On Sat, Aug 12, 2017 at 02:39:08AM +1000, Nicholas Piggin wrote: > POWER9 CPUs have independent MMU contexts per thread, so KVM does not > need to quiesce secondary threads, so the hwthread_req/hwthread_state > protocol does not have to be used. So patch it away on POWER9, and patch > away the branch from the Linux idle wakeup to kvm_start_guest that is > never used. If/when we add support for running HPT guests on a radix host, we will have to run the host in single-threaded mode (since POWER9 doesn't support having some threads of a core using HPT and some using radix simultaneously). We'll then need some sort of thing like kvmppc_grab_hwthread to coordinate with the threads so that guests can use the secondary threads. So I think most of this code should stay. We will still need to have a way to make sure that the secondaries are in real mode and not in a guest, because all threads will need to be in real mode when switching the core between radix and HPT mode. Maybe we can optimize it a bit at present given that we don't yet support running HPT guests on a radix host, but I don't want to make it harder to do that in future. Paul. -- To unsubscribe from this list: send the line "unsubscribe kvm-ppc" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
