[Swan-commit] Changes to ref refs/heads/master

[Paul Wouters]

New commits:
commit 447a362a71287396824958d4e45f2eaa781ddcff
Author: Paul Wouters <pwouters at redhat.com>
Date:   Sun Feb 10 22:34:07 2019 -0500

    documentation: updated CHANGES

commit 63da55e25ac2881657e1ab2720a4bbff488181c7
Author: Paul Wouters <pwouters at redhat.com>
Date:   Sun Feb 10 22:09:00 2019 -0500

    pluto: In IKEv2, do not use delete_dpd_event() as that is IKEv1 only

commit 97fb563c2570c38d1b2470f76e5393ba79639daf
Author: Paul Wouters <pwouters at redhat.com>
Date:   Sun Feb 10 22:18:07 2019 -0500

    IKEv2: in delete_or_replace_state(), do not consider the event type
    
    There was a check for EVENT_SA_REPLACE which was no longer true after
    a recent change to use EVENT_SA_REKY, causing an auto=start connection
    receiveing a delete to no longer initiate the connection from scratch.
    
    Instead of adding or changing the check, it was removed because:
    
    - Whether the connecion is replaced or deleted should not depend on
      an event type, but on a policy of the connection
    - Converting the additional check to use POLICY_DONT_REKEY makes no
      sense, because it is not about to start a rekey, but initiate a
      fresh connection. This scenario maps more closely to auto=start|add
      than to rekey=yes|no

commit 3a7f77d549c28f03c46e7f6053c44772bee21157
Author: Paul Wouters <pwouters at redhat.com>
Date:   Sun Feb 10 22:06:57 2019 -0500

    pluto: initialize stf_status to prevent compile confusion
    
    In ikev2_process_sa_payload() the compiler gets confused when compiled
    with -Og and thinks stf_status could be returned unset. It is wrong.
    So just set it to STF_FAIL to be nice.