From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755398AbYG1W47 (ORCPT ); Mon, 28 Jul 2008 18:56:59 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752404AbYG1W4v (ORCPT ); Mon, 28 Jul 2008 18:56:51 -0400 Received: from 74-93-104-97-Washington.hfc.comcastbusiness.net ([74.93.104.97]:53830 "EHLO sunset.davemloft.net" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1751610AbYG1W4v (ORCPT ); Mon, 28 Jul 2008 18:56:51 -0400 Date: Mon, 28 Jul 2008 15:56:51 -0700 (PDT) Message-Id: <20080728.155651.71572255.davem@davemloft.net> To: jeremy@goop.org Cc: mingo@elte.hu, peterz@infradead.org, linux-kernel@vger.kernel.org, rusty@rustcorp.com.au Subject: Re: combinatorial explosion in lockdep From: David Miller In-Reply-To: <488E4D1C.8000906@goop.org> References: <20080728.153702.103032029.davem@davemloft.net> <488E4D1C.8000906@goop.org> X-Mailer: Mew version 5.2 on Emacs 22.1 / Mule 5.0 (SAKAKI) Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Jeremy Fitzhardinge Date: Mon, 28 Jul 2008 15:50:04 -0700 > David Miller wrote: > > The triggering event seems to be kstopmachine as done by the first > > module load. This seems to create the situation where we do double RQ > > locking on essentially all the run queues, in order to move the > > kstopmachine threads around to their proprer cpus. I think this is > > what starts to make the lockdep dependency chains huge. > > > > Does Rusty's stop_machine rewrite which puts the threads on the correct > cpus to start with at least mitigate the symptoms? I could check that, but it won't keep this from happening and being triggerable by a trivial userland program setting scheduler affinities.