From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932074Ab2AECSG (ORCPT ); Wed, 4 Jan 2012 21:18:06 -0500 Received: from e32.co.us.ibm.com ([32.97.110.150]:44593 "EHLO e32.co.us.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757099Ab2AECSD (ORCPT ); Wed, 4 Jan 2012 21:18:03 -0500 Date: Wed, 4 Jan 2012 18:17:58 -0800 From: "Paul E. McKenney" To: Frederic Weisbecker Cc: Sasha Levin , linux-kernel Subject: Re: [PATCH] rcu: Improve detection of illegal synchronize_rcu() call from RCU read side Message-ID: <20120105021758.GS2448@linux.vnet.ibm.com> Reply-To: paulmck@linux.vnet.ibm.com References: <1324901803.31721.4.camel@lappy> <20111226163148.GC2435@linux.vnet.ibm.com> <20111226163734.GF28309@somewhere.redhat.com> <20111226195656.GD2435@linux.vnet.ibm.com> <20120104190336.GC1143@somewhere> <20120104213035.GF2448@linux.vnet.ibm.com> <20120105014518.GD1143@somewhere> <20120105020108.GQ2448@linux.vnet.ibm.com> <20120105020601.GE1143@somewhere> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20120105020601.GE1143@somewhere> User-Agent: Mutt/1.5.21 (2010-09-15) x-cbid: 12010502-3270-0000-0000-000002FF2031 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Thu, Jan 05, 2012 at 03:06:03AM +0100, Frederic Weisbecker wrote: > On Wed, Jan 04, 2012 at 06:01:08PM -0800, Paul E. McKenney wrote: > > On Thu, Jan 05, 2012 at 02:45:20AM +0100, Frederic Weisbecker wrote: > > > On Wed, Jan 04, 2012 at 01:30:35PM -0800, Paul E. McKenney wrote: > > > > On Wed, Jan 04, 2012 at 08:03:39PM +0100, Frederic Weisbecker wrote: > > > > > Actually for the case of RCU, the wait_for_completion() called by synchronize_rcu() [ . . . ] > > > > > rcutiny seems to be fine with the cond_resched() call, but srcu needs > > > > > a special treatment. > > > > > > > > For the moment, I just applied rcu_lockdep_assert() everywhere -- zero > > > > cost on non-lockdep kernels, and fully handles all of the RCU simple > > > > self-deadlock cases. > > > > > > So, for RCU I'm not sure this is useful given the might_sleep() things. > > > But for srcu it is. > > > > One nice thing about the lockdep approach is that it tracks where the > > conflicting RCU read-side critical section started. But I am planning > > for these to be 3.4 material, so we do have some time to refine them. > > Yeah sure. And in any case it's still good to keep might_sleep() early > to spot other sources of illegal atomic sections (irqs disabled and co) Agreed! Thanx, Paul