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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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().
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)
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.
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.
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.
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.
* 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
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>
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).
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.
Mike Yuan [Sat, 6 Jul 2024 17:05:57 +0000 (19:05 +0200)]
logind-user: take gc_mode into account when reporting user state
Prompted by #33650
Previously, if a user manually starts user@.service (which is
something we support), we'd track it as 'manager' session.
However, since user_get_state() ignores all non-pinning sessions,
if lingering is not enabled, the user state would always be
reported as 'closing', which is spurious.
Let's instead take gc_mode into consideration, and ignore
non-pinning sessions only if USER_GC_BY_PIN.
If a drop-in is set from upper level, e.g. global unit_type.d/,
even if a unit is masked, its dropin_paths would still be partially
populated. However, unit_need_daemon_reload() would always
compare u->dropin_paths with empty strv in case of masked units,
resulting in it always returning true. Instead, let's ignore
dropins entirely here.
Adrian Vovk [Wed, 19 Jul 2023 18:43:58 +0000 (14:43 -0400)]
sysupdate: Support changelogs & appstream metadata
Makes it possible to specify URLs to a changelog and an appstream
catalog XML in the sysupdate.d/*.conf files. This will be passed along
to the clients of systemd-sysupdated, which can then present this data.
Adrian Vovk [Tue, 11 Jul 2023 22:35:54 +0000 (18:35 -0400)]
sysupdate: Add --offline mode
This prevents sysupdate from going out to the network to enumerate
available instances. When combined with the list command, this lets us
query installed instances