hwdb: move key 66/65 handling from specific to generic HP laptop coverage
This takes the idea from #18595 and implements it based on our current
hwdb: the original PR suggested the keys 66/65 are a generic HP thing,
and not limited to specific laptops. The current specific laptop entries
do not contradict that claim.
Hence, let's move them from the specific sections matching some HP
laptops to the generic section matching all.
This uses the correct key names, which have long been fixed (which used
to be a problem our CI was tripped off by).
This is not tested, but I think fairly risk-less, and should allow us to
get rid of a really old PR.
Chen Guanqiao [Wed, 2 Oct 2024 05:10:21 +0000 (13:10 +0800)]
mount: optimize mountinfo traversal by decoupling device discovery
In mount_load_proc_self_mountinfo(), device_found_node() is synchronously called
during the traversal of mountinfo entries. When there are a large number of
mount points, and the device types are not significantly different, this results
in excessive time consumption during device discovery, causing a performance
bottleneck. This issue is particularly prominent on servers with a large number
of cores in IDC.
This patch decouples device discovery from the mountinfo traversal process,
avoiding redundant device operations. As a result, it significantly improves
performance, especially in environments with numerous mount points.
Yu Watanabe [Mon, 7 Oct 2024 21:19:04 +0000 (06:19 +0900)]
core: suppress one debugging log
Otherwise, the log is shown even when getting properties.
Even though it is in the debug level, that's quite noisy.
[ 338.785847] TEST-55-OOMD.sh[1624]: Oct 07 16:35:15 H systemd[1]: TEST-55-OOMD-testmunch.service: Unit not running in private mount namespace, cannot live mount
[ 338.786985] TEST-55-OOMD.sh[1624]: Oct 07 16:35:17 H systemd[1]: TEST-55-OOMD-testmunch.service: Unit not running in private mount namespace, cannot live mount
[ 338.787412] TEST-55-OOMD.sh[1624]: Oct 07 16:35:20 H systemd[1]: TEST-55-OOMD-testmunch.service: Unit not running in private mount namespace, cannot live mount
[ 338.791776] TEST-55-OOMD.sh[1624]: Oct 07 16:35:22 H systemd[1]: TEST-55-OOMD-testmunch.service: Unit not running in private mount namespace, cannot live mount
[ 338.792938] TEST-55-OOMD.sh[1624]: Oct 07 16:35:24 H systemd[1]: TEST-55-OOMD-testmunch.service: Unit not running in private mount namespace, cannot live mount
[ 338.793225] TEST-55-OOMD.sh[1624]: Oct 07 16:35:26 H systemd[1]: TEST-55-OOMD-testmunch.service: Unit not running in private mount namespace, cannot live mount
[ 338.793424] TEST-55-OOMD.sh[1624]: Oct 07 16:35:28 H systemd[1]: TEST-55-OOMD-testmunch.service: Unit not running in private mount namespace, cannot live mount
[ 338.796448] TEST-55-OOMD.sh[1624]: Oct 07 16:35:31 H systemd[1]: TEST-55-OOMD-testmunch.service: Unit not running in private mount namespace, cannot live mount
[ 338.797997] TEST-55-OOMD.sh[1624]: Oct 07 16:35:33 H systemd[1]: TEST-55-OOMD-testmunch.service: Unit not running in private mount namespace, cannot live mount
[ 338.799206] TEST-55-OOMD.sh[1624]: Oct 07 16:35:35 H systemd[1]: TEST-55-OOMD-testmunch.service: Unit not running in private mount namespace, cannot live mount
The method was added with migration of resources in mind (e.g. process's
allocated memory will follow it to the new scope), however, such a
resource migration is not in cgroup semantics. The method may thus have
the intended users and others could be guided to StartTransientUnit().
Since this API was advertised in a regular release, start the removal
with a deprecation message to callers.
Eventually, the goal is to remove the method to clean up DBus API and
simplify code (removal of cgroup_context_copy()).
Part of DBus docs is retained to satisfy build checks.
Catch up with the nice little toys the kernel fs developers have added
for us. Preferably, let's make use of the new F_DUPFD_QUERY fcntl() call
that checks whether two fds are just duplicates of each other
(duplicates as in dup(), not as in open() of the same inode, i.e.
whether they share a single file offset and so on).
This API is much nicer, since it is a core kernel feature, unlike the
kcmp() call we so far used, which is part of the (optional)
checkpoint/restore stuff.
Luca Boccassi [Mon, 7 Oct 2024 15:48:55 +0000 (16:48 +0100)]
mkosi: mark test as skipped when QEMU crashes
On Ubuntu/Debian infrastructure QEMU crashes a lot, so mark the test
as skipped in that case as there's nothing we can do about it and
we shouldn't mark runs as failed
Ryan Wilson [Mon, 30 Sep 2024 16:58:34 +0000 (09:58 -0700)]
Add ExtraFileDescriptor property to StartTransientUnit dbus API
This adds the ExtraFileDescriptor property to StartTransient dbus API
with format "a(hs)" - array of (file descriptor, name) pairs. The FD
will be passed to the unit via sd_notify like Socket and OpenFile.
systemctl show also shows ExtraFileDescriptorName for these transient
units. We only show the name passed to dbus as the FD numbers will
change once passed over the unix socket and are duplicated, so its
confusing to display the numbers.
We do not add this functionality for systemd-run or general systemd
service units as it is not useful for general systemd services.
Arguably, it could be useful for systemd-run in bash scripts but we
prefer to be cautious and not expose the API yet.
Yu Watanabe [Sun, 6 Oct 2024 05:43:45 +0000 (14:43 +0900)]
udev-node: skip stack directory creation for diskseq
The disk/by-diskseq symlink should not be shared with multiple block
devices. Hence, it is not necessary to create stack directory for the
symlink that manages which device owns the symlink.
This is not just a optimization.
If a service unit tries to mount a disk image but the service fails, then
the diskseq of the loop device for the image may be continuously increased
during restart, and inodes in /run may increase rapidly, as the stack
directories are cleaned up only when udev queue is empty.
Mike Yuan [Sat, 14 Sep 2024 17:14:33 +0000 (19:14 +0200)]
path-lookup: move xdg_user_dirs() to xdg-autostart-generator
This is the only place where xdg_user_dir() is needed and
makes sense. All other invocations have been replaced with
user_search_dirs() - see previous commits for details.
* Rename user_dirs() -> user_unit_search_dirs() and port to
user_search_dirs()
* Use STRV_IFNOTNULL to guard paths that could be NULL,
assert otherwise
xdg_user_dirs() doesn't seem well-organized currently.
In all other xdg_user_*() funcs we assume /etc/xdg/systemd
to be a symlink to /etc/systemd/, hence it is the odd one out.
Also, when the relevant envvar is unset, it only returns
the global search dirs.
sd_path_lookup() actually covers this nicely with SD_PATH_SEARCH_*,
where the combined search paths (from user home and system) are used.
Therefore, let's introduce a wrapper for that, and deprecate xdg_user_dirs()
(would be removed in later commits).
Note that -ENXIO reported by xdg_user_config_dir() is now properly
propagated rather than ignored, as unlike XDG_RUNTIME_DIR, XDG_CONFIG_HOME
has a default value hence ENXIO is not really expected.
Daniel Martinez [Sun, 6 Oct 2024 03:39:43 +0000 (23:39 -0400)]
Use case insensitive comparison for the machine's architechture
boot loader specification states:
architecture: refers to the architecture this entry is for. The argument
should be an architecture identifier, using the architecture vocabulary
defined by the EFI specification (i.e. IA32, x64, IA64, ARM, AA64, …).
If specified and it does not match the local system architecture this
entry should be hidden. The comparison should be done case-insensitively.
Daan De Meyer [Tue, 1 Oct 2024 12:30:15 +0000 (14:30 +0200)]
ukify: Rework multi-profile UKIs
The API introduced in https://github.com/systemd/systemd/pull/34295
is less than ideal:
- It doesn't consider signing at all (ukify can't sign separately yet)
- Measurement is completely broken (all profile sections are marked to
not be measured)
- It focuses on a very niche use case of extending existing UKIs and makes
the more common use case of building a UKI with several profiles included
much harder than needed.
Let's instead rework the API to focus on the primary use case of building
a UKI with multiple profiles added to it immediately. We require the profiles
to be built upfront as separate PE binaries with UKI. There's no need to sign
or measure these, they're solely vehicles for profile sections. This saves us
from having to complicate the command line and config parsing to support defining
multiple profiles.
To add the profiles when building a UKI, we introduce the new --add-profile
switch which takes a path to a PE binary describing a profile. The required
sections are read from each PE binary, measured and added as a profile.
The integration test is disabled until the new API is merged and exposed in
mkosi so that building a UKI with profiles can be left to mkosi and the integration
test will only test the switching between profiles and not the building of UKIs
with profiles.