----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: http://reviews.gem5.org/r/1749/#review4127 -----------------------------------------------------------
Looks good, again just a few minor suggestions. src/cpu/kvm/base.hh <http://reviews.gem5.org/r/1749/#comment3966> Is this comment really accurate? It seems like sometimes the up-to-date state is stored in the thread context, else you wouldn't have to update the kvm state... src/cpu/kvm/base.hh <http://reviews.gem5.org/r/1749/#comment3968> How about syncThreadContext() and syncKvmState()? src/cpu/kvm/base.hh <http://reviews.gem5.org/r/1749/#comment3967> I'd call this threadContextDirty so it's parallel with updateThreadContext() the way kvmStateDirty is parallel with updateKvmState(). src/cpu/kvm/base.cc <http://reviews.gem5.org/r/1749/#comment3969> What do you think of adding this here: assert(!m5ContextDirty); There's an invariant here that at most one of m5ContextDirty and kvmStateDirty should be true at the same time, and it would be nice to check that once in a while. src/cpu/kvm/base.cc <http://reviews.gem5.org/r/1749/#comment3970> Similarly you could assert !kvmStateDirty here. - Steve Reinhardt On March 9, 2013, 5 p.m., Ali Saidi wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > http://reviews.gem5.org/r/1749/ > ----------------------------------------------------------- > > (Updated March 9, 2013, 5 p.m.) > > > Review request for Default. > > > Description > ------- > > Changeset 9621:4560fb29509f > --------------------------- > kvm: Avoid synchronizing the TC on every KVM exit > > Reduce the number of KVM->TC synchronizations by overloading the > getContext() method and only request an update when the TC is > requested as opposed to every time KVM returns to gem5. > > > Diffs > ----- > > src/cpu/base.hh e6347e559e8f > src/cpu/kvm/base.hh PRE-CREATION > src/cpu/kvm/base.cc PRE-CREATION > > Diff: http://reviews.gem5.org/r/1749/diff/ > > > Testing > ------- > > > Thanks, > > Ali Saidi > > _______________________________________________ gem5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/gem5-dev
