shared/unit-file: make sure the old hashmaps and sets are freed upon replacement
Possibly fixes #15220. (There might be another leak. I'm still investigating.)
The leak would occur when the path cache was rebuilt. So in normal circumstances
it wouldn't be too bad, since usually the path cache is not rebuilt too often. But
the case in #15220, where new unit files are created in a loop and started, the leak
occurs once for each unit file:
$ for i in {1..300}; do cp ~/.config/systemd/user/test0001.service ~/.config/systemd/user/test$(printf %04d $i).service; systemctl --user start test$(printf %04d $i).service;done
run: when waiting for unit, also check if no job is pending anymore
This is a fix-up for a7c71d214c37797d82de2f66cfe0a0a79c3a5c92: since we
now don't wait for the job to finish anymore right after enqueuing it,
we should not exit our ptyfwd logic before the unit is back to inactive
*and* no job pending anymore.
Chris Down [Tue, 26 May 2020 13:35:18 +0000 (14:35 +0100)]
service: Display updated WatchdogUSec from sd_notify
Suppose a service has WatchdogSec set to 2 seconds in its unit file. I
then start the service and WatchdogUSec is set correctly:
% systemctl --user show psi-notify -p WatchdogUSec
WatchdogUSec=2s
Now I call `sd_notify(0, "WATCHDOG_USEC=10000000")`. The new timer seems
to have taken effect, since I only send `WATCHDOG=1` every 4 seconds,
and systemd isn't triggering the watchdog handler. However, `systemctl
show` still shows WatchdogUSec as 2s:
% systemctl --user show psi-notify -p WatchdogUSec
WatchdogUSec=2s
This seems surprising, since this "original" watchdog timer isn't the
one taking effect any more. This patch makes it so that we instead
display the new watchdog timer after sd_notify(WATCHDOG_USEC):
% systemctl --user show psi-notify -p WatchdogUSec
WatchdogUSec=10s
Benjamin Berg [Wed, 25 Mar 2020 15:59:40 +0000 (16:59 +0100)]
xdg-autostart-generator: Add a generator for XDG autostart files
This generator can be used by desktop environments to launch autostart
applications and services. The feature is an opt-in, triggered by
xdg-desktop-autostart.target being activated.
Also included is the new binary xdg-autostart-condition. This binary is
used as an ExecCondition to test the OnlyShowIn and NotShowIn XDG
desktop file keys. These need to be evaluated against the
XDG_CURRENT_DESKTOP environment variable which may not be known at
generation time.
Co-authored-by: Henri Chain <henri.chain@enioka.com>
Frantisek Sumsal [Mon, 25 May 2020 20:48:01 +0000 (22:48 +0200)]
test: create public images by default
Let's create new images public by default and then symlink/copy them
into the respective private directories afterwards, not the other way
around. This should fix a nasty race condition in parallel runs where
one tests attempts to copy the backing public image at the same moment
another test is already modifying it.
units: remove KillMode=process in our regular gettys
In our regular gettys the actual shell commands live the the session
scope anyway (as long as logind is used). Hence, let's avoid
KillMode=process, it serves no purpose and is simply unsafe since it
disables systemd's own process lifecycle management.
units: don't set PrivateNetwork= in systemd-homed.service
We want to watch USB sticks being plugged in, and that requires
AF_NETLINK to work correctly and get the host's events. But if we live
in a network namespace AF_NETLINK is disconnected too and we'll not get
the host udev events.
tmpfiles: deal with kernel errno change if quota is not enabled
Old kernels returned EINVAL if quota was off but we tried to manipulate
it anyway. Since
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=8a36e408d40606e21cd4e2dd9601004a67b14868
this changed: now ENOTCONN is returned. This of course is a kernel API
compat breakage, but let's not make a fuss and just map EINVAL to
ENOTCONN to make it recognizable the same way everywhere.
The original commit doesn't make sense to me, none of the listed units
have an [Install] section, they hence are not subject to enable/disable
and hence not preset either. This commit hence has no effect whatsoever,
let's undo it to avoid further confusion.
Resolved can't reliably determine on whether "it makes sense" to query
AAAA records when not explicitly specifying it in the request, so we
shouldn't remove them.
After having done the resolving, applications can use RFC6724 to
determine whether that address is reachable.
We can't know whether an address is reachable before having resolved it
and inspecting the routing table, and not resolving AAAA just because
there's no IPv6 default route on the main interface link them breaks
various setups, including IPv6-providing wireguard tunnels on a
non-dualstacked environment.
Andreas Rammhold [Wed, 13 May 2020 22:54:37 +0000 (00:54 +0200)]
networkd: Add support for setting a preferred subnet id for IPv6 PD leases
This allows users to configure a subnet id that should be used instead
of automatically (sequentially) assigned subnets. The previous attempt
had the downside that the subnet id would not be the same between
networkd restarts. In some setups it is desirable to have predictable
subnet ids across restarts of services and systems.
The code for the assignment had to be broken up into two pieces. One of
them is the old (sequential) assignment of prefixes and the other is the
new assignment based on configured subnet ids. The new assignment code
has to be executed first and has to be taken into account when (later
on) allocating the "old" subnets from the same pool.
Instead of having one iteration through the links we are now trying to
allocate a prefix for every link on every delegated prefix, unless they
received an assignment in a previous iteration.
We'd start writing an entry line, then another one, then another one,
and then output the rest of the first one, and then some other random
stuff, and the rest of some other lines... Results were ...eh... random.
Let's define a helper to avoid some of the copy&paste madness, and separate
blocks that output a single line with /**********************************/.
This rework doesn't change what data is written, it only tries to fix the
format of the output. The fact that some entries only write data from
link->network, and some from either link->network or link, some stuff only
for dhpc4 leases while some for both dhpc4 and dhcp6, etc, looks rather
suspicious too, but I didn't touch this.
busctl: verify args early and always print results to stdout
We would print the error sometimes to stdout and sometimes to stderr. It *is*
useful to get the message if one of the names is not found on the bus to
stdout, so that this shows out in the pager. So let's do verification of args
early to catch invalid arguments, and then if we receive an error over the bus
(most likely that the name is not activatable), let's print to stdout so it
gets paged. E.g. 'busctl tree org.freedesktop.systemd1 org.freedesktop.systemd2'
gives a nicely usable output.
network: allow empty assignment to PreferredLifetime=
Users might want to use that to unset a previous setting. The docs seem OK as
they are: we don't need to explictly mention the empty value, since it is
almost always allowed.
network: use consistent type when parsing lifetimes
Those fields are both uint32_t, so we should use the same type when parsing.
Having a different type didn't change the result, but let's be consistent.
sd-bus: internalize setting of bus is_system/is_user
Each of bus_set_address_{user,system} had two users, and each of the two users
would set the internal flag manually. We should do that internally in the
functions instead.
While at it, only set the flag when setting the address is actually successful.
This doesn't change anything for current users, but it seems more correct.
Those are fairly trivial to reimplement, but any non-trivial user of sd-bus
is likely to need them. So let's expose them to save everyone the trouble.
I'm keeping the internal functions and making the public ones thin wrappers,
because for the internal uses we don't need the additional asserts, and also we
can't expose _pure_ annotation easily, and dropping it would likely make the
compiled code a bit less efficient.