From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751941Ab1GZMQb (ORCPT ); Tue, 26 Jul 2011 08:16:31 -0400 Received: from mail-ew0-f46.google.com ([209.85.215.46]:44078 "EHLO mail-ew0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751832Ab1GZMQZ (ORCPT ); Tue, 26 Jul 2011 08:16:25 -0400 Date: Tue, 26 Jul 2011 15:15:36 +0300 From: Sergey Senozhatsky To: Peter Zijlstra Cc: Arnaud Lacombe , Ingo Molnar , Linus Torvalds , linux-kernel@vger.kernel.org, Thomas Gleixner , Andrew Morton , DaveJones Subject: Re: [GIT PULL] core/printk changes for v3.1 Message-ID: <20110726121536.GA3469@swordfish.minsk.epam.com> References: <20110722125934.GA25433@elte.hu> <1311679697.24752.28.camel@twins> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1311679697.24752.28.camel@twins> User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On (07/26/11 13:28), Peter Zijlstra wrote: > > WARNING: at /src/linux/linux/kernel/lockdep.c:2529 > > Not actually having reproduced the problem, does the below cure things? > Able to reproduce bad_area()-cases (http://lkml.org/lkml/2011/7/25/144). Previously: kernel: [ 3634.499710] Call Trace: kernel: [ 3634.499718] [] warn_slowpath_common+0x7e/0x96 kernel: [ 3634.499726] [] ? __bad_area_nosemaphore+0x35/0x1b9 kernel: [ 3634.499730] [] warn_slowpath_null+0x15/0x17 kernel: [ 3634.499735] [] trace_hardirqs_on_caller+0x6d/0x166 kernel: [ 3634.499741] [] trace_hardirqs_on+0xd/0xf kernel: [ 3634.499746] [] __bad_area_nosemaphore+0x35/0x1b9 kernel: [ 3634.499751] [] bad_area+0x45/0x4c kernel: [ 3634.499759] [] do_page_fault+0x2da/0x48a kernel: [ 3634.499766] [] ? do_brk+0x24e/0x2a2 kernel: [ 3634.499774] [] ? trace_hardirqs_off_thunk+0x3a/0x3c kernel: [ 3634.499780] [] page_fault+0x25/0x30 kernel: [ 3634.499784] ---[ end trace 642edd2b5d1b3694 ]--- kernel: [ 3634.499791] conftest[21907]: segfault at 0 ip 00007f591a6e7144 sp 00007fff03423550 error 4 in libc-2.14.so[7f591a62a000+157000] which is not true for the patched kernel. That said, that at least bad_area() is cured for now. Here is mine: Tested-by: Sergey Senozhatsky Best, Sergey > --- > Subject: lockdep: Fix trace_hardirqs_on_caller() > From: Peter Zijlstra > Date: Tue Jul 26 13:13:44 CEST 2011 > > Commit dd4e5d3ac4a ("lockdep: Fix trace_[soft,hard]irqs_[on,off]() > recursion") made a bit of a mess of the various checks and error > conditions. > > In particular it moved the check for !irqs_disabled() before the > spurious enable test, resulting in some warnings. > > Reported-by: Arnaud Lacombe > Reported-by: Dave Jones > Reported-by: Sergey Senozhatsky > Signed-off-by: Peter Zijlstra > --- > kernel/lockdep.c | 30 ++++++++++++++++-------------- > 1 file changed, 16 insertions(+), 14 deletions(-) > Index: linux-2.6/kernel/lockdep.c > =================================================================== > --- linux-2.6.orig/kernel/lockdep.c > +++ linux-2.6/kernel/lockdep.c > @@ -2485,23 +2485,9 @@ static void __trace_hardirqs_on_caller(u > { > struct task_struct *curr = current; > > - if (DEBUG_LOCKS_WARN_ON(unlikely(early_boot_irqs_disabled))) > - return; > - > - if (unlikely(curr->hardirqs_enabled)) { > - /* > - * Neither irq nor preemption are disabled here > - * so this is racy by nature but losing one hit > - * in a stat is not a big deal. > - */ > - __debug_atomic_inc(redundant_hardirqs_on); > - return; > - } > /* we'll do an OFF -> ON transition: */ > curr->hardirqs_enabled = 1; > > - if (DEBUG_LOCKS_WARN_ON(current->hardirq_context)) > - return; > /* > * We are going to turn hardirqs on, so set the > * usage bit for all held locks: > @@ -2529,9 +2515,25 @@ void trace_hardirqs_on_caller(unsigned l > if (unlikely(!debug_locks || current->lockdep_recursion)) > return; > > + if (unlikely(current->hardirqs_enabled)) { > + /* > + * Neither irq nor preemption are disabled here > + * so this is racy by nature but losing one hit > + * in a stat is not a big deal. > + */ > + __debug_atomic_inc(redundant_hardirqs_on); > + return; > + } > + > if (DEBUG_LOCKS_WARN_ON(!irqs_disabled())) > return; > > + if (DEBUG_LOCKS_WARN_ON(unlikely(early_boot_irqs_disabled))) > + return; > + > + if (DEBUG_LOCKS_WARN_ON(current->hardirq_context)) > + return; > + > current->lockdep_recursion = 1; > __trace_hardirqs_on_caller(ip); > current->lockdep_recursion = 0; >