> On 2011-09-10 08:08:28, Ali Saidi wrote: > > src/cpu/inorder/inorder_dyn_inst.cc, line 320 > > <http://reviews.m5sim.org/r/851/diff/1/?file=15002#file15002line320> > > > > Why not have this be a if () instead of a #if? I'm assuming the > > ultimate plan is to make FULLL_SYSTEM a boolean and not a processor > > directive. > > > >
That is the plan, but what I'm going for here is to make part of the exec_context consistent across modes so I can combine things in the ISA descriptions. By using an #if instead of an if (), I can constrain my change to just the exec_contexts and not have to also modify all the backing objects. In the long term I want to change how system calls are routed anyway, so it wouldn't necessarily be useful to push the change farther back. Also, these SE+FS changes are going to tend to be bite sized. Since being able to specialize for SE or FS has been really deeply baked into certain things over the years, it's easy for these changes to spread and spread and spread and get to be entirely unmanageable if they aren't constrained carefully. It's true that SE and FS aren't fundamentally that different, but where they are different there are little threads that spider all over the place. - Gabe ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: http://reviews.m5sim.org/r/851/#review1520 ----------------------------------------------------------- On 2011-09-10 03:12:58, Gabe Black wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > http://reviews.m5sim.org/r/851/ > ----------------------------------------------------------- > > (Updated 2011-09-10 03:12:58) > > > Review request for Default, Ali Saidi, Gabe Black, Steve Reinhardt, and > Nathan Binkert. > > > Summary > ------- > > Syscall: Make the syscall function available in both SE and FS modes. > > In FS mode the syscall function will panic, but the interface will be > consistent and code which calls syscall can be compiled in. This will allow, > for instance, instructions that use syscall to be built unconditionally but > then not returned by the decoder. > > > Diffs > ----- > > src/cpu/inorder/inorder_dyn_inst.hh 9678812ccb62 > src/cpu/inorder/inorder_dyn_inst.cc 9678812ccb62 > src/cpu/o3/dyn_inst.hh 9678812ccb62 > src/cpu/o3/dyn_inst_impl.hh 9678812ccb62 > src/cpu/simple/base.hh 9678812ccb62 > > Diff: http://reviews.m5sim.org/r/851/diff > > > Testing > ------- > > > Thanks, > > Gabe > > _______________________________________________ gem5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/gem5-dev
