Hi all, A brief update for the few people interested in the InOrder model working in full system mode (okay maybe just me!).
For ALPHA_FS, I've updated support for interrupt handling, fault handling, and alpha-specific functionality (e.g. hwrei) to the InOrder model. There are about 20 patches or so that I will eventually push to the repo but I'd rather get everything working first then push things bit by bit. The model is *maddeningly* close to booting linux. If the output messages are a indicator of progress, then out of 108 lines in our 10.linux-boot regression, 105 are printing out correctly before things break. Currently, it seems as though the model livelocks on a fault in the data translation buffer (e.g. DTB "dfault" on addr X, then 10000 cycles later the same fault on addr X, over and over). Unfortunately, I wont be able to get back to this until the early part of next week but in my view progress is coming along pretty good and is on track to be ready for the big ruby-m5 merge. That is all! -- - Korey _______________________________________________ m5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/m5-dev
