From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S933271Ab1INUGe (ORCPT ); Wed, 14 Sep 2011 16:06:34 -0400 Received: from mailhub.sw.ru ([195.214.232.25]:19197 "EHLO relay.sw.ru" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S933151Ab1INUGc (ORCPT ); Wed, 14 Sep 2011 16:06:32 -0400 From: Glauber Costa To: linux-kernel@vger.kernel.org Cc: xemul@parallels.com, paul@paulmenage.org, lizf@cn.fujitsu.com, daniel.lezcano@free.fr, mingo@elte.hu, a.p.zijlstra@chello.nl, jbottomley@parallels.com, Glauber Costa Subject: [PATCH 9/9] Report steal time for cgroup Date: Wed, 14 Sep 2011 17:04:55 -0300 Message-Id: <1316030695-19826-10-git-send-email-glommer@parallels.com> X-Mailer: git-send-email 1.7.6 In-Reply-To: <1316030695-19826-1-git-send-email-glommer@parallels.com> References: <1316030695-19826-1-git-send-email-glommer@parallels.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org This patch introduces a functionality commonly found in hypervisors: steal time. For those not particularly familiar with it, steal time is defined as any time in which a virtual machine (or container) wanted to perform cpu work, but could not due to another VM/container being scheduled in its place. Note that idle time is never defined as steal time. Assuming each container will live in its cgroup, we can very easily and nicely calculate steal time as all user/system time recorded in our sibling cgroups. Signed-off-by: Glauber Costa --- kernel/sched.c | 30 ++++++++++++++++++++++++++++++ 1 files changed, 30 insertions(+), 0 deletions(-) diff --git a/kernel/sched.c b/kernel/sched.c index 8f254d0..2dad1e0 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -9321,6 +9321,7 @@ static int cpuacct_proc_stat(struct cgroup *cgrp, struct cftype *cft, struct cpuacct *ca = cgroup_ca(cgrp); struct timespec ts; u64 *cpustat; + struct cgroup *sib; user = nice = system = idle = iowait = irq = softirq = steal = cputime64_zero; @@ -9343,6 +9344,20 @@ static int cpuacct_proc_stat(struct cgroup *cgrp, struct cftype *cft, irq = cputime64_add(irq, cpustat[CPUACCT_STAT_IRQ]); softirq = cputime64_add(softirq, cpustat[CPUACCT_STAT_SOFTIRQ]); steal = cputime64_add(steal, kstat_cpu(i).cpustat.steal); + rcu_read_lock(); + list_for_each_entry(sib, &ca->css.cgroup->sibling, sibling) { + u64 *cpustat_sib; + struct cpuacct *ca_sib = cgroup_ca(sib); + if (!ca_sib) + continue; + + cpustat_sib = per_cpu_ptr(ca_sib->cpustat, i); + steal = cputime64_add(steal, cpustat_sib[CPUACCT_STAT_USER]); + steal = cputime64_add(steal, cpustat_sib[CPUACCT_STAT_SYSTEM]); + steal = cputime64_add(steal, cpustat_sib[CPUACCT_STAT_IRQ]); + steal = cputime64_add(steal, cpustat_sib[CPUACCT_STAT_SOFTIRQ]); + } + rcu_read_unlock(); guest = cputime64_add(guest, cpustat[CPUACCT_STAT_GUEST]); guest_nice = cputime64_add(guest_nice, cpustat[CPUACCT_STAT_GUEST_NICE]); @@ -9384,6 +9399,21 @@ static int cpuacct_proc_stat(struct cgroup *cgrp, struct cftype *cft, irq = cpustat[CPUACCT_STAT_IRQ]; softirq = cpustat[CPUACCT_STAT_SOFTIRQ]; steal = kstat_cpu(i).cpustat.steal; + rcu_read_lock(); + list_for_each_entry(sib, &ca->css.cgroup->sibling, sibling) { + u64 *cpustat_sib; + struct cpuacct *ca_sib = cgroup_ca(sib); + + if (!ca_sib) + continue; + cpustat_sib = per_cpu_ptr(ca_sib->cpustat, i); + steal = cputime64_add(steal, cpustat_sib[CPUACCT_STAT_USER]); + steal = cputime64_add(steal, cpustat_sib[CPUACCT_STAT_SYSTEM]); + steal = cputime64_add(steal, cpustat_sib[CPUACCT_STAT_IRQ]); + steal = cputime64_add(steal, cpustat_sib[CPUACCT_STAT_SOFTIRQ]); + } + rcu_read_unlock(); + guest = cpustat[CPUACCT_STAT_GUEST]; guest_nice = cpustat[CPUACCT_STAT_GUEST_NICE]; seq_printf(p, -- 1.7.6