From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id S932584Ab0DGN5f (ORCPT <rfc822;w@1wt.eu>);
	Wed, 7 Apr 2010 09:57:35 -0400
Received: from mx1.redhat.com ([209.132.183.28]:12844 "EHLO mx1.redhat.com"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1752556Ab0DGN5e (ORCPT <rfc822;linux-kernel@vger.kernel.org>);
	Wed, 7 Apr 2010 09:57:34 -0400
Date: Wed, 7 Apr 2010 15:54:56 +0200
From: Oleg Nesterov <oleg@redhat.com>
To: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Gautham R Shenoy <ego@in.ibm.com>, Rusty Russell <rusty@rustcorp.com.au>,
       Benjamin Herrenschmidt <benh@kernel.crashing.org>,
       Hugh Dickins <hugh.dickins@tiscali.co.uk>, Ingo Molnar <mingo@elte.hu>,
       "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>,
       Nathan Fontenot <nfont@austin.ibm.com>,
       Peter Zijlstra <peterz@infradead.org>,
       Andrew Morton <akpm@linux-foundation.org>,
       Thomas Gleixner <tglx@linutronix.de>, Sachin Sant <sachinp@in.ibm.com>,
       "H. Peter Anvin" <hpa@zytor.com>, Shane Wang <shane.wang@intel.com>,
       Roland McGrath <roland@redhat.com>, linux-kernel@vger.kernel.org
Subject: Re: [PATCH 2/2] cpuhotplug: make get_online_cpus() scalability by
	using percpu counter
Message-ID: <20100407135456.GA12029@redhat.com>
References: <4BB9BD8A.9040209@cn.fujitsu.com> <20100405162901.GA3567@redhat.com> <20100406120039.GC5680@redhat.com> <4BBC8A11.3040501@cn.fujitsu.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <4BBC8A11.3040501@cn.fujitsu.com>
User-Agent: Mutt/1.5.18 (2008-05-17)
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On 04/07, Lai Jiangshan wrote:
>
> Oleg Nesterov wrote:
> > On 04/05, Oleg Nesterov wrote:
> >> On 04/05, Lai Jiangshan wrote:
> >>> 1) get_online_cpus() must be allowed to be called recursively, so I added
> >>>    get_online_cpus_nest for every task for new code.
> >> Well, iirc one of the goals of
> >>
> >> 	cpu-hotplug: replace lock_cpu_hotplug() with get_online_cpus()
> >> 	86ef5c9a8edd78e6bf92879f32329d89b2d55b5a
> >>
> >> was avoiding the new members in task_struct. I leave this up to you
> >> and Gautham.
>
> Old get_online_cpus() is read-preference, I think the goal of this ability
> is allow get_online_cpus()/put_online_cpus() to be called nested.

Sure, I understand why you added task_struct->get_online_cpus_nest.

> and use per-task counter for allowing get_online_cpus()/put_online_cpus()
> to be called nested, I think this deal is absolutely worth.

As I said, I am not going to argue. I can't justify this tradeoff.

> >>>  void put_online_cpus(void)
> >>>  {
> >>> ...
> >>> +	if (!--current->get_online_cpus_nest) {
> >>> +		preempt_disable();
> >>> +		__get_cpu_var(refcount)--;
> >>> +		if (cpu_hotplug_task)
> >>> +			wake_up_process(cpu_hotplug_task);
> >> This looks unsafe. In theory nothing protects cpu_hotplug_task from
> >> exiting if refcount_sum() becomes zero, this means wake_up_process()
> >> can hit the freed/reused/unmapped task_struct. Probably cpu_hotplug_done()
> >> needs another synhronize_sched() before return.
> >
> > Yes, I think this is true, at least in theory.
>
> preempt_disable() prevent cpu_hotplug_task from exiting.

If the cpu_down() is the caller of cpu_hotplug_begin/done, then yes.

But unless I missed something, nothing protects from cpu_up() which
takes this lock too.

Just in case... I am not saying this is really possible in practice.

Oleg.