-----BEGIN PGP SIGNED MESSAGE-----
>>>>> "Jeff" == Jeff Dike <jdike@...> writes:
>> For the situation with multiple UMLs, the amount of time skip forward
>> is the minima of each systems' view.
Jeff> OK, I got it. You want to skip the network from event to event with the
Jeff> outside tool not knowing what each UML has scheduled.
Jeff> I was considering the case that the tool knows how far ahead to skip, in which
Jeff> case, it can just tell the UMLs to do it.
I was considering this situation as well, and realized that I was going to
have to enhance the testing jig some more - it would have to advance the time
based upon feedback from either the network or the consoles, and my brain was
starting to hurt here.
Werner's suggestion of just invading the scheduler itself to axe out idle
time is a good one - it should work for many situations where the timeouts
being tested are static - i.e. not the result of a network make/break.
e.g. whenever one can really just run the simulation at normal speed and
just wait a long time.
I think that we (FreeSWAN) will still have to deal with changing the
network configuration half-way through the test. But, I hope that this will
just be playing additional sets of packets in the input interfaces.
] ON HUMILITY: to err is human. To moo, bovine. | firewalls [
] Michael Richardson, Sandelman Software Works, Ottawa, ON |net architect[
] mcr@... http://www.sandelman.ottawa.on.ca/ |device driver[
] panic("Just another NetBSD/notebook using, kernel hacking, security guy"); [
-----BEGIN PGP SIGNATURE-----
Comment: Finger me for keys
-----END PGP SIGNATURE-----