From mboxrd@z Thu Jan 1 00:00:00 1970 From: Vivek Goyal Subject: Re: [PATCH 12/12] cfq-iosched: add hierarchical cfq_group statistics Date: Tue, 18 Dec 2012 14:26:33 -0500 Message-ID: <20121218192633.GC25908@redhat.com> References: <1355524885-22719-1-git-send-email-tj@kernel.org> <1355524885-22719-13-git-send-email-tj@kernel.org> <20121218191117.GD24050@redhat.com> <20121218191425.GO1844@htj.dyndns.org> <20121218191854.GB25908@redhat.com> <20121218192155.GQ1844@htj.dyndns.org> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Content-Disposition: inline In-Reply-To: <20121218192155.GQ1844-Gd/HAXX7CRxy/B6EtB590w@public.gmane.org> List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: containers-bounces-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org Errors-To: containers-bounces-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org To: Tejun Heo Cc: axboe-tSWWG44O7X1aa/9Udqfwiw@public.gmane.org, ctalbott-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org, rni-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org, containers-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org, linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, cgroups-u79uwXL29TY76Z2rM5mHXA@public.gmane.org On Tue, Dec 18, 2012 at 11:21:55AM -0800, Tejun Heo wrote: > On Tue, Dec 18, 2012 at 02:18:54PM -0500, Vivek Goyal wrote: > > > Yeah, hmmm, maybe we should add another set of counters to carry stats > > > from dead ones? Avoiding hierarchical accounting overhead in hot path > > > while remembering by-gones shouldn't be that hard. Will work on that. > > > > So are you planning to migrate the stats to parent when a cgroup is being > > deleted? That should make sure we don't do hierarhical update. > > Yeah, something like that. The current statistics being !hierarchical > means that we need to keep them separate from the actual stats, so > it'll end up with another set of stats. Kinda nasty but I don't see > any better way out. Separate stats for hierarchical migration make sense to me. Thanks Vivek