There's no "answer_auxiliary" object anymore, it's all one "answer"
object, and we have per-item flags that tell us which section things are
from, i.e. from the main answer section, or the additional or
authoritative ones.
CNAME following was broken by 775ae35403
where we'd not properly collect RRs along the CNAME path. Good thing
though is that we don't have to anymore: since we nowadays propagate all
sections of the upstream replies into the cache and back to stub clients
all the information should already be available anyway, and there's no
need for us to collect it.
Fixes: #18690
This follows the addition of DEFAULT_HOSTNAME= in os-release.
The distinction between the value from os-release or the env var and
the compile-time setting is not made in the api: HostnameSource is
"default" is all cases. I think that this level of detail is not needed,
because the users of this mostly care whether the hostname was set by
user configuration or not.
This provides a fairly comprehensible fix for
https://bugzilla.redhat.com/show_bug.cgi?id=1893417.
This adds yet-another level of configuration:
- /etc/hostname
- transient hostname
- $SYSTEMD_DEFAULT_HOSTNAME
- DEFAULT_HOSTNAME is os-release
- -Dfallback-hostname=
- "linux"
It's a lot of layers, but each has it's own justification.
See https://bugzilla.redhat.com/show_bug.cgi?id=1893417 for the back story:
the fallback hostname matters a lot in certain environments. Right now the only
way to configure the fallback hostname is by recompiling systemd, which is
obviously problematic in case when the fallback hostname shall differ between
different editions of the same distro that share a single compiled rpm.
By making this overridable through an envvar, we're providing an escape hatch
without making this a top-level api. Later on a way to set this through
os-release is added, but I think the approach with the variable is still
useful. It it very convenient for testing, or to override settings only in a
particular service, etc.
This is useful for various variables that modify process behaviour. This makes
it easy to set it for pid1 without touching the kernel command line. Even for
the *user manager* this also can be convenient for the unprivileged user, who
cannot modify user@.service definition.
Variables that could be set like this include $SD_EVENT_PROFILE_DELAYS,
$SYSTEMD_FALLBACK_HOSTNAME, $SYSTEMD_MEMPOOL, $SYSTMED_RDRAND, etc.
The stub DNS server is authoritative for the RRs we synthesize, such as
localhost, _gateway, and entries from /etc/hosts, and also for trust anchors.
Partially fixes https://github.com/systemd/systemd/issues/17972
This changes the paths we read user manager config from in two ways:
- split-usr-root paths are dropped. The user manager is a poster boy for
non-early-boot, so reading dropins only from /usr is appropriate.
- we look at ~/.config/systemd/user.conf. Users should be allowed to override
their own config.
As user managers become more and more used, it becomes more important for users
to customize their own daemon. By reading from ~/.config, this is possible
without privileges.
As sd_device_get_sysattr_value() may be already used to read uevent
file. The read value of uevent file may be usually quite different
from the action string.
Luca Boccassi
Zbigniew Jędrzejewski-Szmek
Lennart Poettering
Yu Watanabe
Sergey Bugaev