test: wait until `lvm-activate-$vgroup.service` finishes
The new lvm autoactivation method runs `vgchange` via
`systemd-run --no-block`[0], which means that checking if the unit
is in the `active` state is not enough, since the main binary might
still be running. Let's fix this by waiting until the unit reaches
the `exited` sub state.
xdavidwu [Fri, 5 Nov 2021 14:37:06 +0000 (22:37 +0800)]
coredump: fix filename in journal when not compressed
Since 587f2a5e564cf434c2e0a653f52b8f73e86092d8, filename for
not-compressed coredump is missing from save_external_coredump, making
it write COREDUMP_FILENAME= (empty) in journal, making `coredumpctl`
report it missing but it is actually saved.
This fixes it.
Daan De Meyer [Fri, 29 Oct 2021 12:54:31 +0000 (13:54 +0100)]
getty: Pass tty to use by agetty via stdin
If the tty arg is set to "-", agetty uses the stdin fd as the tty.
Let's pass the tty this way so that we keep an fd open to the tty
at all times. If all fd's to a tty are closed, the kernel might
reset the tty which we want to avoid.
Jan Janssen [Thu, 4 Nov 2021 09:37:25 +0000 (10:37 +0100)]
sd-boot: Add .osrel section
This allows starting systemd-boot from \EFI\Linux for easier testing
and bisection without risking an unbootable system as the user does not
need to replace their working loader.
homed: add env var to override dir where we fine stored user records
This adds an env var which we can use to redirect where homed stores and
looks for user records kept on the host. This is useful for debugging
purposes so that one can easily run another homed instnce that doesn't
interfere with the main instance.
Imports are sorted in the usual fashion: stdlib first.
literal_eval() parses string/numbers/lists/sets/dicts, and nothing else, while
eval will execute any python code. Using literal_eval() is generally more
correct, because it avoids the risk of side effects from the parsed expression.
In this case, we generate the parsed strings ourselves, so it's very unlikely
to have anything unexpected in the expressions. But let's do the correct thing
anyway.
loop-util: reopen device node if we shortcut loop device creation
The LoopDevice object supports a shortcut: if the backing fd we are
supposed to create a loopback device of refers to a
block device alrady then we'll use it as is – if we can – instead of
setting up an unnecessary loopback device that would be pretty much
the same as its backing device.
Previously, when doing this we'd just dup() the original backing fd and
use that. But that's problematic in case O_DIRECT was set on the fd,
since we'll keep that flag set on our copy too, which means we can't do
simple, regular IO on it anymore.
Thus, let's reopen the inode in this case with the exact access flags
we'd apply if we'd actually allocate and open a new loopback device.
Jan Janssen [Tue, 2 Nov 2021 09:47:41 +0000 (10:47 +0100)]
sd-boot: Fix efi_arch checks
efi_arch is set to whatever meson returns from host_machine.cpu_family() and
gnu_efi_arch is then set accordingly.
On x86, efi_arch == x86 and gnu_efi_arch == ia32. Surprisingly, compilation
worked (without disabling some instruction sets) and nobody ever complained
about systemd-bootia32.efi not working.
Jan Janssen [Tue, 2 Nov 2021 09:21:19 +0000 (10:21 +0100)]
sd-boot: Detect supported compile args
This brings the final list of compiler arguments more in line with
how meson does things. In particular, --buildtype=plain will not
add any optimizations on its own (and would have to be provided
through CFLAGS=).
Jan Janssen [Mon, 1 Nov 2021 09:36:12 +0000 (10:36 +0100)]
sd-boot: Add some link flags
These flags are used by gnu-efi. Adding --no-undefined even allows
us to drop the no-undefined-symbols.sh test in favor of immediate
compile time errors.
tree-wide: do not print hint about -M if -M is already used
(Or when -H is used, since -H and -M are incompatible.)
Note that the slightly unusual form with separate boolean variables (hint_vars,
hint_addr) instead of e.g. a const char* variable to hold the message, because this
way we don't trigger the warning about non-literal format.
tree-wide: always use TAKE_FD() when calling rearrange_stdio()
rearrange_stdio() invalidates specified fds even on failure, which means
we should always invalidate the fds we pass in no matter what. Let's
make this explicit by using TAKE_FD() for that everywhere.
Note that in many places we such invalidation doesnt get us much
behaviour-wise, since we don't use the variables anymore later. But
TAKE_FD() in a way is also documentation, it encodes explicitly that the
fds are invalidated here, so I think it's a good thing to always make
this explicit here.
Daan De Meyer [Mon, 18 Oct 2021 12:17:02 +0000 (14:17 +0200)]
core: Propagate condition failed state to triggering units.
Alternative to https://github.com/systemd/systemd/pull/20531.
Whenever a service triggered by another unit fails condition checks,
stop the triggering unit to prevent systemd busy looping trying to
start the triggered unit.
I got the logic reversed in 6d9326595592f98e8126eacb4176acd8c3516d5c.
Let's just remove the conditionalization of the status message: if we're
sending something, we might just as well always attach READY=1, the extra
few bytes don't make much of a difference.
FWIW, it seems that this bug didn't cause problems, probably because we'd send
READY=1 either from user_manager_send_ready() or from a later call to
manager_send_ready().
tree-wide: warn when sd_notify fails with READY=1 or FDSTOREREMOVE=1
Most sd_notify() calls are like log_info() — the result is only informative
and if they fail, it's best ignore this. But if a call with READY=1 fails,
the unit may enter a failed state, so we should warn about this. Similarly
for FSTOREREMOVE=1: the manager may be left with a stale fd, at least wasting
resources.
CODING_STYLE: allow joined variable declarations and function calls
… but only for a single variable.
The guidelines already allowed declaring variables at the point of
initialization. But not making a function call to do that. Let's allow that
now. The existing style of declaring and initializing seperate is still
allowed, and whatever makes most sense should be used.
test-process-util: also add EROFS to the list of "good" errors
It is only added in the one place where we actually try to set the
setting to a new value. Before we were testing if we can set to it the
existing value, which was a noop. We could still get a permission error,
but this is the first place where we would propagate EROFS.
procfs-util: fix confusion wrt. quantity limit and maximum value
From packit/rawhide-arm64 logs:
Assertion 'limit >= INT_MAX || get_process_ppid(limit+1, NULL) == -ESRCH' failed at src/test/test-process-util.c:855, function test_get_process_ppid(). Aborting.
――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――
The kernel has a few different limits. In particular kernel.threads-max can be
set to some lower value, and kernel.pid_max can be set to a higher value. This
is nice because it reduces PID reuse, even if the number of threads that is
allowed is limited. But the tests assumed that we cannot have a thread with
PID above MIN(kernel.threads-max, kernel.pid_max-1), which is not valid.
So let's rework the whole thing: let's expose the helpers to read
kernel.threads-max and kernel.pid_max, and print what they return in tests.
procfs_tasks_get_limit() was something that is only used in tests, and wasn't
very well defined, so let's drop it.
Franck Bui [Wed, 19 May 2021 12:14:58 +0000 (14:14 +0200)]
test-keymap-util: always use kbd-model-map we ship
This test makes assumptions on the availability of some mappings contained in
kbd-model-map and therefore strongly relies on the version shipped by
upstream. IOW the test is likely to fail if it's installed on a system with a
more comprehensive kbd-model-map.
This patch makes the upstream kbd-model-map file available via a symlink in
test/testdata/test-keymap-util dir and makes sure that this specific version is
always used by test-keymap-util regardless of whether the test is installed and
run on a different system or directly run (optionally via meson) from the
project working dir.
Topi Miettinen [Sat, 30 Oct 2021 16:58:41 +0000 (19:58 +0300)]
execute: respect selinux_context_ignore
When `SELinuxContext=` parameter is prefixed with `-`, the documentation states
that any errors determining or changing context should be ignored, but this
doesn't actually happen and the service may fail with `229/SELINUX_CONTEXT`.
Fix by adding checks to `context->selinux_context_ignore`.
Daan De Meyer [Tue, 19 Oct 2021 09:45:48 +0000 (10:45 +0100)]
core: Delay start rate limit check when starting a unit
Doing start rate limit checks before doing condition checks made
condition check failures count towards the start rate limit which
broke existing assumptions (see #21025). Run the rate limit checks
after the condition checks again to restore the previous behaviour.
Daan De Meyer [Sat, 30 Oct 2021 10:15:22 +0000 (11:15 +0100)]
mkosi: Install less in the mkosi Fedora image
We're actually falling back to `more` in the mkosi image which doesn't
behave quite the same as less which is somewhat annoying. Let's make
sure `less` is installed so systemd can use it as the pager.
The checks for finding a new available address in the pool were broken in two
ways: not using UINT32_TO_PTR() on hashmap lookups resulted in false negatives,
and the check wasn't skipping the server address if that was part of the pool.
Move the check for available addresses to a small helper function and fix both
issues, and also add a check to the REQUEST code for the server address.
sd-dhcp-server: clear out expired leases when processing requests
The DHCP server configuration supports setting a maximum lease time, but old
leases are never actually cleared out if the client doesn't send a RELEASE.
This causes the pool to run out of addresses on networks where clients just
disappear, which is a fairly common occurrence on wireless networks.
Fix this by cleaning up expired leases before processing client requests, so
addresses can be reused for new clients.