> I don't know for sure, but I would guess that would be a lot less > efficient and more complicated. traceData gets used for every output > operand so you'd be calling through the exec context a lot, even when > it's just going to return NULL. This is how the isa_parser has > generated this code since I started working with it back with SPARC > and probably before that. The real issue is that initiateAcc is > getting the rug pulled out from under it before it finishes running, > so even if it came from the exec context you're trace information > would be printed and discarded prematurely.
Realistically, I don't see it being more complicated since you'd just be replacing traceData with foo->traceData. Efficiency is questionable as well since that traceData had to come from somewhere. Futhermore, the traceData parameter has to be provided even in m5.fast which means we pay extra then. Steve, do you have any idea why this happened? My guess is that it's just historical and that traceData came from somewhere else at some point. That said, if Steve's solution works, we can leave the status quo alone. Nate _______________________________________________ m5-dev mailing list m5-dev@m5sim.org http://m5sim.org/mailman/listinfo/m5-dev
