From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754505Ab0LPLEd (ORCPT ); Thu, 16 Dec 2010 06:04:33 -0500 Received: from casper.infradead.org ([85.118.1.10]:50201 "EHLO casper.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753880Ab0LPLEc convert rfc822-to-8bit (ORCPT ); Thu, 16 Dec 2010 06:04:32 -0500 Subject: Re: [patch 1/2] sched: move periodic share updates to entity_tick() From: Peter Zijlstra To: Paul Turner Cc: linux-kernel@vger.kernel.org, Ingo Molnar , Mike Galbraith , Linus Torvalds In-Reply-To: <20101216031038.067028969@google.com> References: <20101216031016.186364650@google.com> <20101216031038.067028969@google.com> Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 8BIT Date: Thu, 16 Dec 2010 12:03:43 +0100 Message-ID: <1292497423.6803.4572.camel@twins> Mime-Version: 1.0 X-Mailer: Evolution 2.30.3 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org I made that: --- Subject: sched: Move periodic share updates to entity_tick() From: Paul Turner Date: Wed, 15 Dec 2010 19:10:17 -0800 Long running entities that do not block (dequeue) require periodic updates to maintain accurate share values. (Note: group entities with several threads are quite likely to be non-blocking in many circumstances). By virtue of being long-running however, we will see entity ticks (otherwise the required update occurs in dequeue/put and we are done). Thus we can move the detection (and associated work) for these updates into the periodic path. This restores the 'atomicity' of update_curr() with respect to accounting. Signed-off-by: Paul Turner Signed-off-by: Peter Zijlstra LKML-Reference: <20101216031038.067028969@google.com> --- kernel/sched_fair.c | 21 +++++++++++++++++---- 1 file changed, 17 insertions(+), 4 deletions(-) Index: linux-2.6/kernel/sched_fair.c =================================================================== --- linux-2.6.orig/kernel/sched_fair.c +++ linux-2.6/kernel/sched_fair.c @@ -564,10 +564,6 @@ __update_curr(struct cfs_rq *cfs_rq, str #if defined CONFIG_SMP && defined CONFIG_FAIR_GROUP_SCHED cfs_rq->load_unacc_exec_time += delta_exec; - if (cfs_rq->load_unacc_exec_time > sysctl_sched_shares_window) { - update_cfs_load(cfs_rq, 0); - update_cfs_shares(cfs_rq, 0); - } #endif } @@ -809,6 +805,14 @@ static void update_cfs_shares(struct cfs reweight_entity(cfs_rq_of(se), se, shares); } + +static void update_entity_shares_tick(struct cfs_rq *cfs_rq) +{ + if (cfs_rq->load_unacc_exec_time > sysctl_sched_shares_window) { + update_cfs_load(cfs_rq, 0); + update_cfs_shares(cfs_rq, 0); + } +} #else /* CONFIG_FAIR_GROUP_SCHED */ static void update_cfs_load(struct cfs_rq *cfs_rq, int global_update) { @@ -817,6 +821,10 @@ static void update_cfs_load(struct cfs_r static inline void update_cfs_shares(struct cfs_rq *cfs_rq, long weight_delta) { } + +static inline void update_entity_shares_tick(struct cfs_rq *cfs_rq) +{ +} #endif /* CONFIG_FAIR_GROUP_SCHED */ static void enqueue_sleeper(struct cfs_rq *cfs_rq, struct sched_entity *se) @@ -1133,6 +1141,11 @@ entity_tick(struct cfs_rq *cfs_rq, struc */ update_curr(cfs_rq); + /* + * Update share accounting for long-running entities. + */ + update_entity_shares_tick(cfs_rq); + #ifdef CONFIG_SCHED_HRTICK /* * queued ticks are scheduled to match the slice, so don't bother