From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755896AbXGHNlP (ORCPT ); Sun, 8 Jul 2007 09:41:15 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752914AbXGHNlD (ORCPT ); Sun, 8 Jul 2007 09:41:03 -0400 Received: from il.qumranet.com ([82.166.9.18]:55818 "EHLO il.qumranet.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752624AbXGHNlB (ORCPT ); Sun, 8 Jul 2007 09:41:01 -0400 Message-ID: <4690E973.7000606@qumranet.com> Date: Sun, 08 Jul 2007 16:41:07 +0300 From: Avi Kivity User-Agent: Thunderbird 2.0.0.0 (X11/20070419) MIME-Version: 1.0 To: Ingo Molnar CC: kvm-devel@lists.sourceforge.net, linux-kernel@vger.kernel.org, shaohua.li@intel.com Subject: Re: [PATCH][RFC] kvm-scheduler integration References: <11838994974161-git-send-email-avi@qumranet.com> <20070708133539.GA12597@elte.hu> In-Reply-To: <20070708133539.GA12597@elte.hu> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Ingo Molnar wrote: > * Avi Kivity wrote: > > >> +#ifdef CONFIG_SCHED_KVM >> +static __read_mostly struct sched_kvm_hooks kvm_hooks; >> +#endif >> > > please just add a current->put_vcpu() function pointer instead of this > hooks thing. > > Won't that increase task_struct (16 bytes on 64-bit) unnecessarily? The function pointers are common to all virtual machines. >> static inline void prepare_task_switch(struct rq *rq, struct task_struct *next) >> { >> + unload_kvm_vcpu(current); >> prepare_lock_switch(rq, next); >> prepare_arch_switch(next); >> } >> @@ -1860,6 +1912,7 @@ static inline void finish_task_switch(struct rq *rq, struct task_struct *prev) >> prev_state = prev->state; >> finish_arch_switch(prev); >> finish_lock_switch(rq, prev); >> + reload_kvm_vcpu(current); >> > > ok, this looks certainly cheap enough from a scheduler POV, and it > cleans up the whole KVM/scheduling interaction quite nicely. (I'd not > bother with tweaking the migration logic, there's enough incentive for > the scheduler to keep tasks from migrating unnecessarily.) > Okay. -- error compiling committee.c: too many arguments to function