From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S965565Ab2FAQzT (ORCPT ); Fri, 1 Jun 2012 12:55:19 -0400 Received: from forward14.mail.yandex.net ([95.108.130.92]:47954 "EHLO forward14.mail.yandex.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S965541Ab2FAQzR (ORCPT ); Fri, 1 Jun 2012 12:55:17 -0400 X-Greylist: delayed 596 seconds by postgrey-1.27 at vger.kernel.org; Fri, 01 Jun 2012 12:55:16 EDT From: Kirill Tkhai To: Yong Zhang Cc: Steven Rostedt , "linux-kernel@vger.kernel.org" , Ingo Molnar , Peter Zijlstra In-Reply-To: <20120419085440.GC3963@zhy> References: <1334519122.8698.3.camel@hp> <1334592379.28106.4.camel@gandalf.stny.rr.com> <1334773976.28106.49.camel@gandalf.stny.rr.com> <1334783815.28106.56.camel@gandalf.stny.rr.com> <20120419085440.GC3963@zhy> Subject: Re: [sched/rt] Optimization of function pull_rt_task() MIME-Version: 1.0 Message-Id: <69791338569116@web3f.yandex.ru> Date: Fri, 01 Jun 2012 20:45:16 +0400 X-Mailer: Yamail [ http://yandex.ru ] 5.0 Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset=koi8-r Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org 19.04.2012, 12:54, "Yong Zhang" : > On Wed, Apr 18, 2012 at 05:16:55PM -0400, Steven Rostedt wrote: > >> šOn Wed, 2012-04-18 at 14:32 -0400, Steven Rostedt wrote: >>> šOn Mon, 2012-04-16 at 12:06 -0400, Steven Rostedt wrote: >>>> šOn Sun, 2012-04-15 at 23:45 +0400, Kirill Tkhai wrote: >>>>> šThe condition (src_rq->rt.rt_nr_running) is weak because it doesn't >>>>> šconsider the cases when src_rq has only processes bound to it (when >>>>> šsingle cpu is allowed). It may be running kernel thread like >>>>> šmigration/x etc. >>>>> >>>>> šSo it's better to use more stronger condition which is able to exclude >>>>> šabove conditions. The function has_pushable_tasks() complitely does >>>>> šthis. A task may be pullable for another cpu rq only if he is pushable >>>>> šfor his own queue. >>>> šI considered this before, and for some reason I never did the change. >>>> šI'll have to think about it. It seems like this would be the obvious >>>> šcase, but I think there was something not so obvious that caused issues. >>>> šBut I don't remember what it was. >>>> >>>> šI'll have to rethink this again. >>> šI can't find anything wrong with this change. Maybe things change, or I >>> šwas thinking of another change. >>> >>> šI'll apply it and start running my tests against it. >> šNot only does this seem to work fine, I took it one step further :-) > > Hmm... throttle doesn't handle the pushable list, so we may find a > throttled task by pick_next_pushable_task(). > > Thanks, > Yong I don't complitelly understand throttle logic. Is the source patch not-appliable the same reason? Kirill > >> šPeter, do you see anything wrong with this patch? >> >> š-- Steve >> >> šdiff --git a/kernel/sched/rt.c b/kernel/sched/rt.c >> šindex 61e3086..b44fd1b 100644 >> š--- a/kernel/sched/rt.c >> š+++ b/kernel/sched/rt.c >> š@@ -1416,39 +1416,15 @@ static int pick_rt_task(struct rq *rq, struct task_struct *p, int cpu) >> šš/* Return the second highest RT task, NULL otherwise */ >> ššstatic struct task_struct *pick_next_highest_task_rt(struct rq *rq, int cpu) >> šš{ >> š- struct task_struct *next = NULL; >> š- struct sched_rt_entity *rt_se; >> š- struct rt_prio_array *array; >> š- struct rt_rq *rt_rq; >> š- int idx; >> š+ struct plist_head *head = &rq->rt.pushable_tasks; >> š+ struct task_struct *next; >> >> š- for_each_leaf_rt_rq(rt_rq, rq) { >> š- array = &rt_rq->active; >> š- idx = sched_find_first_bit(array->bitmap); >> š-next_idx: >> š- if (idx >= MAX_RT_PRIO) >> š- continue; >> š- if (next && next->prio <= idx) >> š- continue; >> š- list_for_each_entry(rt_se, array->queue + idx, run_list) { >> š- struct task_struct *p; >> š- >> š- if (!rt_entity_is_task(rt_se)) >> š- continue; >> š- >> š- p = rt_task_of(rt_se); >> š- if (pick_rt_task(rq, p, cpu)) { >> š- next = p; >> š- break; >> š- } >> š- } >> š- if (!next) { >> š- idx = find_next_bit(array->bitmap, MAX_RT_PRIO, idx+1); >> š- goto next_idx; >> š- } >> š+ plist_for_each_entry(next, head, pushable_tasks) { >> š+ if (pick_rt_task(rq, next, cpu)) >> š+ return next; >> šššššššššš} >> >> š- return next; >> š+ return NULL; >> šš} >> >> ššstatic DEFINE_PER_CPU(cpumask_var_t, local_cpu_mask); >> >> š-- >> šTo unsubscribe from this list: send the line "unsubscribe linux-kernel" in >> šthe body of a message to majordomo@vger.kernel.org >> šMore majordomo info at šhttp://vger.kernel.org/majordomo-info.html >> šPlease read the FAQ at šhttp://www.tux.org/lkml/ > -- > Only stand for myself