From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752268AbbJTQxi (ORCPT ); Tue, 20 Oct 2015 12:53:38 -0400 Received: from bombadil.infradead.org ([198.137.202.9]:35989 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751042AbbJTQxh (ORCPT ); Tue, 20 Oct 2015 12:53:37 -0400 Date: Tue, 20 Oct 2015 18:53:31 +0200 From: Peter Zijlstra To: Steven Rostedt Cc: "Paul E. McKenney" , LKML , Rusty Russell Subject: Re: [PATCH] module: Prevent recursion bug caused by module RCU check Message-ID: <20151020165331.GU11639@twins.programming.kicks-ass.net> References: <20151020122103.66ab250a@gandalf.local.home> <20151020163952.GI17308@twins.programming.kicks-ass.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20151020163952.GI17308@twins.programming.kicks-ass.net> User-Agent: Mutt/1.5.21 (2012-12-30) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, Oct 20, 2015 at 06:39:52PM +0200, Peter Zijlstra wrote: > On Tue, Oct 20, 2015 at 12:21:03PM -0400, Steven Rostedt wrote: > > +++ b/kernel/module.c > > @@ -284,11 +284,25 @@ static void module_assert_mutex(void) > > static void module_assert_mutex_or_preempt(void) > > { > > #ifdef CONFIG_LOCKDEP > > + static int once; > > + > > if (unlikely(!debug_locks)) > > return; > > > > - WARN_ON(!rcu_read_lock_sched_held() && > > - !lockdep_is_held(&module_mutex)); > > + /* > > + * Would be nice to use WARN_ON_ONCE(), but the warning > > + * that causes a stack trace may call __module_address() > > + * which may call here, and we trigger the warning again, > > + * before the WARN_ON_ONCE() updates its flag. > > + * To prevent the recursion, we need to open code the > > + * once logic. > > + */ > > + if (!once && > > + unlikely(!rcu_read_lock_sched_held() && > > + !lockdep_is_held(&module_mutex))) { > > + once++; > > once = 1; > > is more 'once' :-) Otherwise its once every 4-odd billion. > > > + WARN_ON(1); > > + } > > #endif Also, would it not be better to fix WARN_ON_ONCE() instead? --- include/asm-generic/bug.h | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) diff --git a/include/asm-generic/bug.h b/include/asm-generic/bug.h index 630dd2372238..c41e698613f3 100644 --- a/include/asm-generic/bug.h +++ b/include/asm-generic/bug.h @@ -110,9 +110,11 @@ extern void warn_slowpath_null(const char *file, const int line); static bool __section(.data.unlikely) __warned; \ int __ret_warn_once = !!(condition); \ \ - if (unlikely(__ret_warn_once)) \ - if (WARN_ON(!__warned)) \ + if (unlikely(__ret_warn_once)) { \ + if (!__warned) \ __warned = true; \ + WARN_ON(1); \ + } \ unlikely(__ret_warn_once); \ }) @@ -120,9 +122,11 @@ extern void warn_slowpath_null(const char *file, const int line); static bool __section(.data.unlikely) __warned; \ int __ret_warn_once = !!(condition); \ \ - if (unlikely(__ret_warn_once)) \ - if (WARN(!__warned, format)) \ + if (unlikely(__ret_warn_once)) { \ + if (!__warned) \ __warned = true; \ + WARN(1, format); \ + } \ unlikely(__ret_warn_once); \ })