Frantisek Sumsal [Fri, 11 Mar 2022 17:15:03 +0000 (18:15 +0100)]
test: use flock when calling mkfs.btrfs
As stated in https://github.com/systemd/systemd/issues/21819#issuecomment-1064377645
`mkfs.btrfs` doesn't hold the lock for the whole duration of
`mkfs.btrfs`, thus causing unexpected races & test fails. Let's
wrap the `mkfs.btrfs` calls in an flock wrapper to mitigate this.
Yu Watanabe [Tue, 15 Mar 2022 04:50:06 +0000 (13:50 +0900)]
udev: requeue event when the corresponding block device is locked by another process
Previously, if a block device is locked by another process, then the
corresponding worker skip to process the corresponding event, and does
not broadcast the uevent to libudev listners. This causes several issues:
- During a period of a device being locked by a process, if a user trigger
an event with `udevadm trigger --settle`, then it never returned.
- When there is a delay between close and unlock in a process, then the
synthesized events triggered by inotify may not be processed. This can
happens easily by wrapping mkfs with flock. This causes severe issues
e.g. new devlinks are not created, or old devlinks are not removed.
This commit makes events are requeued with a tiny delay when the corresponding
block devices are locked by other processes. With this way, the triggered
uevent may be delayed but is always processed by udevd. Hence, the above
issues can be solved. Also, it is not necessary to watch a block device
unconditionally when it is already locked. Hence, the logic is dropped.
Yu Watanabe [Sat, 12 Mar 2022 19:45:08 +0000 (04:45 +0900)]
udev: try to reload selinux label database less frequently
Previously, `event_run()` was called repeatedly in one `event_queue_start()`
invocation. Hence, the SELinux label database is reloaded many times needlessly.
Other settings, e.g. udev rules or hwdata, are tried to be reloaded in the
beginning of `event_queue_start()`. Let's also do so for the SELinux database.
Yu Watanabe [Sat, 12 Mar 2022 11:57:15 +0000 (20:57 +0900)]
udev: assume there is no blocker when failed to check event dependencies
Previously, if udevd failed to resolve event dependency, the event is
ignored and libudev listeners did not receive the event. This is
inconsistent with the case when a worker failed to process a event,
in that case, the original uevent sent by the kernel is broadcasted to
listeners.
Yu Watanabe [Sat, 12 Mar 2022 11:40:58 +0000 (20:40 +0900)]
udev: only ignore ENOENT or friends which suggest the block device is not exist
The ENOENT, ENXIO, and ENODEV error can happen easily when a block
device appears and soon removed. So, it is reasonable to ignore the
error. But other errors should not occur here, and hence let's handle
them as critical.
Yu Watanabe [Sun, 13 Mar 2022 12:22:57 +0000 (21:22 +0900)]
udev: remove /run/udev/queue in on_post()
When the last queued event is processed, information about subsequent
events may be already queued in the netlink socket of sd-device-monitor.
In that case, previously we once removed /run/udev/queue and touch the
file soon later, and `udevadm settle` mistakenly considered all events
are processed.
To mitigate such situation, this makes /run/udev/queue removed in on_post().
efi-loader: split efi-api.[ch] from efi-loader.[ch]
Some refactoring: split efi-loader.[ch] in two: isolate the calls that
implement out boot loader interface spec, and those which implement
access to upstream UEFI firmware features.
They are quite different in nature and behaviour, and even semantically
it makes to keep these two separate. At the very least because the
previous name "efi-loader.[ch]" suggests all was about loader-specific
APIs, but much of it is generic uefi stuff...
While we are at it, I renamed a bunch of return parameters to follow our
usual ret_xyz naming. But besides renaming no real code changes.
On systems lacking EFI or the SecureBoot efi var the caching of this
info didn#t work, since we'd see ENOENT when reading the var, and cache
that, which we then use as reason to retry next time.
Let's fix that and convert ENOENT to "secure boot", because that's what
it really means. All other errors are left as is (and reason to retry).
But let's add some debug logging for that case.
TEST-68: instead of calling daemon-reload, just use different cleanup units
On a very slow machine, things are executed out-of-order, and something
pins the previously-exited unit. Instead of fighting with this with daemon-reload,
let's just use a different cleanup unit.
Yu Watanabe [Tue, 22 Mar 2022 13:01:08 +0000 (22:01 +0900)]
network: do not enable IPv4 ACD for IPv4 link-local address if ACD is disabled explicitly
The commit 1cf4ed142d6c1e2b9dc6a0bc74b6a83ae30b0f8e makes the IPv4 ACD
enabled unconditionally for IPv4 link-local addresses even if users
explicitly disable ACD.
This makes the IPv4 ACD is enabled by default, but honor user setting.
Frantisek Sumsal [Fri, 11 Mar 2022 08:19:29 +0000 (17:19 +0900)]
lgtm: disable cpp/missing-return (again)
It looks like the fix for https://github.com/github/codeql/issues/8409
is not yet in production (and the respective query needs to be enabled
in both the main and the PR branch to get results for it, hence why it
passed in #22837).
No functional change intended. The type of the iterator is generally changed to
be 'const char*' instead of 'char*'. Despite the type commonly used, modifying
the string was not allowed.
I adjusted the naming of some short variables for clarity and reduced the scope
of some variable declarations in code that was being touched anyway.
meson: replace sh+find with an internal glob in the python helper
As suggested in https://github.com/systemd/systemd/pull/22810#discussion_r831708052
This makes the whole thing simpler. A glob is passed to helper which then resolves
it on its own. This way it's trivial to call the helper with a different
set of files for testing.
tools/dbus_exporter: deblackify and shorten code a bit
When we do mkdir, we should just use 0o777 and let the umask take care of the
rest. Specifying an explicit mode is inappropriate. And when touching the code,
let's replace black madness with normal python style.
Nishal Kulkarni [Sun, 13 Mar 2022 19:05:18 +0000 (00:35 +0530)]
core/oomd: Use oom-kill ServiceResult for oomd
To notify user of kill events from systemd-oomd we now use
`SERVICE_FAILURE_OOM_KILL` as the failure result.
`unit_check_oomd_kill` now calls `notify_cgroup_oom` to
update the service result to `oom-kill`.
We add a new xattr `user.oomd_ooms` to keep track of the OOM kills
initiated by systemd-oomd, this helps us resolve a race between sending
SIGKILL to processes and checking for OOM kill status from the xattr.
random-util: use correct minimum pool size constant
The actual minimum size of the pool across supported kernel versions is
32 bytes. So adjust this minimum.
I've audited every single usage of random_pool_size(), and cannot see
anywhere that this would have any impact at all on anything. We could
actually just not change the constant and everything would be fine, or
we could change it here and that's fine too. From both a functionality
and crypto perspective, it doesn't really seem to make a substantive
difference any which way, so long as the value is ≥32. However, it's
better to be correct and have the function do what it says, so clamp it
to the right minimum.
systemd-udev-trigger.service by default triggeres all devices regardless
of whether they were already recognized by systemd-udevd.
There are machines (especially in embedded environments) where
systemd-udev-trigger.service is configured to run at a later stage of
the boot sequence, which can lead to quite a lot of devices being
triggered although they were already recognized by systemd-udevd.
Re-triggering a lot of devices is a relatively expensive operation and
therefore should be avoided if unnecessary.
Therefore this patch introduces --initialized-nomatch, which filters out
devices that are already present in the udev database. For consistance
reasons --initialized-match is implemented as well, which filters out devices
that are *not* already present in the udev database.
Yu Watanabe [Sat, 19 Mar 2022 00:35:32 +0000 (09:35 +0900)]
sd-device-enumerator: drop /sys/subsystem support
This addresses the comment by Lennart
https://github.com/systemd/systemd/pull/22662#discussion_r829799863:
> /sys/subsystem is preparation for a future that never came.
> And given that the main proponent of this left Linux kernel
> development (Kay), I doubt this will ever come. So maybe we
> should start dropping references to /sys/subsystem/ given it's
> unlikely to materialize anytime soon.
docs: add /loader/entries.srel to the boot loader spec
This new file is supposed to address conflicts with Fedora/Grub's
frankenbootloaderspec implementation, that squatted the /loader/entries/
dir, but place incompatible files in them (that do variable expansion?).
A simple text file /loader/entries.srel shall indicate which spec is
implemented. If it contains the string "type1\n" then the
/loader/entries/ directory implements our standard spec, otherwise
something else.