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 S1753022Ab3C0Imj (ORCPT <rfc822;w@1wt.eu>);
	Wed, 27 Mar 2013 04:42:39 -0400
Received: from merlin.infradead.org ([205.233.59.134]:43957 "EHLO
	merlin.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1751777Ab3C0Imh (ORCPT
	<rfc822;linux-kernel@vger.kernel.org>);
	Wed, 27 Mar 2013 04:42:37 -0400
Message-ID: <1364373750.5053.54.camel@laptop>
Subject: Re: [PATCH -mm -next] ipc,sem: fix lockdep false positive
From: Peter Zijlstra <peterz@infradead.org>
To: Rik van Riel <riel@surriel.com>
Cc: Michel Lespinasse <walken@google.com>,
        Sasha Levin <sasha.levin@oracle.com>, torvalds@linux-foundation.org,
        davidlohr.bueso@hp.com, linux-kernel@vger.kernel.org,
        akpm@linux-foundation.org, hhuang@redhat.com, jason.low2@hp.com,
        lwoodman@redhat.com, chegu_vinod@hp.com, Dave Jones <davej@redhat.com>,
        benisty.e@gmail.com, Ingo Molnar <mingo@redhat.com>
Date: Wed, 27 Mar 2013 09:42:30 +0100
In-Reply-To: <5151BC78.3030306@surriel.com>
References: <1363809337-29718-1-git-send-email-riel@surriel.com>
	  <5150B1C2.8090607@oracle.com>
	  <20130325163844.042a45ba@annuminas.surriel.com>
	  <CANN689FdtXs0-VKxwWHQKgnnSvcvu5hHDMmGOe0H+bdHA4BhXg@mail.gmail.com>
	  <1364303965.5053.29.camel@laptop>
	  <CANN689HCbuVOc73BpVubYy2AjA2mcz_iyTg6Fo2Zz-NGnQ93TA@mail.gmail.com>
	 <1364308023.5053.40.camel@laptop> <5151BC78.3030306@surriel.com>
Content-Type: text/plain; charset="UTF-8"
X-Mailer: Evolution 3.6.2-0ubuntu0.1 
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On Tue, 2013-03-26 at 11:19 -0400, Rik van Riel wrote:
> > Maybe something like:
> >
> > void sma_lock(struct sem_array *sma) /* global */
> > {
> >       int i;
> >
> >       sma->global_locked = 1;
> >       smp_wmb(); /* can we merge with the LOCK ? */
> >       spin_lock(&sma->global_lock);
> >
> >       /* wait for all local locks to go away */
> >       for (i = 0; i < sma->sem_nsems; i++)
> >               spin_unlock_wait(&sem->sem_base[i]->lock);      
> > }
> >
> > void sma_lock_one(struct sem_array *sma, int nr) /* local */
> > {
> >       smp_rmb(); /* pairs with wmb in sma_lock() */
> >       if (unlikely(sma->global_locked)) { /* wait for global lock */
> >               while (sma->global_locked)
> >                       spin_unlock_wait(&sma->global_lock);
> >       }
> >       spin_lock(&sma->sem_base[nr]->lock);
> > }

I since realized there's an ordering problem with ->global_locked, we
need to use spin_is_locked() or somesuch.

Two competing sma_lock() operations will screw over the separate
variable.

> 
> > This still has the problem of a non-preemptible section of
> O(sem_nsems)
> > (with the avg wait-time on the local lock). Could we make the global
> > lock a sleeping lock?
> 
> Not without breaking your scheme above :)

How would making sma->global_lock a mutex wreck anything?