From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755515AbaFKNn5 (ORCPT ); Wed, 11 Jun 2014 09:43:57 -0400 Received: from forward4m.mail.yandex.net ([37.140.138.4]:49606 "EHLO forward4m.mail.yandex.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751328AbaFKNn4 (ORCPT ); Wed, 11 Jun 2014 09:43:56 -0400 From: Kirill Tkhai To: Srikar Dronamraju Cc: "linux-kernel@vger.kernel.org" , Peter Zijlstra , Ingo Molnar , Tkhai Kirill In-Reply-To: <20140611131536.GB21191@linux.vnet.ibm.com> References: <20140611093417.27807.2288.stgit@tkhai> <1402480330.32126.14.camel@tkhai> <20140611112411.GA21191@linux.vnet.ibm.com> <3732251402489254@web2m.yandex.ru> <20140611131536.GB21191@linux.vnet.ibm.com> Subject: Re: [PATCH 1/2] sched: Rework migrate_tasks() MIME-Version: 1.0 Message-Id: <4032471402494232@web2m.yandex.ru> X-Mailer: Yamail [ http://yandex.ru ] 5.0 Date: Wed, 11 Jun 2014 17:43:52 +0400 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 11.06.2014, 17:15, "Srikar Dronamraju" : >>> š* Kirill Tkhai [2014-06-11 13:52:10]: >>>> ššCurrently migrate_tasks() skips throttled tasks, >>>> ššbecause they are not pickable by pick_next_task(). >>> šBefore migrate_tasks() is called, we do call set_rq_offline(), in >>> šmigration_call(). >>> >>> šShouldnt this take care of unthrottling the tasks and making sure that >>> šthey can be picked by pick_next_task(). >> šIf we do this separate for every class, we'll have to do this 3 times. >> šFurthermore, deadline class does not have a list of throttled tasks. >> šSo we'll have to the same as I did: to lock tasklist_lock and to iterate >> šthrow all of the tasks in the system just to found deadline tasks. > > I think you misread my comment. > > Currently migrate_task() gets called from migration_call() and in the > migration_call() before migrate_tasks(), set_rq_offline() should put > tasks back using unthrottle_cfs_rq(). > > So my question is: Why are these tasks not getting unthrottled > through we are calling set_rq_offline? To me set_rq_offline is > calling the actual sched class routines to do the needful. > > I can understand about deadline tasks, because we don't have a deadline > But thats the only tasks that we need to fix. Hm, I tested that on fair class tasks. They used to disappear from /proc/sched_debug and used to hang. I'll check all once again. I'm agree with you, if set_rq_offline() already presents, we should use it. /me went to clarify why it does not work in my test. >>>> ššThese tasks stay on dead cpu even after they >>>> ššbecomes unthrottled. They are not schedulable >>>> šštill user manually changes their affinity or till >>>> ššcpu becomes alive again. >>> šIf we are still seeing tasks not being picked by pick_next_task(), then >>> šcan it probably mean that rq->rd was NULL? >> šUnthrottle functions dl_task_timer() and unthrottle_cfs_rq() put tasks and >> šqueues back. They do not look at rq->rd. > > What I meant was only if rq->rd isn't set, then we don't call > set_rq_offline, which seems very reasonable. > -- > Thanks and Regards > Srikar Dronamraju