From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757667AbcGKFEE (ORCPT ); Mon, 11 Jul 2016 01:04:04 -0400 Received: from mx1.redhat.com ([209.132.183.28]:53648 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750829AbcGKFEA (ORCPT ); Mon, 11 Jul 2016 01:04:00 -0400 Reply-To: xlpang@redhat.com Subject: Re: [PATCH v3] sched/deadline: remove useless param from setup_new_dl_entity References: <1467977331-31116-1-git-send-email-juri.lelli@arm.com> To: Juri Lelli , peterz@infradead.org Cc: rostedt@goodmis.org, linux-kernel@vger.kernel.org, mingo@redhat.com, luca.abeni@unitn.it From: Xunlei Pang Message-ID: <578328BC.3080504@redhat.com> Date: Mon, 11 Jul 2016 13:03:56 +0800 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101 Thunderbird/38.2.0 MIME-Version: 1.0 In-Reply-To: <1467977331-31116-1-git-send-email-juri.lelli@arm.com> Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.30]); Mon, 11 Jul 2016 05:03:59 +0000 (UTC) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 2016/07/08 at 19:28, Juri Lelli wrote: > setup_new_dl_entity() takes two parameters, but it only actually uses > one of them, under a different name, to setup a new dl_entity, after: > > 2f9f3fdc928 "sched/deadline: Remove dl_new from struct sched_dl_entity" > > as we currently do > > setup_new_dl_entity(&p->dl, &p->dl) > > However, before Luca's change we were doing > > setup_new_dl_entity(dl_se, pi_se) > > in update_dl_entity() for a dl_se->new entity: we were using pi_se's > parameters (the potential PI donor) for setting up a new entity. > > Restore this behaviour (as we want to correctly initialize parameters of > a boosted task that enters DEADLINE) by removing the useless second > parameter of setup_new_dl_entity() and retrieving the top waiter > directly from inside that function. > > Cc: Ingo Molnar > Cc: Peter Zijlstra > Cc: Steven Rostedt > Cc: Luca Abeni > Signed-off-by: Juri Lelli > > --- > Changes from v2: > - optimize by calling rt_mutex_get_top_task only if the task is > boosted (as suggested by Steve) > > Changes from v1: > - Steve pointed out that we were actually using the second parameter > to permorm initialization > - Luca confirmed that behavior is slightly changed w.r.t. before his > change > - changelog updated and original behavior restored > --- > kernel/sched/deadline.c | 16 +++++++++++++--- > 1 file changed, 13 insertions(+), 3 deletions(-) > > diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c > index fcb7f0217ff4..7aeb21963d3c 100644 > --- a/kernel/sched/deadline.c > +++ b/kernel/sched/deadline.c > @@ -346,11 +346,12 @@ static void check_preempt_curr_dl(struct rq *rq, struct task_struct *p, > * one, and to (try to!) reconcile itself with its own scheduling > * parameters. > */ > -static inline void setup_new_dl_entity(struct sched_dl_entity *dl_se, > - struct sched_dl_entity *pi_se) > +static inline void setup_new_dl_entity(struct sched_dl_entity *dl_se) > { > struct dl_rq *dl_rq = dl_rq_of_se(dl_se); > struct rq *rq = rq_of_dl_rq(dl_rq); > + struct task_struct *pi_task; > + struct sched_dl_entity *pi_se = dl_se; > > WARN_ON(dl_time_before(rq_clock(rq), dl_se->deadline)); > > @@ -363,6 +364,15 @@ static inline void setup_new_dl_entity(struct sched_dl_entity *dl_se, > return; > > /* > + * Use the scheduling parameters of the top pi-waiter task, > + * if we have one from which we can inherit a deadline. > + */ > + if (dl_se->dl_boosted && > + (pi_task = rt_mutex_get_top_task(dl_task_of(dl_se))) && > + dl_prio(pi_task->normal_prio)) > + pi_se = &pi_task->dl; > + > + /* > * We use the regular wall clock time to set deadlines in the > * future; in fact, we must consider execution overheads (time > * spent on hardirq context, etc.). > @@ -1721,7 +1731,7 @@ static void switched_from_dl(struct rq *rq, struct task_struct *p) > static void switched_to_dl(struct rq *rq, struct task_struct *p) > { > if (dl_time_before(p->dl.deadline, rq_clock(rq))) > - setup_new_dl_entity(&p->dl, &p->dl); > + setup_new_dl_entity(&p->dl); I'm curious why we even call setup_new_dl_entity() for non-queued cases? It seems more reasonable to do it when it really gets queued. We can see that enqueue_task_dl()->update_dl_entity() also has the same update logic as switched_to_dl(). If so, for already queued and boosted cases, rt_mutex_setprio() will call enqueue_task() with ENQUEUE_REPLENISH set, so enqueue_dl_entity() ->replenish_dl_entity() will advance p->dl.deadline beforehand, see code: replenish_dl_entity(): if (dl_se->dl_deadline == 0) { dl_se->deadline = rq_clock(rq) + pi_se->dl_deadline; dl_se->runtime = pi_se->dl_runtime; } IOW, we don't need to handle !dl boosted cases in setup_new_dl_entity(). Regards, Xunlei > > if (task_on_rq_queued(p) && rq->curr != p) { > #ifdef CONFIG_SMP