[Swan-dev] Matt's changes to informational message handling

[Paul Wouters]

On Tue, 4 Feb 2014, D. Hugh Redelmeier wrote:

> I don't really understand this change.  I imagine it is correct.
>
> Would it be possible to get a test case that works correctly because
> of the change?
>
> I know it is asking a lot: only a few know how to make new tests.
> Maybe Paul could generate a test if you explained what should be
> tested.

I had already asked Matt for a test too. Especially since the change
seemed to hit a passert for me:

 	} else if (type == EVENT_v2_LIVENESS) {
 		passert(st->st_liveness_event == ev);

I am also looking now at why the connection came up with DPD=none.

It seems like we have st->st_dpd_local that will tell us whether we
should/will send out DPD/liveness probes. But it is hardly used,
and only set in ikev1. And the only consumer of this variable is
the DPD=xxxx line, so it might as well check it directly using:

 	if (st->st_connection->dpd_delay && st->st_connection->dpd_timeout)

which is how st->st_dpd_local gets set.

iekv2 has st->st_liveness that I still need to look into.

Paul