexecing (difficult: Reload() | |
977 | currently is properly synchronous, Reexec() is weird, because we | |
978 | cannot delay the response properly until we are back, so instead of | |
979 | being properly synchronous we just keep open the fd and close it | |
980 | when done. That means clients do not get a successful method reply, | |
981 | but much rather a disconnect on success. | |
982 | - when breaking cycles drop sysv services first, then services from /run, then from /etc, then from /usr | |
983 | - when a bus name of a service disappears from the bus make sure to queue further activation requests | |
984 | - maybe introduce CoreScheduling=yes/no to optionally set a PR_SCHED_CORE cookie, so that all | |
985 | processes in a service's cgroup share the same cookie and are guaranteed not to share SMT cores | |
986 | with other units https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/admin-guide/hw-vuln/core-scheduling.rst | |
987 | ||
988 | * unit files: | |
989 | - allow port=0 in .socket units | |
990 | - maybe introduce ExecRestartPre= | |
991 | - add ReloadSignal= for configuring a reload signal to use | |
992 | - implement Register= switch in .socket units to enable registration | |
993 | in Avahi, RPC and other socket registration services. | |
994 | - allow Type=simple with PIDFile= | |
995 | https://bugzilla.redhat.com/show_bug.cgi?id=723942 | |
996 | - allow writing multiple conditions in unit files on one line | |
997 | - introduce Type=pid-file | |
998 | - add a concept of RemainAfterExit= to scope units | |
999 | - Allow multiple ExecStart= for all Type= settings, so that we can cover rescue.service nicely | |
1000 | - add verification of [Install] section to systemd-analyze verify | |
1001 | ||
1002 | * timer units: | |
1003 | - timer units should get the ability to trigger when: | |
1004 | o DST changes | |
1005 | - Modulate timer frequency based on battery state | |
1006 | ||
1007 | * add libsystemd-password or so to query passwords during boot using the password agent logic | |
1008 | ||
1009 | * clean up date formatting and parsing so that all absolute/relative timestamps we format can also be parsed | |
1010 | ||
1011 | * on shutdown: move utmp, wall, audit logic all into PID 1 (or logind?), get rid of systemd-update-utmp-runlevel | |
1012 | ||
1013 | * make repeated alt-ctrl-del presses printing a dump | |
1014 | ||
1015 | * hostnamed: before returning information from /etc/machine-info.conf check the modification data and reread. Similar for localed, ... | |
1016 | ||
1017 | * currently x-systemd.timeout is lost in the initrd, since crypttab is copied into dracut, but fstab is not | |
461bd8e4 | 1018 | |
ab8e074c LP |
1019 | * add a pam module that passes the hdd passphrase into the PAM stack and then expires it, for usage by gdm auto-login. |
1020 | ||
1021 | * add a pam module that on password changes updates any LUKS slot where the password matches | |
1022 | ||
fff87a35 | 1023 | * test/: |
20d52ab6 | 1024 | - add unit tests for config_parse_device_allow() |
b8b4d3dd | 1025 | |
b5c03638 | 1026 | * seems that when we follow symlinks to units we prefer the symlink |
d28315e4 | 1027 | destination path over /etc and /usr. We should not do that. Instead |
b5c03638 LP |
1028 | /etc should always override /run+/usr and also any symlink |
1029 | destination. | |
1030 | ||
eece8c6f LP |
1031 | * when isolating, try to figure out a way how we implicitly can order |
1032 | all units we stop before the isolating unit... | |
1033 | ||
356ce991 LP |
1034 | * teach ConditionKernelCommandLine= globs or regexes (in order to match foobar={no,0,off}) |
1035 | ||
6daebf9e ZJS |
1036 | * Add ConditionDirectoryNotEmpty= handle non-absoute paths as a search path or add |
1037 | ConditionConfigSearchPathNotEmpty= or different syntax? See the discussion starting at | |
1038 | https://github.com/systemd/systemd/pull/15109#issuecomment-607740136. | |
1039 | ||
1ccfb792 LP |
1040 | * BootLoaderSpec: Define a way how an installer can figure out whether a BLS |
1041 | compliant boot loader is installed. | |
795607b2 | 1042 | |
0be8342c LP |
1043 | * think about requeuing jobs when daemon-reload is issued? usecase: |
1044 | the initrd issues a reload after fstab from the host is accessible | |
1045 | and we might want to requeue the mounts local-fs acquired through | |
1046 | that automatically. | |
1047 | ||
e5ec62c5 | 1048 | * systemd-inhibit: make taking delay locks useful: support sending SIGINT or SIGTERM on PrepareForSleep() |
54c31a79 | 1049 | |
ccddd104 | 1050 | * remove any syslog support from log.c — we probably cannot do this before split-off udev is gone for good |
826872b6 | 1051 | |
3679d112 LP |
1052 | * shutdown logging: store to EFI var, and store to USB stick? |
1053 | ||
356ce991 | 1054 | * merge unit_kill_common() and unit_kill_context() |
490b7e47 | 1055 | |
07eabc2b LB |
1056 | * add a dependency on standard-conf.xml and other included files to man pages |
1057 | ||
1058 | * MountFlags=shared acts as MountFlags=slave right now. | |
1059 | ||
1060 | * properly handle loop back mounts via fstab, especially regards to fsck/passno | |
1061 | ||
1062 | * initialize the hostname from the fs label of /, if /etc/hostname does not exist? | |
1063 | ||
1064 | * sd-bus: | |
1065 | - EBADSLT handling | |
1066 | - GetAllProperties() on a non-existing object does not result in a failure currently | |
1067 | - port to sd-resolve for connecting to TCP dbus servers | |
1068 | - 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 | |
1069 | - see if we can drop more message validation on the sending side | |
1070 | - add API to clone sd_bus_message objects | |
1071 | - longer term: priority inheritance | |
1072 | - dbus spec updates: | |
1073 | - NameLost/NameAcquired obsolete | |
1074 | - GVariant | |
1075 | - path escaping | |
1076 | - update systemd.special(7) to mention that dbus.socket is only about the compatibility socket now | |
1077 | ||
1078 | * sd-event | |
1079 | - allow multiple signal handlers per signal? | |
1080 | - document chaining of signal handler for SIGCHLD and child handlers | |
1081 | - define more intervals where we will shift wakeup intervals around in, 1h, 6h, 24h, ... | |
1082 | - maybe support iouring as backend, so that we allow hooking read and write | |
1083 | operations instead of IO ready events into event loops. See considerations | |
1084 | here: | |
1085 | http://blog.vmsplice.net/2020/07/rethinking-event-loop-integration-for.html | |
1086 | ||
1087 | * dbus: when a unit failed to load (i.e. is in UNIT_ERROR state), we | |
1088 | should be able to safely try another attempt when the bus call LoadUnit() is invoked. | |
1089 | ||
1090 | * maybe do not install getty@tty1.service symlink in /etc but in /usr? | |
1091 | ||
1092 | * print a nicer explanation if people use variable/specifier expansion in ExecStart= for the first word | |
1093 | ||
1094 | * mount: turn dependency information from /proc/self/mountinfo into dependency information between systemd units. | |
1095 | ||
1096 | * firstboot: allow provisioning of /etc/hosts entries, so that we can via the | |
1097 | credentials logic insert host name to resolve into containers/hosts. Usecase: | |
1098 | fork a container, and make it ping some specific address which is defined by | |
1099 | the host on invocation | |
1100 | ||
1101 | * systemd-firstboot: make sure to always use chase_symlinks() before | |
1102 | reading/writing files | |
1103 | ||
1104 | * 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 | |
1105 | ||
1106 | * sd-boot: define a drop-in dir in the ESP that may contain X.509 | |
1107 | certificates. If the firmware is detected to be in setup mode, automatically | |
1108 | enroll them as PK/KEK/db, turn off setup mode and proceed. Optionally, | |
1109 | instead of auto-enrolling them add them to the sd-boot menu, giving the user | |
1110 | the option to manually enroll them, after selecting the menu entry. This way, | |
1111 | installer images can just drop the certfiicates in the ESP, and on first boot | |
1112 | can easily enroll the keys without ever booting up. | |
1113 | ||
1114 | * efi stub: optionally, load initrd from disk as a separate file, HMAC check it | |
1115 | with key from TPM, bound to PCR, refusing if failing. This would then allow | |
1116 | traditional distros that generate initrds locally to secure them with TPM: | |
1117 | after generating the initrd, do the HMAC calculation, put result in initrd | |
1118 | filename, done. This would then bind the validity of the initrd to the local | |
1119 | host, and used kernel, and means people cannot change initrd or kernel | |
1120 | without booting the kernel + initrd. | |
1121 | ||
b44be3ec | 1122 | * EFI: |
b44be3ec LP |
1123 | - honor language efi variables for default language selection (if there are any?) |
1124 | - honor timezone efi variables for default timezone selection (if there are any?) | |
466784c8 | 1125 | - change bootctl to be backed by systemd-bootd to control temporary and persistent default boot goal plus efi variables |
631427d6 | 1126 | * bootctl |
631427d6 | 1127 | - recognize the case when not booted on EFI |
e4181484 | 1128 | |
07eabc2b | 1129 | * bootctl,sd-boot: actually honour the "architecture" key |
e9fd44b7 | 1130 | |
07eabc2b LB |
1131 | * bootctl: |
1132 | - teach it to prepare an ESP wholesale, i.e. with mkfs.vfat invocation | |
1133 | - teach it to copy in unified kernel images and maybe type #1 boot loader spec entries from host | |
1134 | - make it operate on loopback files, dissecting enough to find ESP to operate on | |
e6c6e7af | 1135 | |
b44be3ec LP |
1136 | * logind: |
1137 | - logind: optionally, ignore idle-hint logic for autosuspend, block suspend as long as a session is around | |
b44be3ec LP |
1138 | - logind: wakelock/opportunistic suspend support |
1139 | - Add pretty name for seats in logind | |
1140 | - logind: allow showing logout dialog from system? | |
f38afcd0 LP |
1141 | - add Suspend() bus calls which take timestamps to fix double suspend issues when somebody hits suspend and closes laptop quickly. |
1142 | - if pam_systemd is invoked by su from a process that is outside of a | |
1143 | any session we should probably just become a NOP, since that's | |
1144 | usually not a real user session but just some system code that just | |
1145 | needs setuid(). | |
279f0366 LP |
1146 | - logind: make the Suspend()/Hibernate() bus calls wait for the for |
1147 | the job to be completed. before returning, so that clients can wait | |
1148 | for "systemctl suspend" to finish to know when the suspending is | |
1149 | complete. | |
1150 | - logind: when the power button is pressed short, just popup a | |
1151 | logout dialog. If it is pressed for 1s, do the usual | |
1152 | shutdown. Inspiration are Macs here. | |
28423d9a | 1153 | - expose "Locked" property on logind session objects |
e25b5a8d DH |
1154 | - maybe allow configuration of the StopTimeout for session scopes |
1155 | - rename session scope so that it includes the UID. THat way | |
1156 | the session scope can be arranged freely in slices and we don't have | |
1157 | make assumptions about their slice anymore. | |
1158 | - follow PropertiesChanged state more closely, to deal with quick logouts and | |
1159 | relogins | |
77b19caf | 1160 | - (optionally?) spawn seat-manager@$SEAT.service whenever a seat shows up that as CanGraphical set |
e673ad04 | 1161 | |
07eabc2b LB |
1162 | * move logind udev rules to top-level rule.d/ directory |
1163 | ||
1164 | * move multiseat vid/pid matches from logind udev rule to hwdb | |
1165 | ||
1166 | * logind: rework pam_logind to also do a bus call in case of invocation from | |
1167 | user@.service, which returns the XDG_RUNTIME_DIR value, and make this | |
1168 | behaviour selectable via pam module option. | |
1169 | ||
1170 | * delay activation of logind until somebody logs in, or when /dev/tty0 pulls it | |
1171 | in or lingering is on (so that containers don't bother with it until PAM is used). also exit-on-idle | |
1172 | ||
b44be3ec | 1173 | * journal: |
57f2a947 | 1174 | - consider introducing implicit _TTY= + _PPID= + _EUID= + _EGID= + _FSUID= + _FSGID= fields |
b44be3ec LP |
1175 | - journald: also get thread ID from client, plus thread name |
1176 | - journal: when waiting for journal additions in the client always sleep at least 1s or so, in order to minimize wakeups | |
1177 | - add API to close/reopen/get fd for journal client fd in libsystemd-journal. | |
2aed63f4 | 1178 | - fall back to /dev/log based logging in libsystemd-journal, if we cannot log natively? |
b44be3ec | 1179 | - declare the local journal protocol stable in the wiki interface chart |
b44be3ec LP |
1180 | - sd-journal: speed up sd_journal_get_data() with transparent hash table in bg |
1181 | - journald: when dropping msgs due to ratelimit make sure to write | |
1182 | "dropped %u messages" not only when we are about to print the next | |
5238e957 | 1183 | message that works, but already after a short timeout |
b44be3ec LP |
1184 | - check if we can make journalctl by default use --follow mode inside of less if called without args? |
1185 | - maybe add API to send pairs of iovecs via sd_journal_send | |
f47ec8eb | 1186 | - journal: add a setgid "systemd-journal" utility to invoke from libsystemd-journal, which passes fds via STDOUT and does PK access |
b44be3ec LP |
1187 | - journactl: support negative filtering, i.e. FOOBAR!="waldo", |
1188 | and !FOOBAR for events without FOOBAR. | |
06847d0f | 1189 | - journal: store timestamp of journal_file_set_offline() in the header, |
038cf334 | 1190 | so it is possible to display when the file was last synced. |
b44be3ec LP |
1191 | - 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. |
1192 | - journal: find a way to allow dropping history early, based on priority, other rules | |
1193 | - journal: When used on NFS, check payload hashes | |
b44be3ec LP |
1194 | - journald: add kernel cmdline option to disable ratelimiting for debug purposes |
1195 | - refuse taking lower-case variable names in sd_journal_send() and friends. | |
1196 | - journald: we currently rotate only after MaxUse+MaxFilesize has been reached. | |
1197 | - journal: deal nicely with byte-by-byte copied files, especially regards header | |
b44be3ec | 1198 | - journal: sanely deal with entries which are larger than the individual file size, but where the components would fit |
601d9d6f | 1199 | - Replace utmp, wtmp, btmp, and lastlog completely with journal |
f38afcd0 | 1200 | - journalctl: instead --after-cursor= maybe have a --cursor=XYZ+1 syntax? |
f38afcd0 LP |
1201 | - when a kernel driver logs in a tight loop, we should ratelimit that too. |
1202 | - journald: optionally, log debug messages to /run but everything else to /var | |
1203 | - journald: when we drop syslog messages because the syslog socket is | |
1204 | full, make sure to write how many messages are lost as first thing | |
1205 | to syslog when it works again. | |
279f0366 LP |
1206 | - journald: allow per-priority and per-service retention times when rotating/vacuuming |
1207 | - journald: make use of uid-range.h to managed uid ranges to split | |
1208 | journals in. | |
1209 | - journalctl: add the ability to look for the most recent process of a binary. journalctl /usr/bin/X11 --pid=-1 or so... | |
1210 | - improve journalctl performance by loading journal files | |
1211 | lazily. Encode just enough information in the file name, so that we | |
1212 | do not have to open it to know that it is not interesting for us, for | |
1213 | the most common operations. | |
e25b5a8d | 1214 | - man: document that corrupted journal files is nothing to act on |
e25b5a8d DH |
1215 | - rework journald sigbus stuff to use mutex |
1216 | - Set RLIMIT_NPROC for systemd-journal-xyz, and all other of our | |
1217 | services that run under their own user ids, and use User= (but only | |
1218 | in a world where userns is ubiquitous since otherwise we cannot | |
1219 | invoke those daemons on the host AND in a container anymore). Also, | |
1220 | if LimitNPROC= is used without User= we should warn and refuse | |
1221 | operation. | |
1222 | - journalctl --verify: don't show files that are currently being | |
1223 | written to as FAIL, but instead show that their are being written to. | |
1224 | - add journalctl -H that talks via ssh to a remote peer and passes through | |
1225 | binary logs data | |
e25b5a8d | 1226 | - add a version of --merge which also merges /var/log/journal/remote |
e25b5a8d DH |
1227 | - journalctl: -m should access container journals directly by enumerating |
1228 | them via machined, and also watch containers coming and going. | |
1229 | Benefit: nspawn --ephemeral would start working nicely with the journal. | |
1230 | - assign MESSAGE_ID to log messages about failed services | |
06847d0f | 1231 | - check if loop in decompress_blob_xz() is necessary |
b44be3ec | 1232 | |
07eabc2b LB |
1233 | * journald: support RFC3164 fully for the incoming syslog transport, see |
1234 | https://github.com/systemd/systemd/issues/19251#issuecomment-816601955 | |
1235 | ||
1236 | * Hook up journald's FSS logic with TPM2: seal the verification disk by | |
1237 | time-based policy, so that the verification key can remain on host and ve | |
1238 | validated via TPM. | |
1239 | ||
1240 | * build short web pages out of each catalog entry, build them along with man | |
1241 | pages, and include hyperlinks to them in the journal output | |
1242 | ||
1243 | * journald: do journal file writing out-of-process, with one writer process per | |
1244 | client UID, so that synthetic hash table collisions can slow down a specific | |
1245 | user's journal stream down but not the others. | |
1246 | ||
1247 | * tweak journald context caching. In addition to caching per-process attributes | |
1248 | keyed by PID, cache per-cgroup attributes (i.e. the various xattrs we read) | |
1249 | keyed by cgroup path, and guarded by ctime changes. This should provide us | |
1250 | with a nice speed-up on services that have many processes running in the same | |
1251 | cgroup. | |
1252 | ||
1253 | * maybe add call sd_journal_set_block_timeout() or so to set SO_SNDTIMEO for | |
1254 | the sd-journal logging socket, and, if the timeout is set to 0, sets | |
1255 | O_NONBLOCK on it. That way people can control if and when to block for | |
1256 | logging. | |
1257 | ||
1258 | * journalctl: make sure -f ends when the container indicated by -M terminates | |
1259 | ||
1260 | * journald: sigbus API via a signal-handler safe function that people may call | |
1261 | from the SIGBUS handler | |
1262 | ||
590171d1 ZJS |
1263 | * add a test if all entries in the catalog are properly formatted. |
1264 | (Adding dashes in a catalog entry currently results in the catalog entry | |
1265 | being silently skipped. journalctl --update-catalog must warn about this, | |
1266 | and we should also have a unit test to check that all our message are OK.) | |
1267 | ||
07eabc2b LB |
1268 | * homed: |
1269 | - when user tries to log into record signed by unrecognized key, automatically add key to our chain after polkit auth | |
1270 | - rollback when resize fails mid-operation | |
1271 | - GNOME's side for forget key on suspend (requires rework so that lock screen runs outside of uid) | |
07eabc2b LB |
1272 | - update LUKS password on login if we find there's a password that unlocks the JSON record but not the LUKS device. |
1273 | - create on activate? | |
1274 | - properties: icon url?, preferred session type?, administrator bool (which translates to 'wheel' membership)?, address?, telephone?, vcard?, samba stuff?, parental controls? | |
1275 | - communicate clearly when usb stick is safe to remove. probably involves | |
1276 | beefing up logind to make pam session close hook synchronous and wait until | |
1277 | systemd --user is shut down. | |
1278 | - logind: maybe keep a "busy fd" as long as there's a non-released session around or the user@.service | |
1279 | - maybe make automatic, read-only, time-based reflink-copies of LUKS disk | |
1280 | images (and btrfs snapshots of subvolumes) (think: time machine) | |
1281 | - distinguish destroy / remove (i.e. currently we can unregister a user, unregister+remove their home directory, but not just remove their home directory) | |
1282 | - in systemd's PAMName= logic: query passwords with ssh-askpassword, so that we can make "loginctl set-linger" mode work | |
1283 | - fingerprint authentication, pattern authentication, … | |
1284 | - make sure "classic" user records can also be managed by homed | |
1285 | - make size of $XDG_RUNTIME_DIR configurable in user record | |
1286 | - query password from kernel keyring first | |
1287 | - update even if record is "absent" | |
07eabc2b LB |
1288 | - move acct mgmt stuff from pam_systemd_home to pam_systemd? |
1289 | - when "homectl --pkcs11-token-uri=" is used, synthesize ssh-authorized-keys records for all keys we have private keys on the stick for | |
1290 | - make slice for users configurable (requires logind rework) | |
1291 | - logind: populate auto-login list bus property from PKCS#11 token | |
1292 | - when determining state of a LUKS home directory, check DM suspended sysfs file | |
a11e7c0b LB |
1293 | - when homed is in use, maybe start the user session manager in a mount namespace with MS_SLAVE, |
1294 | so that mounts propagate down but not up - eg, user A setting up a backup volume | |
1295 | doesn't mean user B sees it | |
9c53de8b | 1296 | - use credentials logic/TPM2 logic to store homed signing key |
9c53de8b LP |
1297 | - permit multiple user record signing keys to be used locally, and pick |
1298 | the right one for signing records automatically depending on a pre-existing | |
1299 | signature | |
1300 | - add a way to "adopt" a home directory, i.e. strip foreign signatures | |
1301 | and insert a local signature instead. | |
1302 | - as an extension to the directory+subvolume backend: if located on | |
1303 | especially marked fs, then sync down password into LUKS header of that fs, | |
1304 | and always verify passwords against it too. Bootstrapping is a problem | |
1305 | though: if no one is logged in (or no other user even exists yet), how do you | |
1306 | unlock the volume in order to create the first user and add the first pw. | |
1307 | - support new FS_IOC_ADD_ENCRYPTION_KEY ioctl for setting up fscrypt | |
1308 | - maybe pre-create ~/.cache as subvol so that it can have separate quota | |
1309 | easily? | |
9c53de8b LP |
1310 | - add a switch to homectl (maybe called --first-boot) where it will check if |
1311 | any non-system users exist, and if not prompts interactively for basic user | |
1312 | info, mimicking systemd-firstboot. Then, place this in a service that runs | |
1313 | after systemd-homed, but before gdm and friends, as a simple, barebones | |
1314 | fallback logic to get a regular user created on uninitialized systems. | |
1315 | - store PKCS#11 + FIDO2 token info in LUKS2 header, compatible with | |
1316 | systemd-cryptsetup, so that it can unlock homed volumes | |
6d975fe7 LP |
1317 | - maybe make all *.home files owned by `systemd-home` user or so, so that we |
1318 | can easily set overall quota for all users | |
1319 | - on login, if we can't fallocate initially, but rebalance is on, then allow | |
1320 | login in discard mode, then immediately rebalance, then turn off discard | |
07eabc2b | 1321 | |
07eabc2b LB |
1322 | * add a new switch --auto-definitions=yes/no or so to systemd-repart. If |
1323 | specified, synthesize a definition automatically if we can: enlarge last | |
1324 | partition on disk, but only if it is marked for growing and not read-only. | |
1325 | ||
1326 | * systemd-repart: read LUKS encryption key from $CREDENTIALS_PATH | |
1327 | ||
1328 | * systemd-repart: add a switch to factory reset the partition table without | |
1329 | immediately applying the new configuration again. i.e. --factory-reset=leave | |
1330 | or so. (this is useful to factory reset an image, then putting it into | |
1331 | another machine, ensuring that luks key is generated on new machine, not old) | |
1332 | ||
1333 | * systemd-repart: support setting up dm-integrity with HMAC | |
1334 | ||
1335 | * systemd-repart: maybe remove half-initialized image on failure. It fails | |
1336 | if the output file exists, so a repeated invocation will usually fail if | |
1337 | something goes wrong on the way. | |
1338 | ||
1339 | * systemd-repart: drop pager mode on normal operation? | |
1340 | ||
1341 | * systemd-repart: by default generate minimized partition tables (i.e. tables | |
1342 | that only cover the space actually used, excluding any free space at the | |
1343 | end), in order to maximize dd'ability. Requires libfdisk work, see | |
1344 | https://github.com/karelzak/util-linux/issues/907 | |
1345 | ||
1346 | * systemd-repart: MBR partition table support. Care needs to be taken regarding | |
1347 | Type=, so that partition definitions can sanely apply to both the GPT and the | |
1348 | MBR case. Idea: accept syntax "Type=gpt:home mbr:0x83" for setting the types | |
1349 | for the two partition types explicitly. And provide an internal mapping so | |
1350 | that "Type=linux-generic" maps to the right types for both partition tables | |
1351 | automatically. | |
1352 | ||
1353 | * systemd-repart: allow sizing partitions as factor of available RAM, so that | |
1354 | we can reasonably size swap partitions for hibernation. | |
1355 | ||
1356 | * systemd-repart: allow boolean option that ensures that if existing partition | |
1357 | doesn't exist within the configured size bounds the whole command fails. This | |
1358 | is useful to implement ESP vs. XBOOTLDR schemes in installers: have one set | |
1359 | of repart files for the case where ESP is large enough and one where it isn't | |
1360 | and XBOOTLDR is added in instead. Then apply the former first, and if it | |
1361 | fails to apply use the latter. | |
1362 | ||
1363 | * systemd-repart: add per-partition option to never reuse existing partition | |
1364 | and always create anew even if matching partition already exists. | |
1365 | ||
1366 | * systemd-repart: add per-partition option to fail if partition already exist, | |
1367 | i.e. is not added new. Similar, add option to fail if partition does not exist yet. | |
1368 | ||
1369 | * systemd-repart: allow disabling growing of specific partitions, or making | |
1370 | them (think ESP: we don't ever want to grow it, since we cannot resize vfat) | |
1371 | ||
1372 | * systemd-repart: make it a static checker during early boot for existence and | |
1373 | absence of other partitions for trusted boot environments | |
1374 | ||
b44be3ec | 1375 | * document: |
8b8f2591 | 1376 | - document that deps in [Unit] sections ignore Alias= fields in |
b44be3ec LP |
1377 | [Install] units of other units, unless those units are disabled |
1378 | - man: clarify that time-sync.target is not only sysv compat but also useful otherwise. Same for similar targets | |
b44be3ec | 1379 | - document that service reload may be implemented as service reexec |
f38afcd0 LP |
1380 | - add a man page containing packaging guidelines and recommending usage of things like Documentation=, PrivateTmp=, PrivateNetwork= and ReadOnlyDirectories=/etc /usr. |
1381 | - document systemd-journal-flush.service properly | |
f38afcd0 LP |
1382 | - documentation: recommend to connect the timer units of a service to the service via Also= in [Install] |
1383 | - man: document the very specific env the shutdown drop-in tools live in | |
5cf821ac ZJS |
1384 | - man: add more examples to man pages, |
1385 | - in particular an example how to do the equivalent of switching runlevels | |
f38afcd0 | 1386 | - man: maybe sort directives in man pages, and take sections from --help and apply them to man too |
17ec531f | 1387 | - document root=gpt-auto properly |
b44be3ec LP |
1388 | |
1389 | * systemctl: | |
b44be3ec LP |
1390 | - add systemctl switch to dump transaction without executing it |
1391 | - Add a verbose mode to "systemctl start" and friends that explains what is being done or not done | |
1392 | - "systemctl disable" on a static unit prints no message and does | |
1393 | nothing. "systemctl enable" does nothing, and gives a bad message | |
1394 | about it. Should fix both to print nice actionable messages. | |
1395 | - print nice message from systemctl --failed if there are no entries shown, and hook that into ExecStartPre of rescue.service/emergency.service | |
1396 | - add new command to systemctl: "systemctl system-reexec" which reexecs as many daemons as virtually possible | |
d28315e4 | 1397 | - systemctl enable: fail if target to alias into does not exist? maybe show how many units are enabled afterwards? |
b44be3ec | 1398 | - systemctl: "Journal has been rotated since unit was started." message is misleading |
61233823 | 1399 | - systemctl status output should include list of triggering units and their status |
f38afcd0 | 1400 | |
07eabc2b LB |
1401 | * introduce an option (or replacement) for "systemctl show" that outputs all |
1402 | properties as JSON, similar to busctl's new JSON output. In contrast to that | |
1403 | it should skip the variant type string though. | |
8b04b925 | 1404 | |
07eabc2b LB |
1405 | * add an explicit "vertical" mode to format-table, so that "systemctl |
1406 | status"-like outputs (i.e. with a series of field names left and values | |
1407 | right) become genuine first class citizens, and we gain automatic, sane JSON | |
1408 | output for them. | |
d2e83c23 | 1409 | |
07eabc2b LB |
1410 | * Add a "systemctl list-units --by-slice" mode or so, which rearranges the |
1411 | output of "systemctl list-units" slightly by showing the tree structure of | |
1412 | the slices, and the units attached to them. | |
a19554ed | 1413 | |
07eabc2b LB |
1414 | * add "systemctl wait" or so, which does what "systemd-run --wait" does, but |
1415 | for all units. It should be both a way to pin units into memory as well as a | |
1416 | wait to retrieve their exit data. | |
a7a3f28b | 1417 | |
07eabc2b LB |
1418 | * show whether a service has out-of-date configuration in "systemctl status" by |
1419 | using mtime data of ConfigurationDirectory=. | |
08f95888 | 1420 | |
07eabc2b LB |
1421 | * "systemctl preset-all" should probably order the unit files it |
1422 | operates on lexicographically before starting to work, in order to | |
1423 | ensure deterministic behaviour if two unit files conflict (like DMs | |
1424 | do, for example) | |
dcfc4b2e | 1425 | |
07eabc2b LB |
1426 | * add "systemctl start -v foobar.service" that shows logs of a service |
1427 | while the start command runs. This is non-trivial to do without | |
1428 | races though, since we should flush out all journal messages before | |
1429 | returning from the "systemctl stop". | |
71ef24d0 | 1430 | |
07eabc2b LB |
1431 | * systemctl: if some operation fails, show log output? |
1432 | ||
1433 | * Add a new verb "systemctl top" | |
1434 | ||
1435 | * unit install: | |
1436 | - "systemctl mask" should find all names by which a unit is accessible | |
1437 | (i.e. by scanning for symlinks to it) and link them all to /dev/null | |
1b89884b | 1438 | |
b44be3ec | 1439 | * nspawn: |
e25b5a8d DH |
1440 | - emulate /dev/kmsg using CUSE and turn off the syslog syscall |
1441 | with seccomp. That should provide us with a useful log buffer that | |
1442 | systemd can log to during early boot, and disconnect container logs | |
1443 | from the kernel's logs. | |
1444 | - as soon as networkd has a bus interface, hook up --network-interface=, | |
1445 | --network-bridge= with networkd, to trigger netdev creation should an | |
1446 | interface be missing | |
e25b5a8d DH |
1447 | - a nice way to boot up without machine id set, so that it is set at boot |
1448 | automatically for supporting --ephemeral. Maybe hash the host machine id | |
1449 | together with the machine name to generate the machine id for the container | |
1450 | - fix logic always print a final newline on output. | |
1451 | https://github.com/systemd/systemd/pull/272#issuecomment-113153176 | |
1452 | - should optionally support receiving WATCHDOG=1 messages from its payload | |
1453 | PID 1... | |
e25b5a8d DH |
1454 | - optionally automatically add FORWARD rules to iptables whenever nspawn is |
1455 | running, remove them when shut down. | |
e25b5a8d | 1456 | |
07eabc2b LB |
1457 | * nspawn: add support for sysext extensions, too. i.e. a new --extension= |
1458 | switch that takes one or more arguments, and applies the extensions already | |
1459 | during startup. | |
1460 | ||
1461 | * when main nspawn supervisor process gets suspended due to SIGSTOP/SIGTTOU or | |
1462 | so, freeze the payload too. | |
1463 | ||
1464 | * machined: add API to acquire UID range. add API to mount/dissect loopback | |
1465 | file. Both protected by PK. Then make nspawn use these APIs to run | |
1466 | unprivileged containers. i.e. push the truly privileged bits into machined, | |
1467 | so that the client side can remain entirely unprivileged, with SUID or | |
1468 | anything like that. | |
1469 | ||
1470 | * nspawn: support time namespaces | |
1471 | ||
1472 | * nspawn: on cgroupsv1 issue cgroup empty handler process based on host events, | |
1473 | so that we make cgroup agent logic safe | |
1474 | ||
1475 | * nspawn/machined: add API to invoke binary in container, then use that as | |
1476 | fallback in "machinectl shell" | |
1477 | ||
1478 | * nspawn: make nspawn suitable for shell pipelines: instead of triggering a | |
1479 | hangup when input is finished, send ^D, which synthesizes an EOF. Then wait | |
1480 | for hangup or ^D before passing on the EOF. | |
1481 | ||
1482 | * nspawn: greater control over selinux label? | |
1483 | ||
1484 | * nspawn: support that /proc, /sys/, /dev are pre-mounted | |
1485 | ||
e25b5a8d | 1486 | * machined: |
e25b5a8d DH |
1487 | - add an API so that libvirt-lxc can inform us about network interfaces being |
1488 | removed or added to an existing machine | |
1489 | - "machinectl migrate" or similar to copy a container from or to a | |
1490 | difference host, via ssh | |
e25b5a8d DH |
1491 | - introduce systemd-nspawn-ephemeral@.service, and hook it into |
1492 | "machinectl start" with a new --ephemeral switch | |
1493 | - "machinectl status" should also show internal logs of the container in | |
1494 | question | |
e25b5a8d DH |
1495 | - "machinectl history" |
1496 | - "machinectl diff" | |
1497 | - "machinectl commit" that takes a writable snapshot of a tree, invokes a | |
1498 | shell in it, and marks it read-only after use | |
1499 | ||
abd55b16 | 1500 | * udev: |
abd55b16 | 1501 | - move to LGPL |
abd55b16 KS |
1502 | - kill scsi_id |
1503 | - add trigger --subsystem-match=usb/usb_device device | |
e8d842a0 | 1504 | - reimport udev db after MOVE events for devices without dev_t |
b8217b7b | 1505 | |
e25b5a8d DH |
1506 | * coredump: |
1507 | - save coredump in Windows/Mozilla minidump format | |
73a99163 | 1508 | - 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 |
1509 | |
1510 | * support crash reporting operation modes (https://live.gnome.org/GnomeOS/Design/Whiteboards/ProblemReporting) | |
1511 | ||
f38afcd0 | 1512 | * tmpfiles: |
f38afcd0 | 1513 | - apply "x" on "D" too (see patch from William Douglas) |
614cc34f | 1514 | - instead of ignoring unknown fields, reject them. |
e25b5a8d DH |
1515 | - creating new directories/subvolumes/fifos/device nodes |
1516 | should not follow symlinks. None of the other adjustment or creation | |
1517 | calls follow symlinks. | |
36f57e02 | 1518 | - add --test mode |
ba405b22 ZJS |
1519 | - teach tmpfiles.d q/Q logic something sensible in the context of XFS/ext4 |
1520 | project quota | |
1258097c | 1521 | |
af6f0d42 TG |
1522 | * udev-link-config: |
1523 | - Make sure ID_PATH is always exported and complete for | |
1524 | network devices where possible, so we can safely rely | |
1525 | on Path= matching | |
af6f0d42 | 1526 | |
88e4d1d7 | 1527 | * sd-rtnl: |
88e4d1d7 | 1528 | - add support for more attribute types |
c589a0e6 | 1529 | - inbuilt piping support (essentially degenerate async)? see loopback-setup.c and other places |
88e4d1d7 | 1530 | |
0a4b9a07 | 1531 | * networkd: |
c74ecd71 TG |
1532 | - add more keys to [Route] and [Address] sections |
1533 | - add support for more DHCPv4 options (and, longer term, other kinds of dynamic config) | |
e8d842a0 | 1534 | - add reduced [Link] support to .network files |
798e174a | 1535 | - properly handle routerless dhcp leases |
a8eaaee7 | 1536 | - work with non-Ethernet devices |
e25b5a8d | 1537 | - dhcp: do we allow configuring dhcp routes on interfaces that are not the one we got the dhcp info from? |
e25b5a8d DH |
1538 | - the DHCP lease data (such as NTP/DNS) is still made available when |
1539 | a carrier is lost on a link. It should be removed instantly. | |
1540 | - expose in the API the following bits: | |
e2da6491 | 1541 | - option 15, domain name |
38b38500 | 1542 | - option 12, hostname and/or option 81, fqdn |
e25b5a8d DH |
1543 | - option 123, 144, geolocation |
1544 | - option 252, configure http proxy (PAC/wpad) | |
1545 | - provide a way to define a per-network interface default metric value | |
1546 | for all routes to it. possibly a second default for DHCP routes. | |
1547 | - allow Name= to be specified repeatedly in the [Match] section. Maybe also | |
1548 | support Name=foo*|bar*|baz ? | |
e25b5a8d | 1549 | - whenever uplink info changes, make DHCP server send out FORCERENEW |
155e8b9a | 1550 | |
07eabc2b LB |
1551 | * in networkd, when matching device types, fix up DEVTYPE rubbish the kernel passes to us |
1552 | ||
d5e172d2 ZJS |
1553 | * Figure out how to do unittests of networkd's state serialization |
1554 | ||
ac976532 | 1555 | * dhcp: |
424a8732 | 1556 | - figure out how much we can increase Maximum Message Size |
ac976532 | 1557 | |
37d8b536 PF |
1558 | * dhcp6: |
1559 | - add functions to set previously stored IPv6 addresses on startup and get | |
1560 | them at shutdown; store them in client->ia_na | |
1561 | - write more test cases | |
37d8b536 | 1562 | - implement reconfigure support, see 5.3., 15.11. and 22.20. |
37d8b536 | 1563 | - implement support for temporary adressess (IA_TA) |
37d8b536 PF |
1564 | - implement dhcpv6 authentication |
1565 | - investigate the usefulness of Confirm messages; i.e. are there any | |
1566 | situations where the link changes without any loss in carrier detection | |
1567 | or interface down | |
1568 | - some servers don't do rapid commit without a filled in IA_NA, verify | |
1569 | this behavior | |
4a77c53d | 1570 | - RouteTable= ? |