This commit introduces a mechanism to attach light weight analyzers to
the root analyzer of sessions in order to tap into the packets delivered
to child analyzer.
* origin/topic/awelzel/defer-more-stuff:
RecordType: Ensure &default fields are always re-initialized
Attr: Deprecate using &default and &optional together on record fields
RecordType: Allow deferring &default=vector(), set(), table() fields
This moves c$service_violation to the deprecated-dpd-log policy script.
This is the only script in the distribution that uses the field, and it
is unlikely to be used externally. It is also responsible for a
significant amount of memory use by itself.
This also restores the field being populated, which was broken in
GH-4362
The conn protocol scripts now assume that new_connection is run before
connection_state_remove. Update the DNS analyzer to raise the
new_connection event.
This class is a parameter of virtual methods of the Backend API for users
to implement and also a parameter to the HookPublishEvent() API. Seems it
shouldn't be in detail and instead we should own it.
Alternatively, could mark the cluster APIs as not-stable-yet, but I
think we can move forward and make it non-detail for 8.0.
* origin/topic/awelzel/4431-zeromq-drop-policy-v2:
cluster.bif: Improve Cluster::publish() docstring
btest/cluster/zeromq: Add tests for overload behavior
cluster/zeromq: Metric for msg errors
cluster/zeromq: Drop events when overloaded
cluster/zeromq: Comments and move lookups to InitPostScript()
cluster/zeromq: Rework lambdas to member functions
cluster/zeromq: Support local XPUB/XSUB hwm and buf configurability
cluster/OnLoop: Support DontBlock and Force flags for queueing
cluster/ThreadedBackend: Injectable OnLoopProcess instance
The overload-drop.zeek and overload-no-drop.zeek tests have proxy,
worker-1 and worker-2 publish to the manager topic. For the drop
case, we verify that both, the senders, but also the manager drops
events. For the no-drop test, the HWMs are set such that all events
are buffered.
The overload-worker-proxy-topic*.zeek tests are similar, but instead
of publishing to the manager topic, proxy, worker-1 and worker-2 publish
to the proxy and worker topics to overload each other. This had
previously resulted in lockups and these tests verify that this doesn't
happen anymore.
When either the XPUB socket's hwm is reached, or the onloop queue is
full, drop the events. Users can set ths xpub_sndhwm and
onloop_queue_hwm to 0 to avoid these drops at the risk of unbounded
memory growth.
Now that Conn::set_conn is guaranteed to be called at the beginning and
at the end of the connection, we can skip re-setting the elements that
we know will not have changed. This prevents repeated lookups, e.g. to
check that addresses are in the local networks.
During `connection_state_remove`, only the duration, number of
packets, service, and history fields are updated.
local_orig and local_resp are updated when the connection is flipped. A
test was added for that purpose. It uses the already existing
http.zeek-image-post-1080-8000-x.pcap, which was slightly rewritten for
this, so that one side of the connection has IP addresses different from
127.0.0.1.
The existing history-flip test also was updated to have one side being
in a local-net, to check that the flipping of local_orig and local_resp
works correctly at the beginning of a connection.
c$conn is often needed for connection events, but it being established
in connection_state_removed can be a problem because event handlers have
to remember to call Con::set_conn().
This commit moves to call Conn::set_conn() in new_connection.
Addresses GH-4202
update logs
