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

[Paul Wouters]

New commits:
commit 896957f821355ddbfab834b6cbd4d1944c81035c
Author: Paul Wouters <pwouters at redhat.com>
Date:   Sun Jun 14 15:32:07 2015 -0400

    testing: updated most kev2-liveness-0* tests
    
    But not ikev2-liveness-07 bcause I think that still shows a problem

commit 150b8c9d4d73eca5d820916bfe54a689cd08c6e4
Author: Paul Wouters <pwouters at redhat.com>
Date:   Sun Jun 14 00:17:24 2015 -0400

    pluto: ikev2 retransmit detection logic was broken
    
    It would false mark informational responses (after the first two) as
    duplicates.
    
    It would also flag duplicates when receiving an IKEv2 Request, and
    not answer, but that would only lead to receibing a duplicated request.
    So it is better to just process the duplicate on the first go and answer.

commit 132b01f472abbc4228c2c6191d5616769b813a9c
Author: Paul Wouters <pwouters at redhat.com>
Date:   Sun Jun 14 00:14:43 2015 -0400

    pluto: ikev2_send_informational() did not update msgid counters
    
    It cannot do this via ikev2_update_msgid_counters() because it
    has no pointer to the message digest (md)
    
    But since we know we are the initiator for this exchange, we know
    which state counters to update, so just do it directly.

commit 44e3bdbbafb219e54bcc3d49c5fd078dda13b006
Author: Paul Wouters <pwouters at redhat.com>
Date:   Sun Jun 14 00:11:48 2015 -0400

    pluto: process_encrypted_informational_ikev2() fix for msgid counter
    
    Only some branches in process_encrypted_informational_ikev2() that
    needed to call ikev2_update_msgid_counters() did so.
    
    To simplify things, don't call these in the branches but just at the
    end of the function. There is one catch, one branch deletes the state,
    so in ikev2_update_msgid_counters() check for st == NULL and just
    do nothing.