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

[Andrew Cagney]

New commits:
commit 3c5e07ca1e2f574b3818456f291ec5e61821a6f7
Merge: b05b885 249d2d1
Author: Andrew Cagney <cagney at gnu.org>
Date:   Tue Nov 19 22:32:36 2019 -0500

    Revert botched push

commit 249d2d16a8c2f26078a9f9502cd108a20c3208de
Author: Andrew Cagney <cagney at gnu.org>
Date:   Tue Nov 19 22:28:29 2019 -0500

    Revert botched push
    
    Reverts 6d791e9641f2fb22364ac43173f180174d1873be ^ce8a1f9c2bb62a31844889a8312c137d46163dbd

commit b05b885b684195d9971d19c23538997527e9f716
Author: Andrew Cagney <cagney at gnu.org>
Date:   Tue Nov 19 21:45:41 2019 -0500

    testing: expect 'v2-auth-hash-policy: none'

commit 0596b4fbeeecf5aeb02b80f70e7128570f4de77b
Author: Andrew Cagney <cagney at gnu.org>
Date:   Tue Nov 19 21:20:23 2019 -0500

    ikev2: set the REPLACE timeout to 2*retransmit timeout when sending IKE_AUTH request
    
    Both retransmit and replace (for half-open IKE) had the same
    value 60s leading to either being delivered first.
    
    If the replace is delivered first, whack was left hanging (the replace
    keeps the socket open, even when the retransmit is trying to close
    it).  Slowing down the replace in once sense avoids the problem but in
    another sense fixes it as the retransmit timeout should take priority.