When we receive an update for a certain neighbor, the update
might be invalid and we want to reject it. However, we still
must create an invalid object to compare whether the update
causes a remove of a previously valid neighbor.
Let lldp_neighbor_new() create an instance as long as the
id fields are present.
Instead of replacing the whole hash with a new one (and all new by a new one,
LldpNeighbor instances), update the existing hash.
One point of this is that our process-all function requires less
comparisons and avoids duplicate work right earlier. E.g. if a neighbor
didn't change, we don't have to put it into a hash to compare later for
equality.
But more importantly, we preserve our LldpNeighbor instance instead
of recreating them all the time. Later, the LldpNeighbor will cache
the GVariant.
lldp_start_timer() was only called during sd_lldp_get_neighbors().
Ensure that the timer is (re-)started when a new neighbor appears.
Otherwise, the timer is not started when relying on the events alone.
https://github.com/systemd/systemd/pull/2826
Otherwise the connection wouldn't verify:
<error> [1458066126.2270] device (eth10): Generated connection does not verify:
connection.type: property type should be set to '802-3-ethernet'
<debug> [1458066126.2271] manager: (eth10): can't assume; no connection
(cherry picked from commit 4b71939e9ac3df93bfe72af0eac42b4ebaf94e15)
We used to pad the lifetime since the beginning (commit
f121995fad).
However, there is not race involved, since our platform cache
is in sync with the messages from kernel (which didn't used to
be the case).
Also, when receiving a RA with a zero preferred time, we must
not extend the address lifetime by 5 seconds, but instead deprecate
the address immediately.
https://bugzilla.gnome.org/show_bug.cgi?id=763513
The prune list is for elements that must be deleted from the list of
available connections. So, when processing all the existing
connections an element must be deleted from the prune list iff it's
available.
Fixes: 8b2abe0e2chttps://bugzilla.redhat.com/show_bug.cgi?id=1316488
Previously, the push/pop API to switch between namespaces would always
switch both the net and mount namespace together.
There are situations, where we want to only switch one namespace.
For example, the function nmp_netns_bind_to_path() introduced next
only wants to switch the net namespace to get /proc/self/ns/net,
but must not switch the mount namespace as it bind-mounds in the
namespace of the caller.
Make the test helper independent from the platform singleton instance.
That way, we can also use them for other platform instances (e.g. in a
different namespace).
GError is not used, the error branch would always result in NULL dereference.
Also, check for the result being zero for clarity -- it's the only allowed
success indication.
CID 75365 (#3 of 3): Explicit null dereferenced (FORWARD_NULL)
12. var_deref_op: Dereferencing null pointer error.
Don't use a "sd_" prefix for our local functions because
then it's not immediately clear which functions are ours
and which are from systemd-logind. Just rename them to have
a "st_sd_*" prefix ("st" like "session tracking").
Now we have:
"nm-sd.h" is a header file of NetworkManager with utilities
related to systemd. It can be used anywhere freely.
Also, systemd headers that are considered public API (like
"sd-event.h") can be used without restrictions.
When compiling the systemd sources, we always must include
"nm-sd-adapt.h" as first. Similarly, systemd headers must
not include "nm-sd-adapt.h", because they are either public
(in which case the adapter is not needed) or they are internal
(in which case they are themself included via a systemd source).
Sometimes, we must internal API (like "dhcp-lease-internal.h").
In this case, we also must include "nm-sd-adapt.h".
As the lldp API changed, adjust "nm-lldp-listener.c".
Note that the commit is not yet functional due to missing
sd_event_source_set_enabled() and sd_event_source_set_time().
