It should be accessible through physmem->start() or something similar. Ali
On Feb 9, 2012, at 6:56 PM, Gabriel Michael Black wrote: > I'd say you're the victim of a bad merge. Would you mind adding the > initialization back in? We'd be implicitly making the assumption that > physical memory starts at 0 which may not always be correct. Can the actual > start be extracted from the physmem thing? > > Gabe > > Quoting Nilay Vaish <[email protected]>: > >> I had the wrong changeset number. It should be 8799. >> >> http://repo.gem5.org/gem5/diff/dac1e33e07b0/src/sim/system.cc >> >> -- >> Nilay >> >> On Thu, 9 Feb 2012, Ali Saidi wrote: >> >>> >>> >>> I don't see that change in the repository. Do you have the revision >>> id? >>> >>> Ali >>> >>> On 09.02.2012 12:31, Nilay Vaish wrote: >>> >>>> This changeset >>> removes the statement that initializes pagePtr variable. >>>> This can, >>> and in my case it does, result in pagePtr getting arbitrary >>>> value >>> that may lead the simulator to believe that it is out of physical >>>> >>> memory. >>>> >>>> -- >>>> Nilay >>>> >>> _______________________________________________ >>>> gem5-dev mailing >>> list >>>> [email protected] >>>> http://m5sim.org/mailman/listinfo/gem5-dev >>> >>> >>> _______________________________________________ >>> gem5-dev mailing list >>> [email protected] >>> http://m5sim.org/mailman/listinfo/gem5-dev >>> >> _______________________________________________ >> gem5-dev mailing list >> [email protected] >> http://m5sim.org/mailman/listinfo/gem5-dev >> > > > _______________________________________________ > gem5-dev mailing list > [email protected] > http://m5sim.org/mailman/listinfo/gem5-dev > _______________________________________________ gem5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/gem5-dev
