From: Jeff Moyer Subject: Re: [PATCH 1/4] cfq-iosched: Keep track of average think time for the sync-noidle workload. Date: Wed, 14 Apr 2010 19:06:38 -0400 Message-ID: References: <1271279826-30294-1-git-send-email-jmoyer@redhat.com> <1271279826-30294-2-git-send-email-jmoyer@redhat.com> <20100414213745.GC3167@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: jens.axboe@oracle.com, linux-kernel@vger.kernel.org, linux-ext4@vger.kernel.org To: Vivek Goyal Return-path: Received: from mx1.redhat.com ([209.132.183.28]:13210 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756198Ab0DNXGo (ORCPT ); Wed, 14 Apr 2010 19:06:44 -0400 In-Reply-To: <20100414213745.GC3167@redhat.com> (Vivek Goyal's message of "Wed, 14 Apr 2010 17:37:45 -0400") Sender: linux-ext4-owner@vger.kernel.org List-ID: Vivek Goyal writes: > On Wed, Apr 14, 2010 at 05:17:03PM -0400, Jeff Moyer wrote: >> @@ -3678,11 +3710,14 @@ static void *cfq_init_queue(struct request_queue *q) >> >> /* Init root service tree */ >> cfqd->grp_service_tree = CFQ_RB_ROOT; >> + cfqd->grp_service_tree.last_end_request = jiffies; >> > > This assignment is not required as we never update think time analsys of > service tree where all the groups are hanging. So for grp_service_tree, > last_end_request can be zero and there will be no harm. OK, thanks. > Otherwise patch looks good to me. Can you please run some simple blkio > cgroup tests to make sure that functionality is not broken. Yes, I'll add that to my list and get back to you. Thanks for the review, Vivek. Cheers, Jeff