From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S965162AbWEaV1O (ORCPT ); Wed, 31 May 2006 17:27:14 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S965169AbWEaV1N (ORCPT ); Wed, 31 May 2006 17:27:13 -0400 Received: from mx3.mail.elte.hu ([157.181.1.138]:58593 "EHLO mx3.mail.elte.hu") by vger.kernel.org with ESMTP id S965163AbWEaV1L (ORCPT ); Wed, 31 May 2006 17:27:11 -0400 Date: Wed, 31 May 2006 23:27:30 +0200 From: Ingo Molnar To: Arjan van de Ven Cc: alan@redhat.com, Andrew Morton , linux-kernel@vger.kernel.org Subject: Re: [patch, -rc5-mm1] locking validator: special rule: 8390.c disable_irq() Message-ID: <20060531212730.GA3174@elte.hu> References: <20060531200236.GA31619@elte.hu> <1149107500.3114.75.camel@laptopd505.fenrus.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1149107500.3114.75.camel@laptopd505.fenrus.org> User-Agent: Mutt/1.4.2.1i X-ELTE-SpamScore: 0.0 X-ELTE-SpamLevel: X-ELTE-SpamCheck: no X-ELTE-SpamVersion: ELTE 2.0 X-ELTE-SpamCheck-Details: score=0.0 required=5.9 tests=AWL,BAYES_50 autolearn=no SpamAssassin version=3.0.3 0.0 BAYES_50 BODY: Bayesian spam probability is 40 to 60% [score: 0.5001] 0.0 AWL AWL: From: address is in the auto white-list X-ELTE-VirusStatus: clean Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org * Arjan van de Ven wrote: > On Wed, 2006-05-31 at 22:02 +0200, Ingo Molnar wrote: > > untested on 8390 hardware, but ought to solve the lockdep false > > positive. > > > > ----------------- > > Subject: locking validator: special rule: 8390.c disable_irq() > > From: Ingo Molnar > > > > 8390.c knows that ei_local->page_lock can only be used by an irq > > context that it disabled - > > btw I think this is no longer correct with the irq polling stuff Alan > added to the kernel recently... hm, indeed. misrouted_irq() goes through all irq descriptors and ignores IRQ_DISABLED flag - rendering disable_irq() useless in essence, and introducing the kind of deadlocks that lockdep warned about. Andrew, as far as i can see with irqfixup this isnt a lockdep false positive but a real deadlock scenario - a spurious IRQ might arrive during vortex_timer() execution and might cause the execution of misrouted_irq(), which could execute vortex_interrupt() => deadlock. Alan, is this a necessary property of irqpoll/irqfixup? Shouldnt irqfixup leave irq descriptors alone that are disabled? Ingo