Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757455Ab0KTCKE (ORCPT ); Fri, 19 Nov 2010 21:10:04 -0500 Received: from e1.ny.us.ibm.com ([32.97.182.141]:57518 "EHLO e1.ny.us.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756567Ab0KTCJW (ORCPT ); Fri, 19 Nov 2010 21:09:22 -0500 From: John Stultz To: lkml Cc: Dima Zavin , Ingo Molnar , Peter Zijlstra , =?UTF-8?q?Arve=20Hj=C3=B8nnev=C3=A5g?= , John Stultz Subject: [PATCH 5/5] sched: use the old min_vruntime when normalizing on dequeue Date: Fri, 19 Nov 2010 18:08:54 -0800 Message-Id: <1290218934-8544-6-git-send-email-john.stultz@linaro.org> X-Mailer: git-send-email 1.7.3.2.146.gca209 In-Reply-To: <1290218934-8544-1-git-send-email-john.stultz@linaro.org> References: <1290218934-8544-1-git-send-email-john.stultz@linaro.org> MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1872 Lines: 56 From: Dima Zavin After pulling the thread off the run-queue during a cgroup change, the cfs_rq.min_vruntime gets recalculated. The dequeued thread's vruntime then gets normalized to this new value. This can then lead to the thread getting an unfair boost in the new group if the vruntime of the next task in the old run-queue was way further ahead. CC: Ingo Molnar CC: Peter Zijlstra Cc: Arve Hjønnevåg Signed-off-by: Dima Zavin Signed-off-by: John Stultz --- kernel/sched_fair.c | 6 +++++- 1 files changed, 5 insertions(+), 1 deletions(-) diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index f4f6a83..72f19ad 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -802,6 +802,8 @@ static void clear_buddies(struct cfs_rq *cfs_rq, struct sched_entity *se) static void dequeue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags) { + u64 min_vruntime; + /* * Update run-time statistics of the 'current'. */ @@ -826,6 +828,8 @@ dequeue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags) if (se != cfs_rq->curr) __dequeue_entity(cfs_rq, se); account_entity_dequeue(cfs_rq, se); + + min_vruntime = cfs_rq->min_vruntime; update_min_vruntime(cfs_rq); /* @@ -834,7 +838,7 @@ dequeue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags) * movement in our normalized position. */ if (!(flags & DEQUEUE_SLEEP)) - se->vruntime -= cfs_rq->min_vruntime; + se->vruntime -= min_vruntime; } /* -- 1.7.3.2.146.gca209 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/