[Swan] With the find_ifaces() call moved from rcv_wack.c to server.c

[Philippe Vouters]

Paul,

If the root cause is a race condition as it appears to be, wouldn't the 
failure be caused by a Libreswan timer issue causing the oriented(c) to 
return false ?

Can you dig onto the pluto log files to pinpoint such an issue ? To 
pinpoint such an issue, rerun your westnet-eastnet-ipv4-psk-ikev2 test 
case with plutodebug=controlmore and in a second step with 
plutodebug="control controlmore" if not too verbose, missing relevant 
facts. Never use plutodebug=all.

Philippe Vouters (Fontainebleau/France)
URL: http://vouters.dyndns.org/
SIP: sip:Vouters at sip.linphone.org

Le 13/01/2013 17:09, Paul Wouters a écrit :
> On Sun, 13 Jan 2013, Philippe Vouters wrote:
>
>> With just the find_ifaces() call moved from rcv_wack.c to server.c, 
>> this gives this signature. This signature matches my roadwarrior tests.
>>
>> Would you be so kind as to rerun your full test set against Openswan 
>> 2.6.38 to ensure there are no regression caused by this find_ifaces() 
>> call move ?
>
> I am investigating a regression, the same orientation problem I've seen
> before, with Antony on testing.libreswan.org. But it's not very
> consistent - it seems to be a race condition that sometimes happens. So
> we are still trying to see if we can pinpoint the change. And we still
> need to test with openswan as well.
>
> Stay tuned,
>
> Paul
>