]> git.ipfire.org Git - thirdparty/systemd.git/log
thirdparty/systemd.git
16 months agoterminal-util: return correct error in chvt()
Lennart Poettering [Wed, 10 Jul 2024 14:32:35 +0000 (16:32 +0200)] 
terminal-util: return correct error in chvt()

16 months agoterminal-util: refuse a few more unexpected open flags in open_terminal()
Lennart Poettering [Wed, 10 Jul 2024 15:01:47 +0000 (17:01 +0200)] 
terminal-util: refuse a few more unexpected open flags in open_terminal()

16 months agoterminal-util: trivial white space fix
Lennart Poettering [Wed, 10 Jul 2024 15:17:20 +0000 (17:17 +0200)] 
terminal-util: trivial white space fix

16 months agoenv-util: suppress unnecessary setenv() in setenvf()
Lennart Poettering [Wed, 17 Jul 2024 15:54:09 +0000 (17:54 +0200)] 
env-util: suppress unnecessary setenv() in setenvf()

16 months agovmspawn: make "-n" just work
Lennart Poettering [Thu, 18 Jul 2024 13:39:39 +0000 (15:39 +0200)] 
vmspawn: make "-n" just work

The tap network device should be called "vt-", so that that the
80-vm-vt.network file we ship by default actually matches against it.

Also, turn off any qemu callout stuff, networkd is smart enough to
handle all this on its own, without ugly callouts.

16 months agoMerge pull request #33591 from teknoraver/o_path
Luca Boccassi [Fri, 19 Jul 2024 09:01:01 +0000 (11:01 +0200)] 
Merge pull request #33591 from teknoraver/o_path

use O_PATH when possible

16 months agoMerge pull request #33770 from bluca/polkit-root
Luca Boccassi [Fri, 19 Jul 2024 08:57:07 +0000 (10:57 +0200)] 
Merge pull request #33770 from bluca/polkit-root

polkit: map POLKIT_ALWAYS_QUERY to new polkit flag

16 months agoman: Mention Type=oneshot timeout directive
Lucas Werkmeister [Tue, 16 Jul 2024 16:15:32 +0000 (18:15 +0200)] 
man: Mention Type=oneshot timeout directive

Make the warning for oneshot services (where RuntimeMaxSec= has no
effect) more actionable by pointing to the directive people can use
instead to effectively limit their runtime.

16 months agoMerge pull request #33773 from kovalev0/fix_aquarius_cmp_ns483_keyb_and_sensor
Luca Boccassi [Thu, 18 Jul 2024 20:28:23 +0000 (22:28 +0200)] 
Merge pull request #33773 from kovalev0/fix_aquarius_cmp_ns483_keyb_and_sensor

Fix aquarius cmp ns483 keyboard and sensor

16 months agohwdb: add backslash and touchpad toggle mapping for Aquarius Cmp NS483 33773/head
Vasiliy Kovalev [Thu, 18 Jul 2024 18:58:55 +0000 (21:58 +0300)] 
hwdb: add backslash and touchpad toggle mapping for Aquarius Cmp NS483

Signed-off-by: Vasiliy Kovalev <kovalev@altlinux.org>
16 months agohwdb: fix accelerometer mount matrix for Aquarius Cmp NS483
Vasiliy Kovalev [Thu, 18 Jul 2024 18:42:18 +0000 (21:42 +0300)] 
hwdb: fix accelerometer mount matrix for Aquarius Cmp NS483

Signed-off-by: Vasiliy Kovalev <kovalev@altlinux.org>
16 months agoMerge pull request #33767 from keszybz/test-ukify-nicer
Luca Boccassi [Thu, 18 Jul 2024 15:59:55 +0000 (17:59 +0200)] 
Merge pull request #33767 from keszybz/test-ukify-nicer

Make test_ukify not use a real initrd

16 months agopolkit: fix typo in enum name 33770/head
Luca Boccassi [Thu, 18 Jul 2024 15:43:56 +0000 (16:43 +0100)] 
polkit: fix typo in enum name

16 months agol10n: fix credits for the French translation
Léane GRASSER [Thu, 18 Jul 2024 13:15:16 +0000 (15:15 +0200)] 
l10n: fix credits for the French translation

16 months agopolkit: map POLKIT_ALWAYS_QUERY to new polkit flag
Luca Boccassi [Thu, 27 Jun 2024 19:55:34 +0000 (20:55 +0100)] 
polkit: map POLKIT_ALWAYS_QUERY to new polkit flag

polkitd by default just waves through requests from a root process.
A new POLKIT_CHECK_AUTHORIZATION_FLAGS_ALWAYS_CHECK flag was added
to main (will be part of v125 when it ships) that forces it to go
through the policy checks for root too. Previous versions will just
ignore it.

Change the flags handling slightly so that we pass this or the
interactive flags through, as the values match what polkit expects.

16 months agoMerge pull request #33763 from DaanDeMeyer/mkosi
Daan De Meyer [Thu, 18 Jul 2024 14:43:15 +0000 (16:43 +0200)] 
Merge pull request #33763 from DaanDeMeyer/mkosi

mkosi: Two fixlets

16 months agoptyfwd: reset color after two tty reset sequences, too
Lennart Poettering [Thu, 18 Jul 2024 12:28:52 +0000 (14:28 +0200)] 
ptyfwd: reset color after two tty reset sequences, too

When we patch in a bg color we must make sure that when certain "reset"
sequences are transferred we fix up the bg color again.

Do so for \033[!p ("soft terminal reset") and \033c ("reset to initial
state" aka "full reset").

16 months agotest_ukify: add instructions 33767/head
Zbigniew Jędrzejewski-Szmek [Thu, 18 Jul 2024 12:50:43 +0000 (14:50 +0200)] 
test_ukify: add instructions

Copied directly from a1d6dbb1c94685d7972f63ed2762fe4ba0251287.

16 months agotest_ukify: use sha384 in the signing tests
Zbigniew Jędrzejewski-Szmek [Thu, 18 Jul 2024 12:48:37 +0000 (14:48 +0200)] 
test_ukify: use sha384 in the signing tests

On Fedora, with crypto policy TEST-FEDORA41, sha1 is not allowed:

$ SYSTEMD_LOG_LEVEL=debug build/systemd-measure sign
  --linux=/lib/modules/6.9.7-200.fc40.x86_64/vmlinuz
  --osrel=/tmp/tmp.osrelbl2sr77f
  --cmdline=/tmp/tmp.cmdlineouc7hqtj
  --uname=/tmp/tmp.unamecbjgesty
  --pcrpkey=/tmp/tmpufiadu8l
  --initrd=/boot/3a9d668b4db749398a4a5e78a03bffa5/6.9.7-200.fc40.x86_64/initrd
  --sbat=/tmp/tmp.sbataz9arpy0
  --private-key=/tmp/tmppyf0gx6w
  --public-key=/tmp/tmpufiadu8l
  --bank=sha1
Measuring boot phases: enter-initrd, enter-initrd:leave-initrd, enter-initrd:leave-initrd:sysinit, enter-initrd:leave-initrd:sysinit:ready
Loaded 'libtss2-esys.so.0' via dlopen()
Loaded 'libtss2-rc.so.0' via dlopen()
Loaded 'libtss2-mu.so.0' via dlopen()
PolicyPCR calculated digest: cec1a2ccb188ddd171a2be7bfa6b31cb9148776647354eb1069e0f891ed2dbe7
Failed to initialize signature context: error:03000098:digital envelope routines::invalid digest
Failed to sign PCR policy: Input/output error

16 months agotest_ukify: do not use files from /boot
Zbigniew Jędrzejewski-Szmek [Thu, 18 Jul 2024 12:43:32 +0000 (14:43 +0200)] 
test_ukify: do not use files from /boot

They might not be readable to the unprivileged user running the tests
and it shouldn't really matter what is used. OTOH, we need a real kernel
because we look at the header.

16 months agomkosi: update debian commit reference
Luca Boccassi [Thu, 18 Jul 2024 11:44:41 +0000 (12:44 +0100)] 
mkosi: update debian commit reference

2d10c12bd5 Drop /etc/sysctl.d/99-sysctl.conf symlink
fb73af0d22 update changelog
9d74923dd7 Move systemd-time-wait-sync to systemd-timesyncd package
8e0914aa0d Update changelog for 256.2-1 release
6dabf74a81 Install run0 zsh completion file
*   a72e454b2b Update upstream source from tag 'upstream/256.2'
|\
| * cf928e099f New upstream version 256.2
c473d940f7 d/e/checkout-upstream: switch packaging branch on upstream stable PRs
13586fc76f d/e/checkout-upstream: do not fail if rebase fails
19785960d1 d/e/checkout-upstream: fix shellcheck warnings
72b215c3f0 Install zsh completion for run0
51f87a981e initramfs-tools: copy network drop-ins too

16 months agomkosi: Fix indentation 33763/head
Daan De Meyer [Thu, 18 Jul 2024 11:01:27 +0000 (13:01 +0200)] 
mkosi: Fix indentation

16 months agomkosi: Fix typo
Daan De Meyer [Thu, 18 Jul 2024 10:53:53 +0000 (12:53 +0200)] 
mkosi: Fix typo

Our config parsing is flexible enough that this kind of worked surprisingly
enough.

16 months agomkosi: Drop util-linux from centos/fedora packages
Daan De Meyer [Thu, 18 Jul 2024 10:51:44 +0000 (12:51 +0200)] 
mkosi: Drop util-linux from centos/fedora packages

It's already included in the packages list in mkosi.conf.

16 months agomkosi: Drop udev from Packages= list
Daan De Meyer [Thu, 18 Jul 2024 10:48:48 +0000 (12:48 +0200)] 
mkosi: Drop udev from Packages= list

It's pulled in via VolatilePackages=, no need to put it in Packages=.

16 months agomkosi: Fix formatting
Daan De Meyer [Thu, 18 Jul 2024 10:47:51 +0000 (12:47 +0200)] 
mkosi: Fix formatting

All of our lists start on the next line, so let's make KernelCommandLine=
fit that as well.

16 months agoMerge pull request #33761 from DaanDeMeyer/mkosi
Daan De Meyer [Thu, 18 Jul 2024 10:46:58 +0000 (12:46 +0200)] 
Merge pull request #33761 from DaanDeMeyer/mkosi

mkosi: Various NO_BUILD improvements

16 months agomkosi: Build CentOS Stream 10 images by default 33761/head
Daan De Meyer [Thu, 18 Jul 2024 06:46:14 +0000 (08:46 +0200)] 
mkosi: Build CentOS Stream 10 images by default

CentOS Stream 10 has a newer util-linux which means the terminal
gets correctly resized to the size specified by mkosi. This is a
much nicer experience than CentOS Stream 9 where you're stuck on
80x24 so let's make CentOS Stream 10 the default release to build.

16 months agoupdate TODO
Lennart Poettering [Thu, 18 Jul 2024 09:40:10 +0000 (11:40 +0200)] 
update TODO

16 months agomkosi: Streamline running the integration tests without building systemd
Daan De Meyer [Thu, 11 Jul 2024 14:09:47 +0000 (16:09 +0200)] 
mkosi: Streamline running the integration tests without building systemd

Let's document in detail how to build the integration test image and run
the integration tests without building systemd. To streamline the process,
we stop automatically using binaries from build/ when invoking mkosi directly
and don't automatically use a tools tree anymore if systemd on the host is too
old. Instead, we document these options in HACKING.md and change the mkosi meson
target to automatically use the current build directory as an extra binary search
path for mkosi.

16 months agomkosi: Skip sync script if NO_BUILD is enabled
Daan De Meyer [Thu, 11 Jul 2024 14:09:15 +0000 (16:09 +0200)] 
mkosi: Skip sync script if NO_BUILD is enabled

If we're not doing a build, there's no point in syncing either so
lets skip it.

16 months agoMerge pull request #33092 from YHNdnzj/freezer-cleanup
Mike Yuan [Wed, 17 Jul 2024 21:45:32 +0000 (23:45 +0200)] 
Merge pull request #33092 from YHNdnzj/freezer-cleanup

UnitFreezer: several cleanups

16 months agosd-device: remove debug log message when dirs are missing
Lennart Poettering [Wed, 17 Jul 2024 13:56:59 +0000 (15:56 +0200)] 
sd-device: remove debug log message when dirs are missing

This is a common case, and nothing noteworthy at all. For example, if we
establish an enumerator for listing all devices tagged by some tag, then
the per-tag dir is not going to exist if there are currently no devices
tagged that way, but that's a really common case, and doesn't really
deserve any mention, not even at debug level.

16 months agomain: show different welcome msg in initrd than on the host
Lennart Poettering [Wed, 17 Jul 2024 13:07:29 +0000 (15:07 +0200)] 
main: show different welcome msg in initrd than on the host

It has bugged me for a while that we show the exact same welcome message
at boot twice: once in the initrd, and once after the initrd→host
transition. That's very confusing.

Let's change the text a bit, and tone down the initrd message a bit (by
removing the empty line before and after it), because it is the less
relevant one.

16 months agovarlink-util: minor tweak
Lennart Poettering [Wed, 17 Jul 2024 16:56:31 +0000 (18:56 +0200)] 
varlink-util: minor tweak

Apparently I forgot that in the initial PR. Fix that.

https://github.com/systemd/systemd/pull/33714#pullrequestreview-2177885475

16 months agoMerge pull request #33609 from AdrianVovk/table-json
Lennart Poettering [Wed, 17 Jul 2024 21:03:54 +0000 (23:03 +0200)] 
Merge pull request #33609 from AdrianVovk/table-json

Improve table JSON name mangling

16 months agotable: Fix JSON name mangling breaking changes 33609/head
Adrian Vovk [Thu, 11 Jul 2024 23:35:10 +0000 (19:35 -0400)] 
table: Fix JSON name mangling breaking changes

In previous commits, we've changed the JSON name mangling logic. This,
of course, will cause breaking changes to occur on anything that relied
on the JSON mangling logic.

This commit fixes those breaking changes by manually forcing the JSON
name back to what it was before.

16 months agotable: Improve mangling of JSON field names
Adrian Vovk [Wed, 3 Jul 2024 21:57:42 +0000 (17:57 -0400)] 
table: Improve mangling of JSON field names

First, when displaying JSON we convert dashes into underscores. We want
to avoid using dashes in JSON field names in new code, because some
JSON parsers don't support dashes very well.

Second, we make the first character of every word lower-case. This
better matches our JSON field name style, and makes the automatic
JSON name mangling a lot more useful for vertical tables, where fields
are given a display name. For example, "Foo Bar" would be converted into
"foo_bar" instead of "Foo_Bar", which much better matches our style.
We don't make the whole string lowercase to support cases like:
"fooBar" should stay as "fooBar".

Some situations don't behave quite perfectly, such as "Foo BarBaz" gets
converted into "foo_barBaz", or all-caps headings get mangled
incorrectly. In these situations, the JSON field should be overridden
manually. In most cases, or at least more cases than before, this
heuristic does good enough.

16 months agotable: Add TABLE_SET_JSON_FIELD_NAME
Adrian Vovk [Wed, 3 Jul 2024 21:51:35 +0000 (17:51 -0400)] 
table: Add TABLE_SET_JSON_FIELD_NAME

Lets you conveniently set JSON field names in table_add_many. Especially
useful for vertical tables. For example:

table_add_many(t,
               TABLE_FIELD, "Display Name",
               TABLE_STRING, obj->display_name,
               TABLE_SET_JSON_FIELD_NAME, "displayName",
               TABLE_FIELD, "Timestamp",
               TABLE_TIMESTAMP, obj->timestamp,
               TABLE_SET_JSON_FIELD_NAME, "timestampUSec");

16 months agoMerge pull request #33752 from DaanDeMeyer/lsm
Daan De Meyer [Wed, 17 Jul 2024 18:15:17 +0000 (20:15 +0200)] 
Merge pull request #33752 from DaanDeMeyer/lsm

Two mkosi improvements

16 months agomkosi: Remove enforcing=0 from default kernel command line 33752/head
Daan De Meyer [Wed, 17 Jul 2024 16:56:02 +0000 (18:56 +0200)] 
mkosi: Remove enforcing=0 from default kernel command line

We already have selinux=0 in the default kernel command line so
enforcing=0 is redundant. Instead, pass in enforcing=0 when we
enable selinux in TEST-06-SELINUX.

16 months agomkosi: Stop setting apparmor=0
Daan De Meyer [Wed, 17 Jul 2024 16:55:24 +0000 (18:55 +0200)] 
mkosi: Stop setting apparmor=0

It doesn't get pulled in as a dependency anyway and kernel command
line space is precious so let's remove apparmor=0 as it's a noop
anyway.

16 months agosleep,home: always initialize UnitFreezer if used 33092/head
Mike Yuan [Tue, 11 Jun 2024 14:00:22 +0000 (16:00 +0200)] 
sleep,home: always initialize UnitFreezer if used

Previously, unit_freezer_new_freeze() would only return
UnitFreezer object if FreezeUnit() succeeds. This is not
ideal though, as a failed bus call doesn't mean the action
actually failed. E.g. a timeout might occur because pid1
is waiting for cgroup event from kernel, while the bus call
timeout was exceeded (#33269). In such a case, ThawUnit()
will never be called, resulting in frozen units remain that
way after resuming from sleep.

Therefore, let's get rid of unit_freezer_new_freeze(),
and make sure as long as unit freezer is involved, we'll
call ThawUnit() when we're done. This should make things
a lot more robust.

16 months agosleep: also log about errno when getenv_bool fails
Mike Yuan [Tue, 11 Jun 2024 16:17:01 +0000 (18:17 +0200)] 
sleep: also log about errno when getenv_bool fails

16 months agosleep: explicitly list valid sleep operations in switch
Mike Yuan [Fri, 31 May 2024 04:18:44 +0000 (12:18 +0800)] 
sleep: explicitly list valid sleep operations in switch

To follow our usual coding style.

16 months agocore/dbus-unit: add an explicit bus error when unit is frozen by parent
Mike Yuan [Tue, 11 Jun 2024 17:09:05 +0000 (19:09 +0200)] 
core/dbus-unit: add an explicit bus error when unit is frozen by parent

While at it, use more accurate errno (EDEADLK) instead of ECHILD.

16 months agocore/unit: introduce unit_freezer_complete, correctly report end state
Mike Yuan [Fri, 31 May 2024 12:43:53 +0000 (20:43 +0800)] 
core/unit: introduce unit_freezer_complete, correctly report end state

16 months agocore/unit: introduce unit_set_freezer_state, make logging consistent
Mike Yuan [Fri, 31 May 2024 12:46:16 +0000 (20:46 +0800)] 
core/unit: introduce unit_set_freezer_state, make logging consistent

Also, emit PropertiesChanged signal for FreezerState too.

Fixes #31115

16 months agocore/cgroup: skip freezer action wholly if current == objective
Mike Yuan [Sat, 25 May 2024 10:46:55 +0000 (18:46 +0800)] 
core/cgroup: skip freezer action wholly if current == objective

16 months agocore/cgroup: replace hardcoded state set with freezer_state_finish()
Mike Yuan [Mon, 15 Jul 2024 18:32:47 +0000 (20:32 +0200)] 
core/cgroup: replace hardcoded state set with freezer_state_finish()

This makes code simpler and more readable.

16 months agocore/unit: rename a few more vars for unit_next_freezer_state()
Mike Yuan [Wed, 17 Jul 2024 15:34:16 +0000 (17:34 +0200)] 
core/unit: rename a few more vars for unit_next_freezer_state()

16 months agoid128: add 'var-partition-uuid' verb
Luca Boccassi [Tue, 16 Jul 2024 15:53:08 +0000 (16:53 +0100)] 
id128: add 'var-partition-uuid' verb

As per DPS the UUID for /var/ should be keyed by the local machine-id,
which is non-trivial to do in a script. Enhance 'systemd-id128' to
take 'var-partition-uuid' as a verb, and if so perform the
calculation.

16 months agocore/unit: use switch for unit_next_freezer_state
Mike Yuan [Sat, 25 May 2024 10:46:23 +0000 (18:46 +0800)] 
core/unit: use switch for unit_next_freezer_state

16 months agocore/unit: rename freezer "target" to "objective"
Mike Yuan [Sat, 25 May 2024 10:44:51 +0000 (18:44 +0800)] 
core/unit: rename freezer "target" to "objective"

16 months agocore/slice: simplify slice_freezer_action a bit
Mike Yuan [Sat, 25 May 2024 10:45:31 +0000 (18:45 +0800)] 
core/slice: simplify slice_freezer_action a bit

16 months agocore: make unit_can_freeze take const Unit*
Mike Yuan [Fri, 24 May 2024 14:57:18 +0000 (22:57 +0800)] 
core: make unit_can_freeze take const Unit*

16 months agocore,unit-def: use our usual way of asserting enums
Mike Yuan [Fri, 24 May 2024 14:52:28 +0000 (22:52 +0800)] 
core,unit-def: use our usual way of asserting enums

16 months agocore/unit: drop pointless unit_freezer_state wrapper
Mike Yuan [Fri, 24 May 2024 14:27:52 +0000 (22:27 +0800)] 
core/unit: drop pointless unit_freezer_state wrapper

16 months agocore/dbus-unit: use UNIT_IS_LOAD_ERROR where appropriate
Mike Yuan [Thu, 30 May 2024 12:54:24 +0000 (20:54 +0800)] 
core/dbus-unit: use UNIT_IS_LOAD_ERROR where appropriate

16 months agoMerge pull request #33714 from poettering/make-varlink-public
Daan De Meyer [Wed, 17 Jul 2024 10:58:50 +0000 (12:58 +0200)] 
Merge pull request #33714 from poettering/make-varlink-public

varlink: make API public as "sd-varlink.h"

16 months agodocs: update mkosi version mentioned in HACKING.md (#33723)
migleeson [Wed, 17 Jul 2024 07:45:26 +0000 (03:45 -0400)] 
docs: update mkosi version mentioned in HACKING.md (#33723)

* fix: update docs since the default config uses a setting only available in v23

* fix: update docs to only refer to installing from the mkosi repo

16 months agorules: Add uaccess tag to /dev/udmabuf
Hans de Goede [Tue, 16 Jul 2024 15:46:09 +0000 (17:46 +0200)] 
rules: Add uaccess tag to /dev/udmabuf

In some cases userspace may need to create dmabuffers from userspace
on such example is the software ISP part of libcamera which needs to
allocate dma-buffers for the output of the software ISP.

At first the plan was to allow console users access to /dev/dma_heap/*,
this was discussed with various kernel folks here:
https://lore.kernel.org/all/bb372250-e8b8-4458-bc99-dd8365b06991@redhat.com/

Giving console users access to the dma_heap's was deemed a bad idea
because memory allocated this way is not accounted in cgroup limits.

Giving access to /dev/udmabuf OTOH was deemed acceptable so that
is what this patch adds.

Resolves: #32662

16 months agoid128: refuse --app-specific= if we're listing GPT types
Mike Yuan [Tue, 16 Jul 2024 19:10:46 +0000 (21:10 +0200)] 
id128: refuse --app-specific= if we're listing GPT types

Prompted by #33737

The intention of b37e8184a5a376749fbf68674ed6d7a4fc9901aa
is to expose sd_id128_get_app_specific() on command line.
But combining that with GPT type list makes little sense.

16 months agogpt: add more architecture aliases
Luca Boccassi [Tue, 16 Jul 2024 16:46:54 +0000 (17:46 +0100)] 
gpt: add more architecture aliases

Same as the other aliases. Allows chaining commands like:

$ systemd-id128 show -P root-$(dpkg-architecture --query DEB_HOST_ARCH)
4f68bce3e8cd4db196e7fbcaf984b709

16 months agohomed: open blob dir with O_PATH 33591/head
Matteo Croce [Tue, 9 Jul 2024 00:05:29 +0000 (02:05 +0200)] 
homed: open blob dir with O_PATH

`sys_base_dfd` is only used as dirfd argument to *_at functions,
open it with O_PATH.

16 months agohomed: use O_PATH for FDs
Matteo Croce [Mon, 8 Jul 2024 23:50:53 +0000 (01:50 +0200)] 
homed: use O_PATH for FDs

`Home->pin_fd` is used only for pinning, open it with O_PATH.

16 months agouse O_PATH to get the root fds
Matteo Croce [Mon, 8 Jul 2024 23:24:50 +0000 (01:24 +0200)] 
use O_PATH to get the root fds

`switch_root()` opens two file descriptors to do the switch but never
reads from them. Open them with O_PATH.

16 months agouse O_PATH to obtain cgroupfs fd
Matteo Croce [Thu, 4 Jul 2024 01:16:29 +0000 (03:16 +0200)] 
use O_PATH to obtain cgroupfs fd

We get a fd to cgroupfs just to avoid it being umounted, O_PATH
is enough, use it instead of O_RDONLY.

16 months agouse O_PATH when getting cgroup fd
Matteo Croce [Thu, 4 Jul 2024 01:14:22 +0000 (03:14 +0200)] 
use O_PATH when getting cgroup fd

O_PATH mode is enough to attach a bpf program to a cgroup, use it
instead of O_RDONLY

16 months agomkosi: List library packages explicitly in VolatilePackages=
Daan De Meyer [Tue, 16 Jul 2024 12:15:38 +0000 (14:15 +0200)] 
mkosi: List library packages explicitly in VolatilePackages=

These should already get pulled in as dependencies, but it doesn't
hurt to be explicit.

16 months agomkosi: Build initrd as a subimage
Daan De Meyer [Tue, 16 Jul 2024 10:07:38 +0000 (12:07 +0200)] 
mkosi: Build initrd as a subimage

Let's make things a little more consistent and build the initrd
explicitly as a subimage as well instead of relying on mkosi building
it as part of the main image build.

We drop the opensuse initrd postinst script as we don't use erofs by
default anymore. We can always reintroduce it again later if needed.

16 months agomkosi: Drop CacheOnly=always from two subimages
Daan De Meyer [Tue, 16 Jul 2024 09:45:43 +0000 (11:45 +0200)] 
mkosi: Drop CacheOnly=always from two subimages

This is a universal setting and can only be configured in the main
image so let's drop it from the subimages.

16 months agoupdate TODO 33714/head
Lennart Poettering [Fri, 12 Jul 2024 14:38:20 +0000 (16:38 +0200)] 
update TODO

16 months agosd-json/sd-varlink: downgrade assert() → assert_ret()
Lennart Poettering [Fri, 12 Jul 2024 14:39:01 +0000 (16:39 +0200)] 
sd-json/sd-varlink: downgrade assert() → assert_ret()

Our usual rule is that we are more lenient towards misuse for public
users of our code than for ourselves. Or in other words: when validating
parameters of our public functions (those starting with sd_…) we prefer
assert_ret() over assert().

16 months agolibsystemd: use C23 explicit enum types where available
Lennart Poettering [Fri, 12 Jul 2024 14:27:25 +0000 (16:27 +0200)] 
libsystemd: use C23 explicit enum types where available

In C23 we can explicitly choose the integer type for an enum. Let's do
so to make our requirements for 64bit integers explicitly. Previously,
we'd rely on a GNU extension that would size the enum to 64bit if at
least one value outside the 32bit range is in the enum. Let's keep that
too, for compat with older compilers.

(Also, add the support for older compilers to the definition of
sd_json_dispatch_flags_t, where it was forgotten so far)

16 months agosd-varlink: make our internal Varlink API public as sd-varlink.[ch]
Lennart Poettering [Fri, 12 Jul 2024 09:22:58 +0000 (11:22 +0200)] 
sd-varlink: make our internal Varlink API public as sd-varlink.[ch]

It's time. sd-json was already done earlier in this cycle, let's now
make sd-varlink public too.

This is mostly just a search/replace job of epical proportions.

I left some functions internal (mostly IDL handling), and I turned some
static inline calls into regular calls.

16 months agoMerge pull request #33695 from DaanDeMeyer/epel
Daan De Meyer [Tue, 16 Jul 2024 09:43:49 +0000 (11:43 +0200)] 
Merge pull request #33695 from DaanDeMeyer/epel

mkosi: Add CI for CentOS Stream 10

16 months agoTEST-13-NSPAWN: make sure we don't load libnss_systemd 33695/head
Daan De Meyer [Mon, 15 Jul 2024 21:07:52 +0000 (23:07 +0200)] 
TEST-13-NSPAWN: make sure we don't load libnss_systemd

Let's make sure we don't load libnss_systemd.so from bash as the
necessary environment variables aren't set to make that work when
we're running with sanitizers enabled.

We can't add a sanitizer wrapper for bash as the wrapper runs using
bash so you end up in a loop.

16 months agomkosi: Disable unique debug source names
Daan De Meyer [Mon, 15 Jul 2024 15:15:43 +0000 (17:15 +0200)] 
mkosi: Disable unique debug source names

We use -fdebug-prefix-map= because debugedit doesn't work for us (for
a currently unknown reason since it's the most obtuse code I've ever
had the pleasure of reading). With all the unique macros enabled, the
destination directory we pass to -fdebug-prefix-map= includes the package
release. The release is either the timestamp of the current commit or
the current time if the working tree is dirty. This means it generally
changes every time we rerun the build script. However, meson only reads
compiler arguments the first time it is invoked or if --wipe is specified.
This means that on a rerun -fdebug-prefix-map= will be configured wrong
and the build will fail.

Let's prevent this from happening by disabling the unique debug source
names by overriding the --unique-debug-src-base option that is passed to
find-debuginfo.sh by rpm via the _find_debuginfo_opts macro.

16 months agosystemd-networkd-tests: Skip tests requiring dhcpd if it is not available
Daan De Meyer [Mon, 15 Jul 2024 10:26:04 +0000 (12:26 +0200)] 
systemd-networkd-tests: Skip tests requiring dhcpd if it is not available

dhcpd is not available on CentOS Stream 10

See https://github.com/systemd/systemd/issues/33717

16 months agoTEST-55-OOMD: Remove the opensuse user@ dropin
Daan De Meyer [Mon, 15 Jul 2024 07:21:51 +0000 (09:21 +0200)] 
TEST-55-OOMD: Remove the opensuse user@ dropin

Required to make TEST-55-OOMD pass on OpenSUSE.

16 months agomkosi: Install binutils
Daan De Meyer [Sun, 14 Jul 2024 20:56:43 +0000 (22:56 +0200)] 
mkosi: Install binutils

16 months agoTEST-64-UDEV-STORAGE: Use max_ioqpairs instead of num_queues
Daan De Meyer [Sun, 14 Jul 2024 20:27:09 +0000 (22:27 +0200)] 
TEST-64-UDEV-STORAGE: Use max_ioqpairs instead of num_queues

Fixes a deprecation warning from qemu.

16 months agomkosi: Use clang --print-runtime-dir
Daan De Meyer [Fri, 12 Jul 2024 14:45:04 +0000 (16:45 +0200)] 
mkosi: Use clang --print-runtime-dir

16 months agomkosi: Extend arch build script comment about symlinks
Daan De Meyer [Fri, 12 Jul 2024 14:42:00 +0000 (16:42 +0200)] 
mkosi: Extend arch build script comment about symlinks

16 months agotest: do not attempt to set xattr on tmpfs
Daan De Meyer [Fri, 12 Jul 2024 14:04:22 +0000 (16:04 +0200)] 
test: do not attempt to set xattr on tmpfs

This is only possible since a recent kernel version, and fails otherwise,
like on CentOS 9

16 months agomkosi: Add CI for CentOS Stream 10
Daan De Meyer [Thu, 11 Jul 2024 13:49:17 +0000 (15:49 +0200)] 
mkosi: Add CI for CentOS Stream 10

We switch to the c10s-sig-hyperscale branch of the spec repository
as it will receive all the latest changes the earliest before they
end up in the c9s-sig-hyperscale branch.

16 months agomkosi: Use squashfs for sysext if mkfs.erofs is not available
Daan De Meyer [Thu, 11 Jul 2024 14:19:20 +0000 (16:19 +0200)] 
mkosi: Use squashfs for sysext if mkfs.erofs is not available

CentOS Stream 10 does not have erofs-utils so let's add a fallback
to squashfs when building the sysext.

16 months agomkosi: Make epel repositories optional for CentOS Stream 9
Daan De Meyer [Fri, 5 Jul 2024 09:56:44 +0000 (11:56 +0200)] 
mkosi: Make epel repositories optional for CentOS Stream 9

This allows us to add CI for CentOS Stream 10 as EPEL 10 doesn't
exist yet and won't exist for quite some time.

CentOS Stream 10 will be enabled later as soon as
https://issues.redhat.com/browse/RHEL-46604 is resolved.

16 months agomkosi: Introduce build image
Daan De Meyer [Fri, 12 Jul 2024 13:33:49 +0000 (15:33 +0200)] 
mkosi: Introduce build image

We want the exitrd image to be built with the latest systemd as well.
As the exitrd image is built as part of mkosi.images, and all subimages
are built before the main image, this implies the packages must be built
as a subimage in mkosi.images/ as well. So we introduce the build image and
move all logic related to building distribution packages there.

This also has the nice side effect of slimming down the main image as the
build dependencies are not installed into the main image anymore. It also
makes sure the packages are built in a "clean" chroot without any of the
other packages which we install in the main image available.

16 months agomkosi: update fedora commit reference
Daan De Meyer [Sun, 14 Jul 2024 15:50:23 +0000 (17:50 +0200)] 
mkosi: update fedora commit reference

a3524fc837 Use a more precise Recommends for libkxbcommon
980ede8c0f Drop machined revert
d569018a92 Rebuilt for the bin-sbin merge
8881fa94ee Version 256.2
1cc4f83002 Link systemd-executor statically
0319e62d9c Update dracut workaround
c96f54de22 Fix ELN build
3f68c5d802 Only exclude dracut conflicts on non-fedora on upstream builds
7db154308b Conditionalize dracut Conflicts more

16 months agomkosi: Update to latest
Daan De Meyer [Fri, 12 Jul 2024 14:09:21 +0000 (16:09 +0200)] 
mkosi: Update to latest

16 months agoRevert "hwdb: Added StarLabs StarLite position sensor mapping"
Sean Rhodes [Sun, 14 Jul 2024 19:28:00 +0000 (20:28 +0100)] 
Revert "hwdb: Added StarLabs StarLite position sensor mapping"

This reverts commit 1bd5db86f50feb2efd8e54e8156816d9727f6f2f.

The `kxcjk-1013` driver in Linux will parse the rotation matrix
from ACPI. This quirk is not specific enough to exist without
causing issues on different variations.

Signed-off-by: Sean Rhodes <sean@starlabs.systems>
16 months agoMerge pull request #33718 from YHNdnzj/machine-id-followup
Lennart Poettering [Mon, 15 Jul 2024 09:16:59 +0000 (11:16 +0200)] 
Merge pull request #33718 from YHNdnzj/machine-id-followup

core/main: several cleanups

16 months agosysupdate: fix return types according to style guide
Abderrahim Kitouni [Sun, 14 Jul 2024 08:29:34 +0000 (09:29 +0100)] 
sysupdate: fix return types according to style guide

16 months agofstab-generator: add x-systemd.wants
chayleaf [Sun, 14 Jul 2024 02:15:39 +0000 (09:15 +0700)] 
fstab-generator: add x-systemd.wants

This is an analog of x-systemd.requires that adds a Wants dependency
instead. This is useful for filesystems that support mounting in
degraded states (such as multi-device filesystems).

16 months agokernel-install: Remove existing loader entries and UKIs
Jörg Behrmann [Fri, 12 Jul 2024 08:43:54 +0000 (10:43 +0200)] 
kernel-install: Remove existing loader entries and UKIs

When boot counting is enabled, adding a new loader entry or UKI can conflict
with an existing one that has booted successfully and therefore has its boot
counter removed. systemd-bless-boot will fail to bless the new successful boot,
since a file without a boot counter already exists. Since kernel-install will
clobber existing files without boot counting, we should therefore remove files
without a boot count as well, when we add a file with one.

Fixes: #33504
16 months agosysupdate: fix size_t specifier
Abderrahim Kitouni [Sat, 13 Jul 2024 08:10:10 +0000 (09:10 +0100)] 
sysupdate: fix size_t specifier

16 months agocore/main: refuse to run if /usr/ is not populated 33718/head
Mike Yuan [Sat, 13 Jul 2024 17:10:58 +0000 (19:10 +0200)] 
core/main: refuse to run if /usr/ is not populated

The support for split-usr has been dropped in v255.
So, rather than emitting a warning, let's refuse
such unsupported setups completely.

16 months agocore/main: move capability_ambient_set_apply() to initialize_runtime()
Mike Yuan [Sat, 13 Jul 2024 20:36:22 +0000 (22:36 +0200)] 
core/main: move capability_ambient_set_apply() to initialize_runtime()

This is preparation for (user) manager, and we apply capability
settings for the system manager in initialize_runtime() too.