> I was actually going to suggest passing a flag to startup rather than a > new function. Heh, after that e-mail to Ali, now I'm starting to believe that we need construction, init, unserialization, post unserialization init with flag, and startup(). (Don't forget that regStats happens in there too. It currently happens before the unserialization.) I'd have to think about it, but the multiphase stuff might be unnecessary if the second phase were all done in post unserialization init (which is quite likely where we wanted it to happen anyway).
Nate _______________________________________________ m5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/m5-dev
