]> git.ipfire.org Git - thirdparty/systemd.git/log
thirdparty/systemd.git
5 weeks agohwdb/60-tpm2: correct prefix in comment (tpm -> tpm2)
Mike Yuan [Thu, 5 Mar 2026 00:02:41 +0000 (01:02 +0100)] 
hwdb/60-tpm2: correct prefix in comment (tpm -> tpm2)

Follow-up for f2eed3fa25e8c38b7a90d6ab3d22ee90e3569271

5 weeks agojournalctl-filter: drop doubled newline
Mike Yuan [Wed, 4 Mar 2026 21:33:47 +0000 (22:33 +0100)] 
journalctl-filter: drop doubled newline

Follow-up for d8302c2fd92602eae780511037ca08ed8cb0667d

5 weeks agoTranslations update from Fedora Weblate (#40952)
Luca Boccassi [Wed, 4 Mar 2026 20:11:59 +0000 (20:11 +0000)] 
Translations update from Fedora Weblate (#40952)

Translations update from [Fedora
Weblate](https://translate.fedoraproject.org) for
[systemd/main](https://translate.fedoraproject.org/projects/systemd/main/).

Current translation status:

![Weblate translation
status](https://translate.fedoraproject.org/widget/systemd/main/horizontal-auto.svg)

5 weeks agopo: Translated using Weblate (Slovenian) 40952/head
Martin Srebotnjak [Wed, 4 Mar 2026 19:58:26 +0000 (19:58 +0000)] 
po: Translated using Weblate (Slovenian)

Currently translated at 100.0% (266 of 266 strings)

Co-authored-by: Martin Srebotnjak <miles@filmsi.net>
Translate-URL: https://translate.fedoraproject.org/projects/systemd/main/sl/
Translation: systemd/main

5 weeks agopo: Translated using Weblate (Finnish)
Jan Kuparinen [Wed, 4 Mar 2026 19:58:26 +0000 (19:58 +0000)] 
po: Translated using Weblate (Finnish)

Currently translated at 100.0% (266 of 266 strings)

Co-authored-by: Jan Kuparinen <copper_fin@hotmail.com>
Translate-URL: https://translate.fedoraproject.org/projects/systemd/main/fi/
Translation: systemd/main

5 weeks agoupdate TODO
Lennart Poettering [Wed, 4 Mar 2026 16:06:23 +0000 (17:06 +0100)] 
update TODO

5 weeks agobuild(deps): bump ninja from 1.11.1.4 to 1.13.0 in /.github/workflows
dependabot[bot] [Wed, 4 Mar 2026 12:58:47 +0000 (12:58 +0000)] 
build(deps): bump ninja from 1.11.1.4 to 1.13.0 in /.github/workflows

Bumps [ninja](https://github.com/scikit-build/ninja-python-distributions) from 1.11.1.4 to 1.13.0.
- [Release notes](https://github.com/scikit-build/ninja-python-distributions/releases)
- [Changelog](https://github.com/scikit-build/ninja-python-distributions/blob/master/HISTORY.rst)
- [Commits](https://github.com/scikit-build/ninja-python-distributions/compare/1.11.1.4...1.13.0)

---
updated-dependencies:
- dependency-name: ninja
  dependency-version: 1.13.0
  dependency-type: direct:production
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <support@github.com>
5 weeks agopcrextend-util: fix log message
Lennart Poettering [Wed, 4 Mar 2026 14:18:44 +0000 (15:18 +0100)] 
pcrextend-util: fix log message

The NvPCR is actually named differently from what the log msg said. Fix it.

5 weeks agoclang-tidy fixes (#40947)
Mike Yuan [Wed, 4 Mar 2026 15:27:51 +0000 (16:27 +0100)] 
clang-tidy fixes (#40947)

5 weeks agomachine-dbus: do not check for overlapping condition 40947/head
Mike Yuan [Wed, 4 Mar 2026 13:31:54 +0000 (14:31 +0100)] 
machine-dbus: do not check for overlapping condition

Follow-up for c5e48e3a66b23313cd4931b9dc25a8f48cfb1035

This also makes things in line with the varlink method.

5 weeks agosd-journal/catalog: make catalog_file_dirs static
Mike Yuan [Wed, 4 Mar 2026 13:11:52 +0000 (14:11 +0100)] 
sd-journal/catalog: make catalog_file_dirs static

5 weeks agobpf-dlopen: mark bpf_get_error() as static
Mike Yuan [Wed, 4 Mar 2026 13:09:03 +0000 (14:09 +0100)] 
bpf-dlopen: mark bpf_get_error() as static

Any use of this function should go via bpf_get_error_translated().

5 weeks agoshared/password-quality-util-*: make dl functions static
Mike Yuan [Wed, 4 Mar 2026 12:12:41 +0000 (13:12 +0100)] 
shared/password-quality-util-*: make dl functions static

5 weeks agomachined: move declaration of manager_object to machined-dbus.h
Mike Yuan [Wed, 4 Mar 2026 12:49:59 +0000 (13:49 +0100)] 
machined: move declaration of manager_object to machined-dbus.h

5 weeks agovarious: make bus objects static
Mike Yuan [Wed, 4 Mar 2026 12:26:42 +0000 (13:26 +0100)] 
various: make bus objects static

5 weeks agocore/varlink-metrics: make metric_family_table static
Mike Yuan [Wed, 4 Mar 2026 12:11:50 +0000 (13:11 +0100)] 
core/varlink-metrics: make metric_family_table static

5 weeks agoshared/varlink-*: normalize includes
Mike Yuan [Wed, 4 Mar 2026 12:10:31 +0000 (13:10 +0100)] 
shared/varlink-*: normalize includes

5 weeks agovarious: mark arg_* as static
Mike Yuan [Wed, 4 Mar 2026 12:25:37 +0000 (13:25 +0100)] 
various: mark arg_* as static

5 weeks agobootctl: declare missing arg_* in header
Mike Yuan [Wed, 4 Mar 2026 12:25:03 +0000 (13:25 +0100)] 
bootctl: declare missing arg_* in header

Follow-up for 90cf998875a2cfac2cdfe3e659c96d25457bf24b

5 weeks agotree-wide: make private hash ops static
Mike Yuan [Wed, 4 Mar 2026 11:53:12 +0000 (12:53 +0100)] 
tree-wide: make private hash ops static

5 weeks agoshared/acpi-fpdt: set _packed_ attr properly
Mike Yuan [Wed, 4 Mar 2026 11:51:48 +0000 (12:51 +0100)] 
shared/acpi-fpdt: set _packed_ attr properly

We were declaring a variable named _packed...

5 weeks agomeson: bump version to v260~rc2 v260-rc2
Luca Boccassi [Wed, 4 Mar 2026 12:24:14 +0000 (12:24 +0000)] 
meson: bump version to v260~rc2

5 weeks agoNEWS: update contributors list
Luca Boccassi [Wed, 4 Mar 2026 13:19:11 +0000 (13:19 +0000)] 
NEWS: update contributors list

5 weeks agoNEWS: finalize place and date
Luca Boccassi [Wed, 4 Mar 2026 12:23:59 +0000 (12:23 +0000)] 
NEWS: finalize place and date

5 weeks agomachined: do not allow unprivileged users to shell into the root namespace via varlink
Luca Boccassi [Sat, 28 Feb 2026 00:46:21 +0000 (00:46 +0000)] 
machined: do not allow unprivileged users to shell into the root namespace via varlink

Forbid non-root from shelling into a machine that is running in
the root user namespace.

Follow-up for adaff8eb35d9c471af81fddaa4403bc5843a256f

5 weeks agotest: don't register short-living containers with machined, again
Frantisek Sumsal [Wed, 4 Mar 2026 11:32:15 +0000 (12:32 +0100)] 
test: don't register short-living containers with machined, again

Otherwise we might try to register the same scope again before the
previous instance gets a chance to be cleaned up:

[   54.378392] systemd-nspawn[2554]: ░ Spawning container TEST-13-NSPAWN.defaultinaccessiblepaths.nxs on /var/lib/machines/TEST-13-NSPAWN.default_inaccessible_paths.nxs.
[   54.382202] systemd-nspawn[2554]: Failed to allocate scope: Unit TEST-13-NSPAWN.defaultinaccessiblepaths.nxs.scope was already loaded or has a fragment file.
[   54.411211] systemd[1]: TEST-13-NSPAWN.service: Main process exited, code=exited, status=1/FAILURE
[   54.411413] systemd[1]: TEST-13-NSPAWN.service: Failed with result 'exit-code'.
[   54.411885] systemd[1]: Failed to start TEST-13-NSPAWN.service - TEST-13-NSPAWN.

This is basically the same change as in
6a05abb9b49900774bc0323316103dceab0c1a7d but for the newly added tests.

Follow-up for 83b8daa032cd0adb538cfd9467e6acf2c44aa661.

Resolves: #40945

5 weeks agonetwork: Rename ModemManager .network section WRT tech, not project...
Oleksandr Andrushchenko [Mon, 2 Mar 2026 15:01:31 +0000 (17:01 +0200)] 
network: Rename ModemManager .network section WRT tech, not project...

and use dedicated knobs for every option used in
former SimpleConnectProperties.
New section is [MobileNetwork] with the following configuration options:

APN=
AllowedAuthenticationMechanisms==
User=
Password=
IPFamily=
AllowRoaming=
PIN=
OperatorId=

5 weeks agoportable: Add ExtensionImages drop-in for any extension (#40911)
Zbigniew Jędrzejewski-Szmek [Wed, 4 Mar 2026 10:44:00 +0000 (11:44 +0100)] 
portable: Add ExtensionImages drop-in for any extension (#40911)

The diff seems big, but it's just removing and if and deleting the
indentation. With `git show -w` you can see clearly that is just one
line. In the files tab, in github, you can click "hide whitespaces" and
it is clear too.

I don't know if there is some better way, but when using a portable
service with an extension image that just overlays files I need to
manually add the `ExtensionImages=`. And that is added automatically for
other extension images that provide a service unit.

So, this patch just makes sure we add the `ExtensionImages=` for any
extension.

Any thoughts?

Below the commit msg for more details:

---

    Before this patch, when running:

            portablectl attach --extension ext.raw ./base.raw

No drop-in is added for the "ExtensionImages" if there aren't units from
    the extension loaded.

But the extension can just overlay files, as in my case. So before this
patch, I also need to manually add a drop-in with "ExtensionImages=" for
    it to really be loaded.

Let's just always add the drop-in for extensions. This way, it works for
    extensions that just overlay files too.

Please note this commit just removes the if (simpler to view the diff
    with git show -w). Also, the if checked for m->image_path being not
NULL, but removing it shouldn't cause a NULL pointer dereference.
Because
m->image_path is not used inside the if (it was needed just for the if
itself) and image_path is asserted at the beginning of the function to
    be non-NULL too.

    This was like this since the beginning of time in 907952bbc9
    ("portabled: add --extension parameter for layered images support")

5 weeks agotpm2: add hwdb quirk db, to deal with borked TPMs (#40931)
Lennart Poettering [Wed, 4 Mar 2026 08:54:10 +0000 (09:54 +0100)] 
tpm2: add hwdb quirk db, to deal with borked TPMs (#40931)

This adds infrastructure so that we can maintain quirks for TPM devices
in hwdb, to deal with issues such as #40485.

I expected that we need this sooner or later anyway, so here we go.

Fixes: #40485
5 weeks agoupdate NEWS 40931/head
Lennart Poettering [Tue, 3 Mar 2026 20:24:51 +0000 (21:24 +0100)] 
update NEWS

5 weeks agotpm2-setup: don't fail service on two more types of failures
Lennart Poettering [Tue, 3 Mar 2026 20:51:38 +0000 (21:51 +0100)] 
tpm2-setup: don't fail service on two more types of failures

Let's bubble up failures all the way until they reach the services, but
then let's carefully gracefully handle some of them, that are about
issues not immediately actionable to the admin, even if they are
potentially quite problematic.

5 weeks agotpm2-setup: use symbolic exit code
Lennart Poettering [Tue, 3 Mar 2026 20:50:49 +0000 (21:50 +0100)] 
tpm2-setup: use symbolic exit code

76 is the bsd exit code EX_PROTOCOL, which is kinda fitting here. Let#s
hence use the symbolic exit code here.

5 weeks agotest: add superficial testcase for tpm2 identification
Lennart Poettering [Tue, 3 Mar 2026 20:20:13 +0000 (21:20 +0100)] 
test: add superficial testcase for tpm2 identification

5 weeks agotpm2-util: check udev db to determine if NvPCRs are going to work
Lennart Poettering [Tue, 3 Mar 2026 14:59:40 +0000 (15:59 +0100)] 
tpm2-util: check udev db to determine if NvPCRs are going to work

Fixes: #40485
5 weeks agotpm2-util: remove strjoina() usage on user-controlled data
Lennart Poettering [Tue, 3 Mar 2026 14:53:09 +0000 (15:53 +0100)] 
tpm2-util: remove strjoina() usage on user-controlled data

5 weeks agohwdb: introduce hwdb for tpm2 devices
Lennart Poettering [Tue, 3 Mar 2026 13:06:44 +0000 (14:06 +0100)] 
hwdb: introduce hwdb for tpm2 devices

This hwdb is can carry hw quirks and similar for us, in particular tell
us if nvpcrs have a chance of working.

5 weeks agoudev: add tpm2_id builtin
Lennart Poettering [Tue, 3 Mar 2026 11:32:11 +0000 (12:32 +0100)] 
udev: add tpm2_id builtin

We need to be able to look up tpm2 metadata from hwdb, hence add a way
to synthesize a whdb lookup key from with udev rules.

5 weeks agotpm2-util,analyze: add helper for generating hwdb lookup key from TPM2 vendor data
Lennart Poettering [Tue, 3 Mar 2026 11:28:45 +0000 (12:28 +0100)] 
tpm2-util,analyze: add helper for generating hwdb lookup key from TPM2 vendor data

Our goal is to find TPM2 metadata in hwdb, hence let's compile a
"modalias"-style string from the TPM2 metadata, we can use as hwdb
lookup key.

5 weeks agoanalyze: add "identify-tpm2" command that shows TPM2 chip information
Lennart Poettering [Tue, 3 Mar 2026 11:28:22 +0000 (12:28 +0100)] 
analyze: add "identify-tpm2" command that shows TPM2 chip information

5 weeks agotpm2-util: add tpm2_get_vendor_info() helper for getting TPM2 vendor info from the...
Lennart Poettering [Tue, 3 Mar 2026 11:27:55 +0000 (12:27 +0100)] 
tpm2-util: add tpm2_get_vendor_info() helper for getting TPM2 vendor info from the device

5 weeks agotpm2-util: also load libtss2-tcti-device.so.0 in dlopen_tpm2()
Lennart Poettering [Tue, 3 Mar 2026 13:27:36 +0000 (14:27 +0100)] 
tpm2-util: also load libtss2-tcti-device.so.0 in dlopen_tpm2()

This TCTI module is the one we need to actually access a Linux TPM
device, we'll hence pretty much always need it if we do TPM at all.
Given that we nowadays turn off dlopen() after fork() in the child,
let's explicitly load it as part of dlopen_tpm2() so that it is
available whenever TPM2 is used.

5 weeks agoparse_hwdb: sort general matches
Lennart Poettering [Tue, 3 Mar 2026 20:09:27 +0000 (21:09 +0100)] 
parse_hwdb: sort general matches

5 weeks agoparse_hwdb: introduce local variable for boolean syntax
Lennart Poettering [Tue, 3 Mar 2026 20:07:57 +0000 (21:07 +0100)] 
parse_hwdb: introduce local variable for boolean syntax

5 weeks agotpm2-util: mark two functions as static that are not used outside of tpm2-util.c
Lennart Poettering [Tue, 3 Mar 2026 15:00:00 +0000 (16:00 +0100)] 
tpm2-util: mark two functions as static that are not used outside of tpm2-util.c

5 weeks agoudevadm: fix --help text for udevadm test-builtin
Lennart Poettering [Tue, 3 Mar 2026 13:28:46 +0000 (14:28 +0100)] 
udevadm: fix --help text for udevadm test-builtin

5 weeks agoansi-color: in 256 mode, always set the fallback color first
Mike Yuan [Sun, 1 Mar 2026 13:20:53 +0000 (14:20 +0100)] 
ansi-color: in 256 mode, always set the fallback color first

Linux console is very weird when it comes to ANSI color sequences.
Not only that it isn't aware of ':' separator (c.f.
https://github.com/systemd/systemd/pull/40878#issuecomment-3979826739),
it even skips the whole CSI-m sequence if it contains parts it cannot
parse. Hence when color mode is set to 256 (i.e. default when no
extra info is available) let's always emit two distinct CSI-m sequences,
and set the fallback 16 color first in case the terminal doesn't have
complete support for the 256 one.

Replaces #40905

5 weeks agotree-wide: use ALPHANUMERICAL where appropriate
Mike Yuan [Tue, 3 Mar 2026 21:24:39 +0000 (22:24 +0100)] 
tree-wide: use ALPHANUMERICAL where appropriate

Prompted by 43116c56621c7317852c01e67dbcaa78b5e7ff70

5 weeks agoudev: rules: improve usb integration detection
David Santamaría Rogado [Sun, 1 Mar 2026 18:55:27 +0000 (19:55 +0100)] 
udev: rules: improve usb integration detection

usb hubs tend to expose removable attribute as unknown. This makes some
problems like a hub for external usb ports in pogo pins is unknown and
also soldered hubs in laptops for keyboard+touchpad.

Let's set internal when the device removable attribute is fixed and
external when removable, but when it's unknown lets check the parent
ports (not the host devpath!=0) attribute to decide.

This makes us to missdetect pogo ping connected external usb hubs but
let us to correctly detect laptop internal keyboards and touchpads that
are wired through hubs instead directly.

This behaviour is more desirable, as actually there are a bunch of
laptops with this setup.

Fixes: a4381cae8bfacb1160967ac499c2919da7ff8c2b.
5 weeks agovmspawn: change order of fields in --extra-drive=
Zbigniew Jędrzejewski-Szmek [Tue, 3 Mar 2026 15:32:29 +0000 (16:32 +0100)] 
vmspawn: change order of fields in --extra-drive=

Closes #40877. As requested, --extra-drive=path[:format] is changed
to --extra-drive=[format:]path, so that the parsing is less ambiguous.
(In the original request, it was requested that the empty format can be
used also, but that was dropped in the second version of the patch.)

5 weeks agoChores for RC2 (#40939)
Luca Boccassi [Tue, 3 Mar 2026 21:39:43 +0000 (21:39 +0000)] 
Chores for RC2 (#40939)

5 weeks agotree-wide: use DIGITS and friends
Yu Watanabe [Tue, 3 Mar 2026 13:34:41 +0000 (22:34 +0900)] 
tree-wide: use DIGITS and friends

5 weeks agoudev/scsi: use hexchar()
Yu Watanabe [Tue, 3 Mar 2026 13:34:27 +0000 (22:34 +0900)] 
udev/scsi: use hexchar()

5 weeks agosd-messages: fix typo
Lennart Poettering [Tue, 3 Mar 2026 16:02:48 +0000 (17:02 +0100)] 
sd-messages: fix typo

(This was introduced in v260, i.e. not yet released, hence not API
break)

5 weeks agomachined: add comment explaining access to machine objects a bit
Lennart Poettering [Tue, 3 Mar 2026 17:22:29 +0000 (18:22 +0100)] 
machined: add comment explaining access to machine objects a bit

5 weeks agoNEWS: move interesting items closer to top and mention PrivateTmp changes
Zbigniew Jędrzejewski-Szmek [Tue, 3 Mar 2026 17:45:29 +0000 (18:45 +0100)] 
NEWS: move interesting items closer to top and mention PrivateTmp changes

In https://bugzilla.redhat.com/show_bug.cgi?id=2443620 it was
reported that the changes to unit ordering were surprising. Let's
add a note about the PrivateTmp= handling changes.
Follow-up for https://github.com/systemd/systemd/pull/39790.

5 weeks agoNEWS: update contributors list 40939/head
Luca Boccassi [Tue, 3 Mar 2026 17:44:49 +0000 (17:44 +0000)] 
NEWS: update contributors list

5 weeks agoUpdate hwdb
Luca Boccassi [Tue, 3 Mar 2026 17:42:40 +0000 (17:42 +0000)] 
Update hwdb

ninja -C build update-hwdb

5 weeks agosystemctl-start-unit: enclose command in single quotes
Mike Yuan [Tue, 3 Mar 2026 16:50:19 +0000 (17:50 +0100)] 
systemctl-start-unit: enclose command in single quotes

5 weeks agosystemctl: rename enqueue-marked-jobs to enqueue-marked (#40930)
Zbigniew Jędrzejewski-Szmek [Tue, 3 Mar 2026 16:39:23 +0000 (17:39 +0100)] 
systemctl: rename enqueue-marked-jobs to enqueue-marked (#40930)

Closes #40883. As described in the issue, it's not "jobs" that are
marked, and also the name is unnecessarilly long.

I think we don't need any compatibility measures here. At least in the
rpm world, package upgrade scripts go through the helper which is part
of the package so the new systemctl and the new helper are upgraded
together.

5 weeks agohwdb/keyboard: Map FN key on Positron Proxima 15
Mikhail Novosyolov [Tue, 3 Mar 2026 13:57:30 +0000 (16:57 +0300)] 
hwdb/keyboard: Map FN key on Positron Proxima 15

After kernel commit 907bc9268a ("Input: atkbd - map F23 key to support default copilot shortcut")
Fn+F5 combination (switch touchpad on/off) stopped working correctly.

Fn produces F23, it is probably a bug in BIOS, ther eis no "Copilot" key.
It was ignored before that commit, but now we have to remap it here in hwdb.

This workaround is similar to systemd commit d2502f55a2d
("hwdb/keyboard: Map FN key on TUXEDO InfinityFlex 14 Gen1")

Hardware probe of this notebook: https://linux-hardware.org/?probe=7aca7ed668
See also: https://bugzilla.rosa.ru/show_bug.cgi?id=19950

5 weeks agosystemctl: rename enqueue-marked-jobs to enqueue-marked 40930/head
Zbigniew Jędrzejewski-Szmek [Tue, 3 Mar 2026 14:22:08 +0000 (15:22 +0100)] 
systemctl: rename enqueue-marked-jobs to enqueue-marked

Closes #40883. As described in the issue, it's not "jobs" that are
marked, and also the name is unnecessarilly long.

I think we don't need any compatibility measures here. At least in the
rpm world, package upgrade scripts go through the helper which is part
of the package so the new systemctl and the new helper are upgraded
together.

5 weeks agoNEWS: mention the sd_varlink_field_type_t breakage
Zbigniew Jędrzejewski-Szmek [Tue, 3 Mar 2026 14:05:31 +0000 (15:05 +0100)] 
NEWS: mention the sd_varlink_field_type_t breakage

Follow-up for 93d768e0f36a62afed7ebbf3abe3385cfd186480. The commit with
the fix didn't mention this, but the reported reproducer was:
> Install openSUSE Tumbleweed with account-utils and systemd v258.
> Compile and install systemd v260. Run "varlinkctl list-methods
> /run/account/newidmapd-socket" -> the newidmap service crashes in
> varlink_idl_format_all_fields(). Recompile newidmap with systemd v260
> headers -> varlinkctl list-methods works again.

Other people might hit the same issue, so let's mention that this was
fixed.

5 weeks agosome chaseat() love (#40825)
Zbigniew Jędrzejewski-Szmek [Tue, 3 Mar 2026 13:53:20 +0000 (14:53 +0100)] 
some chaseat() love (#40825)

5 weeks agoTODO: fix typo
Yu Watanabe [Tue, 3 Mar 2026 13:11:44 +0000 (22:11 +0900)] 
TODO: fix typo

5 weeks agoman: fix typo in docs for notify-ready option
Bouke van der Bijl [Tue, 3 Mar 2026 12:21:15 +0000 (13:21 +0100)] 
man: fix typo in docs for notify-ready option

5 weeks agoshell-completion: update run0 completion
Eisuke Kawashima [Tue, 3 Mar 2026 08:25:55 +0000 (17:25 +0900)] 
shell-completion: update run0 completion

5 weeks agopo: Translated using Weblate (Catalan)
naly zzwd [Tue, 3 Mar 2026 08:58:26 +0000 (08:58 +0000)] 
po: Translated using Weblate (Catalan)

Currently translated at 100.0% (266 of 266 strings)

Co-authored-by: naly zzwd <xeanhort007@gmail.com>
Translate-URL: https://translate.fedoraproject.org/projects/systemd/main/ca/
Translation: systemd/main

5 weeks agopo: Translated using Weblate (Greek)
Jim Spentzos [Tue, 3 Mar 2026 08:58:26 +0000 (08:58 +0000)] 
po: Translated using Weblate (Greek)

Currently translated at 100.0% (266 of 266 strings)

Co-authored-by: Jim Spentzos <jimspentzos2000@gmail.com>
Translate-URL: https://translate.fedoraproject.org/projects/systemd/main/el/
Translation: systemd/main

5 weeks agotest/portable: Ensure ExtensionImages is set for any dep 40911/head
Rodrigo Campos [Mon, 2 Mar 2026 15:02:06 +0000 (16:02 +0100)] 
test/portable: Ensure ExtensionImages is set for any dep

The previous commit made portablectl attach add a drop-in for any
extension image. Let's add a test for that too.

5 weeks agoportable: Add ExtensionImage drop-in for any extension
Rodrigo Campos [Fri, 27 Feb 2026 17:10:40 +0000 (18:10 +0100)] 
portable: Add ExtensionImage drop-in for any extension

Before this patch, when running:

portablectl attach --extension ext.raw ./base.raw

No drop-in is added for the "ExtensionImages" if there aren't units from
the extension loaded.

But the extension can just overlay files, as in my case. So before this
patch, I also need to manually add a drop-in with "ExtensionImages=" for
it to really be loaded.

Let's just always add the drop-in for extensions. This way, it works for
extensions that just overlay files too.

Please note this commit just removes the if (simpler to view the diff
with git show -w). Also, the if checked for m->image_path being not
NULL, but removing it shouldn't cause a NULL pointer dereference.
Because m->image_path is not used inside the if (it was needed just for
the if itself) and image_path is asserted at the beginning of the
function to be non-NULL too.

This was like this since the beginning of time in 907952bbc9
("portabled: add --extension parameter for layered images support")

5 weeks agoxstatx_full(): add flag to acquire STATX_MNT_ID_UNIQUE if we can, with fallback. 40825/head
Lennart Poettering [Wed, 25 Feb 2026 13:16:47 +0000 (14:16 +0100)] 
xstatx_full(): add flag to acquire STATX_MNT_ID_UNIQUE if we can, with fallback.

5 weeks agostat-util: teach statx_mount_same() STATX_MNT_ID_UNIQUE
Lennart Poettering [Wed, 25 Feb 2026 14:10:12 +0000 (15:10 +0100)] 
stat-util: teach statx_mount_same() STATX_MNT_ID_UNIQUE

So far statx_mount_same() assumed STATX_MNT_ID_UNIQUE didn't exist.
However it does exist, hence do something useful if we see it set. Note
that this creates a certain ambiguity: if we compare one statx struct
with STATX_MNT_ID_UNIQUE and one without it (but with the regular mnt
id), then we cnanot really come to a clear conclusion, hence need to
introduce a third, unknown state.

Note that we don't request STATX_MNT_ID_UNIQUE yet wherever we call
statx_mount_same(). THis will be added in a later commit.

5 weeks agomountwork: use statx_mount_same() where appropriate
Lennart Poettering [Wed, 25 Feb 2026 12:08:48 +0000 (13:08 +0100)] 
mountwork: use statx_mount_same() where appropriate

5 weeks agofd-util: minor shortcut
Lennart Poettering [Wed, 25 Feb 2026 12:08:34 +0000 (13:08 +0100)] 
fd-util: minor shortcut

5 weeks agofd-util: rename fds_are_same_mount() fds_inode_and_mount_same()
Lennart Poettering [Wed, 25 Feb 2026 12:06:32 +0000 (13:06 +0100)] 
fd-util: rename fds_are_same_mount() fds_inode_and_mount_same()

The old name suggested this would only check if the two inodes are on
the same mount, but it actually checks if they are the same inodes too.

Let's rename it to make this clearer, in particular as we have
both statx_inode_same() and statx_mount_same() already, and they are
even used here, and hence very confusing.

This also drops two checks from the test case, which are simply wrong.
Given they apparently weren't load bearing (since no CI tripped up),
let's just drop them.

5 weeks agochase: drop wrong optimization
Lennart Poettering [Wed, 25 Feb 2026 11:51:29 +0000 (12:51 +0100)] 
chase: drop wrong optimization

The same optimization is already applied in the very similar
dir_fd_is_root() check a few lines up – with the exception that it
doesn't accept AT_FCWD there. And frankly turning off
CHASE_AT_RESOLVE_IN_ROOT if we operate on AT_FCWD is simply wrong. Hence
remove this code.

5 weeks agochase: tighten checks on ".." once we hit the root of an CHASE_AT_RESOLVE_IN_ROOT...
Lennart Poettering [Wed, 25 Feb 2026 11:26:42 +0000 (12:26 +0100)] 
chase: tighten checks on ".." once we hit the root of an CHASE_AT_RESOLVE_IN_ROOT root tree

Let's harden things in case concurrent access is allowed to a root tree
passed via CHASE_AT_RESOLVE_IN_ROOT: let's not just validate via the
path if we hit the root of the tree, but also by comparing inodes +
mount ids.

Hardening opportunity reported by Sebastian Wick.

5 weeks agochase: port to statx()
Lennart Poettering [Wed, 25 Feb 2026 11:15:14 +0000 (12:15 +0100)] 
chase: port to statx()

In one of the next commits we want to acquire .stx_mnt_id from statx()
for each inode we traverse (plain fstat() doesn't provide that field).
Hence let's port chase() over to statx() as preparation for that.

No change in behaviour.

5 weeks agochase: put limit on overall chase cycles
Lennart Poettering [Wed, 25 Feb 2026 10:43:11 +0000 (11:43 +0100)] 
chase: put limit on overall chase cycles

Let's add some protections in case we deal with inodes owned by an
untrusted person, with concurrent access: let's put a limit on how long
we traverse, and fail eventually so that live changes cannot send us in
circles indefinitely.

This reworks the current CHASE_MAX logic so that it not only applies to
symlinks transitions, but to any transitions.

This also bumps CHASE_MAX a bit, given that it's now bumped on every
single iteration of the loop.

5 weeks agostat-util: add statx() flavours of stat_verify_regular() + stat_verify_socket()
Lennart Poettering [Wed, 25 Feb 2026 11:13:31 +0000 (12:13 +0100)] 
stat-util: add statx() flavours of stat_verify_regular() + stat_verify_socket()

5 weeks agofd-util: make use of XAT_FDROOT in path_is_root_at()
Lennart Poettering [Mon, 2 Mar 2026 17:09:23 +0000 (18:09 +0100)] 
fd-util: make use of XAT_FDROOT in path_is_root_at()

5 weeks agomeson: Work around Meson install_subdir limitation
Kai Lüke [Tue, 3 Mar 2026 01:37:49 +0000 (10:37 +0900)] 
meson: Work around Meson install_subdir limitation

When install_subdir encounters a mkosi.tools tree with a /bin to
/usr/bin symlink it fails to copy it because it dereferences but still
treats it like a file.
Work around the Meson bug by excluding the mkosi.tools tree from
installation like mkosi.local is excluded. We anyway don't want the
tools tree end up there.

5 weeks agonetwork: fix error aggregation in wwan_check_and_set_configuration()
noxiouz [Mon, 2 Mar 2026 22:42:16 +0000 (22:42 +0000)] 
network: fix error aggregation in wwan_check_and_set_configuration()

When removing marked routes, the condition `if (ret)` incorrectly
overwrites any previously accumulated error in `ret` with the latest
return value `r`, even if `r >= 0` (success). This means an earlier
real error can be silently cleared by a subsequent successful
route_remove() call.

The parallel address_remove() block just above uses the correct
`if (r < 0)` pattern. Apply the same fix to the route_remove() block.

5 weeks agodocs: document the "verity" NvPCR measurements
Lennart Poettering [Mon, 2 Mar 2026 21:57:56 +0000 (22:57 +0100)] 
docs: document the "verity" NvPCR measurements

I forgot this when I posted 32f405074a3aa221982ad92a7f61560b9f6a2b03,
let's add it now.

5 weeks agonetwork: fix LLDP field type in Interface Varlink IDL
noxiouz [Thu, 26 Feb 2026 03:31:24 +0000 (03:31 +0000)] 
network: fix LLDP field type in Interface Varlink IDL

sd_lldp_tx_describe() returns a single object (the LLDP TX configuration),
but the IDL declared LLDP as SD_VARLINK_ARRAY|SD_VARLINK_NULLABLE. This
caused server-side validation failures ("Field 'LLDP' should be an array,
but it is of type 'object'") whenever networkctl status was called on an
interface with LLDP TX active.

Also fix the field comment: the LLDP field represents the transmit
configuration, not received neighbors.

Follow-up for dd2934d44e2c9cd1a92ae0fd6806985c4bc031e6.

5 weeks agoimport,nspawn: fix a couple of typos in mountfsd
Frantisek Sumsal [Mon, 2 Mar 2026 19:58:09 +0000 (20:58 +0100)] 
import,nspawn: fix a couple of typos in mountfsd

5 weeks agonspawn: actually mask certain files under /proc/
Frantisek Sumsal [Wed, 25 Feb 2026 18:13:37 +0000 (19:13 +0100)] 
nspawn: actually mask certain files under /proc/

/run/systemd/inaccessible/ exists only on host - in the container we have
/run/host/inaccessible/, and since all the inaccessible mounts have
MOUNT_IN_USERNS we need to use the latter one, otherwise the masking
gets silently skipped:

~# SYSTEMD_LOG_LEVEL=debug systemd-nspawn -q --directory=foo ls -la /proc/kallsyms
...
Bind-mounting /run/systemd/inaccessible/reg on /proc/kallsyms (MS_BIND "")...
Failed to mount /run/systemd/inaccessible/reg (type n/a) on /proc/kallsyms (MS_BIND ""): No such file or directory
Changing mount flags /proc/kallsyms (MS_RDONLY|MS_NOSUID|MS_NODEV|MS_NOEXEC|MS_REMOUNT|MS_BIND "")...
Failed to mount n/a (type n/a) on /proc/kallsyms (MS_RDONLY|MS_NOSUID|MS_NODEV|MS_NOEXEC|MS_REMOUNT|MS_BIND ""): Invalid argument
Bind-mounting /run/systemd/inaccessible/reg on /proc/kcore (MS_BIND "")...
Failed to mount /run/systemd/inaccessible/reg (type n/a) on /proc/kcore (MS_BIND ""): No such file or directory
Changing mount flags /proc/kcore (MS_RDONLY|MS_NOSUID|MS_NODEV|MS_NOEXEC|MS_REMOUNT|MS_BIND "")...
Failed to mount n/a (type n/a) on /proc/kcore (MS_RDONLY|MS_NOSUID|MS_NODEV|MS_NOEXEC|MS_REMOUNT|MS_BIND ""): Invalid argument
...
Inner child finished, invoking payload.
-r--r--r--. 1 root root 0 Feb 25 13:19 /proc/kallsyms

5 weeks agoudev/varlink: ignore polkit related field
Yu Watanabe [Mon, 2 Mar 2026 14:41:26 +0000 (23:41 +0900)] 
udev/varlink: ignore polkit related field

Follow-up for da7374b2ae07b4d3801f5187aacc199978793680.

5 weeks agoupdate TODO
Lennart Poettering [Mon, 2 Mar 2026 16:56:53 +0000 (17:56 +0100)] 
update TODO

5 weeks agonetwork: add DHCPv6 message types to string table (#40912)
Skye Soss [Mon, 2 Mar 2026 16:50:12 +0000 (10:50 -0600)] 
network: add DHCPv6 message types to string table (#40912)

Adds the DHCPv6 message types ADDR-REG-INFORM and
ADDR-REG-REPLY to the DHCPv6 message types string table.

Follow-up for 1e55da38aab0a7e7d5ba4de3243512fa70401df9.

5 weeks agotree-wide: use _contains() rather than _get()/_find()
Yu Watanabe [Mon, 2 Mar 2026 07:06:33 +0000 (16:06 +0900)] 
tree-wide: use _contains() rather than _get()/_find()

5 weeks agojournal-remote: fix error number confusion
Lennart Poettering [Mon, 2 Mar 2026 10:34:42 +0000 (11:34 +0100)] 
journal-remote: fix error number confusion

See: https://lists.freedesktop.org/archives/systemd-devel/2026-February/051924.html

5 weeks agologin: fix two io.systemd.Login Varlink IDL issues (#40857)
Yu Watanabe [Mon, 2 Mar 2026 14:09:19 +0000 (23:09 +0900)] 
login: fix two io.systemd.Login Varlink IDL issues (#40857)

Add missing NoSessionPIDFD error to the interface definition; the error
is emitted by vl_method_create_session() when the session leader process
does not have a pidfd available, but was never declared in the IDL.

Drop SD_JSON_MANDATORY from the ReleaseSession.Id dispatch entry so that
omitting the field (which the IDL allows via SD_VARLINK_NULLABLE)
resolves to the caller's own session via session_is_self(NULL).

5 weeks agosystemd/sd-varlink-idl.h: fix ABI breakage (#40903)
Zbigniew Jędrzejewski-Szmek [Mon, 2 Mar 2026 13:08:22 +0000 (14:08 +0100)] 
systemd/sd-varlink-idl.h: fix ABI breakage (#40903)

5 weeks agologin: add missing NoSessionPIDFD error to io.systemd.Login IDL 40857/head
noxiouz [Mon, 2 Mar 2026 10:04:48 +0000 (10:04 +0000)] 
login: add missing NoSessionPIDFD error to io.systemd.Login IDL

The error is emitted by vl_method_create_session() when the session leader
process does not have a pidfd available, but was never declared in the IDL.

Fixes: 3180c4d46151673a9c985e60f205d4c76a81573f
5 weeks agologin: fix ReleaseSession.Id dispatch flag in io.systemd.Login Varlink handler
noxiouz [Mon, 2 Mar 2026 10:04:37 +0000 (10:04 +0000)] 
login: fix ReleaseSession.Id dispatch flag in io.systemd.Login Varlink handler

The ReleaseSession method's Id field is declared as nullable (?string) in
the IDL, allowing callers to omit it so that the method releases the
caller's own session via session_is_self(NULL). The SD_JSON_MANDATORY flag
in the dispatch table contradicts this and makes omitting Id return -EINVAL
("Invalid argument", parameter "Id") instead.

Drop the flag so omitting Id is treated as passing NULL.

Fixes: 2baca6c22b2d75b8ba2d0bd8a9e7f4a8579752ed
5 weeks agosystemd/sd-varlink-idl.h: fix ABI breakage 40903/head
Mike Yuan [Mon, 2 Mar 2026 09:49:17 +0000 (10:49 +0100)] 
systemd/sd-varlink-idl.h: fix ABI breakage

Follow-up for 2e51ed7fcb8a4215432ca189f8b3d2ad848ea93b

5 weeks agocore/varlink-unit: use VARLINK_ERROR_UNIT_NO_SUCH_UNIT macro
Mike Yuan [Mon, 2 Mar 2026 09:45:43 +0000 (10:45 +0100)] 
core/varlink-unit: use VARLINK_ERROR_UNIT_NO_SUCH_UNIT macro

Follow-up for 1fc868ac6b74d61c75d00a62aa4331961dead3ed

5 weeks agonspawn: allow cachestat systemcall
Thomas Weißschuh [Thu, 26 Feb 2026 17:30:32 +0000 (18:30 +0100)] 
nspawn: allow cachestat systemcall

cachestat(2) is essentially a better variant of mincore(2).
As mincore(2) is already allowed, also allow cachestat(2).

5 weeks agopidref: do not tamper with pidref_acquire_pidfd_id() retval
Mike Yuan [Thu, 19 Feb 2026 23:14:19 +0000 (00:14 +0100)] 
pidref: do not tamper with pidref_acquire_pidfd_id() retval

If a pidfd id is passed to us, it must have originated from
somewhere, i.e. pidfs should be supported on the system.
Hence let's be honest about -EOPNOTSUPP, which might indicate
some bigger problems under the hood, rather than allowing
it to be turned into mild -ESRCH.