https://github.com/systemd/systemd/pull/14133 made
capability_ambient_set_apply() acquire capabilities that were explicitly
asked for and drop all others. This change means the function is called
even with an empty capability set, opening up a code path for users
without ambient capabilities to call this function. This function will
error with EINVAL out on kernels < 4.3 because PR_CAP_AMBIENT is not
understood. This turns capability_ambient_set_apply() into a noop for
kernels < 4.3
Fixes https://github.com/systemd/systemd/issues/15225
Functions called from device_setup_unit() already make sure that unit is
enqueued in case it is a new unit or properties exported on the bus have
changed.
This should prevent unnecessary DBus wakeups and associated DBus traffic
when device_setup_unit() was called while reparsing /proc/self/mountinfo
due to the mountinfo notifications. Note that we parse
/proc/self/mountinfo quite often on the busy systems (e.g. k8s container
hosts) but majority of the time mounts didn't change, only some mount
got added. Thus we don't need to generate PropertiesChanged for devices
associated with the mounts that didn't change.
Thanks to Renaud Métrich <rmetrich@redhat.com> for debugging the
problem and providing draft version of the patch.
journal_file_fstat() returns an error if we call it on already unlinked
journal file and hence we never reach remove_file_real() which is the
entire point.
I must have made some mistake while testing the fix that got me thinking
the issue is gone while opposite was true.
Fixes#14695
So far we had various ad hoc APIs to query search paths:
systemd-analyze unit-paths, lookup_paths_log(), the pkgconfig file,
debug logs emitted by systemd-analyze cat-config.
But answering a simple question "what is the search path for tmpfiles,
sysusers, .network files, ..." is surprisingly hard.
I think we should have an api that makes it easy to query this. Pkgconfig is
not bad, but it is primarily a development tool, so it's not available in many
context. Also it can't provide support for paths which are influenced by
environment variables, and I'd like to be able to answer the question "what is
the search path for ..., assuming that VAR_FOO=... is set?".
Extending sd-path to support more of our internal paths seems to be most
flexible solution. We already have systemd-path which provides a nice
way to query, and we can add stuff like optional descriptions later on.
We we essentially get a nice programmatic and commmandline apis for the price
of one.
The two functions were duplicating a lot of functionality and more
importantly, they both had explicit lists of types which are search
paths. I want to add more types, and I don't want to have to remember
to add them to both lists.
I think the two names were both pretty bad. They did not give a proper hint
what the difference between the two functions is, and sd_path_home sounds like
it is somehow related to /home or home directories or whatever, when in fact
both functions return the same set of paths as either a colon-delimited string
or a strv. "_strv" suffix is used by various functions in sd-bus, so let's
reuse that.
Those functions are not public yet, so let's rename.
The names with multiple lowercase words run together are hard to read. We
started that way with very short names like rootprefix, but then same pattern
was applied to longer and longer names. Looking at the body of .pc files
available on my machine, many packages use underscores; let's do the same. Old
names are kept for compatiblity, so this is backwards compatible.
In 1a29610f5f the change inadvertedly
disabled names with digit as the first character. This follow-up change
allows a digit as the first character in compat mode.
Fixes: #15141
When there are hundreds of mounts on the server, it will take a long
time to analyze the failure of a certain mount unit. So it is useful
to print the reason why unit_add_name() failed.
Add a simple check on the number of unit files that were found: return
`-ENOENT` when none is found from the function and thus `systemctl`
consequently exits with `1` (`EXIT_FAILURE`) if none were found.
Verification:
```bash
root@image:~# systemctl list-unit-files dbus-nonexistant.service; echo
$?
UNIT FILE STATE VENDOR PRESET
0 unit files listed.
1
root@image:~# systemctl list-unit-files dbus.service; echo $?
UNIT FILE STATE VENDOR PRESET
dbus.service static enabled
1 unit files listed.
0
```
Fixes#15082.
A common pattern in the codebase is reading a cgroup memory value
and converting it to a uint64_t. Let's make it a helper and refactor a
few places to use it so it's more concise.
Commit 08c588d18b moved attachment of the
sd-event into creation of the dhcp_client or ipv4ll in dhcp4_configure()
or ipv4ll_configure(), but these can also be created in link_load(),
so that creation needs to also perform sd-event attachment.
Without this, dhcp_client or ipv4ll created in link_load() will not have
an ->event and will fail assertion, causing networkd to fail, e.g.:
Assertion 'client->event' failed at src/libsystemd-network/sd-dhcp-client.c:1283, function client_start_delayed(). Ignoring.
ens2: Could not acquire DHCPv4 lease: Invalid argument
ens2: Failed
