Previously, if an [Address] section is configured with a null address,
e.g. Address=0.0.0.0/24, then we acquired a free address in
link_request_address().
With this commit, we queue a request with the null address as is, and
acquire a free address later in address_process_request(). Similary,
now IPv4ACD daemon is configured in address_process_request().
With this change, we can make the address acquisition depend on other
conditions, e.g. if the persistent storage is ready or not.
Otherwise, even if the persistent storage is not ready, the DHCP server
may be started e.g. by unplugging and plugging cable.
Follow-up for 5582b36c38.
- prefix length and preference should be fit in uint8_t, and actually
the kernel and networkd uses uint8_t to store them.
- captive portal is now stored as a NUL-terminated string. Hence, it
is not necessary to also provide its length.
It is not used in this commit, but will be used for parsing NDisc
options in Router Advertisement message and friends.
The parser does mostly equivalent to what currently we do in
sd-ndisc-router.c. Several notable differences are:
- also perse source and target link-layer address,
- refuse multiple captive portals,
- check if the captive portal is in safe characters, as previously we
checked that in networkd-ndisc.c,
- dedup prefixes, routes, and pref64,
- limit the total number of options, for safety.
ProtectSystem=full remounts /boot and/or /efi read-only, but that
may trigger automount for the paths and delay the service being started.
===
systemd[1]: boot.automount: Got automount request for /boot, triggered by 720 ((networkd))
===
The service does not need to access the paths, so let's hide them.
Follow-up for f90eb08627.
Fixes#31742.
Then, we can read the lease file on restart, and the DHCP server will be
able to manage previously assigned addresses.
To save leases in the state directory /var/lib/systemd/network/, this
adds systemd-networkd-dhcp-server.service, and by default
systemd-networkd does not start the DHCP server without the heler
service started.
Closes#29991.
Currently, the received hostname is not used for assigning an address to
the host, or options in the subsequent reply message. But, the parsed
hostname is exposed through DBus, and possibly Varlink in the future.
Let's ignore spurious hostname.
We oftem save parsed DHCP options into a file, or expose them
through DBus or Varlink. In such case, control characters or non-UTF8
characters may cause many kind of unexpected errors. In general, a DHCP
message that have string options with spurious characters is mostly
malformed or broken. Let's refuse them.
This also makes dhcp_option_parse_string() do not free 'ret' argument,
to follow our usual coding style. So, callers now need to free the
pre-exisitng string if necessary.
Fixes#31708.
Then, this introduces systemd-networkd-persistent-storage.service.
systemd-networkd.service is an early starting service. So, at the time
it is started, the persistent storage for the service may not be ready,
and we cannot use StateDirectory=systemd/network in
systemd-networkd.service.
The newly added systemd-networkd-persistent-storage.service creates the
state directory for networkd, and notify systemd-networkd that the
directory is usable.
Otherwise, even if the interface is available, the requested config may
not be applied to the interface yet.
This also merges multiple tests for RPS setting. Hopefully the
performance of the test is improved.
`networkctl lldp` and `networkctl status INTERFACE` now use varlink
call to the networkd to query LLDP neighbors.
Then, this allows to dump LLDP neighbors in JSON format.
Co-authored-by: Tomáš Pecka <tomas.pecka@cesnet.cz>
This effectively reverts 9175002864.
The retrans time field in RA message is for neighbor solicitation,
and the commit d4c8de21a0 makes the value
assigned to the correct sysctl property.
Let's deprecate the option, and drop the redundant functions.
Previously, the function supports packets without IPv6 sender address
for unit tests. However, now unit tests use their own version of
icmp6_receive(). Hence, let's make the check more strict.
- split out radv_setup_recv_event(),
- slightly update log messages,
- use DIV_ROUND_UP(),
- use structured initializer more.
No functional change, just preparation for later commits.
- update several log messages,
- use event_reset_time_relative(),
- split out ndisc_setup_recv_event() and ndisc_setup_timer().
No functional change, just refactoring and preparation for later commits.
This deprecates IPForward= setting, which unconditionally controled
the global setting, even though it is a setting in .network file.
Instead, this introduces new IPv4Forwarding= and IPv6Forwarding=
settings both in .network and networkd.conf.
If these settings are specified in a .network file, then the
per-interface forwarding setting will be configured.
If specified in networkd.conf, then the global IP forwarding setting will
be configured.
Closes#30648.
Yu Watanabe
