From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753914Ab1KDJ1H (ORCPT ); Fri, 4 Nov 2011 05:27:07 -0400 Received: from mail-gx0-f174.google.com ([209.85.161.174]:51987 "EHLO mail-gx0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751262Ab1KDJ1F (ORCPT ); Fri, 4 Nov 2011 05:27:05 -0400 From: Yong Zhang To: linux-kernel@vger.kernel.org Cc: sergey.senozhatsky@gmail.com, bp@alien8.de Subject: [PATCH 0/4] patches to cure race in lock_set_class() Date: Fri, 4 Nov 2011 17:26:26 +0800 Message-Id: <1320398790-21663-1-git-send-email-yong.zhang0@gmail.com> X-Mailer: git-send-email 1.7.5.4 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org This a new version of my previous post[1]. patch#1 is the one which fix the reported bug. So Sergey, Borislav, would you please give it a test? patch#2~4 is an enhancement which fix a potential race in lock_set_class(). But the fact is the only current user in kernel is lock_set_subclass(), we still have no user which change its 'key' in flying. So I mark it with RFC. More detail in each patch :) Thanks, Yong [1]:http://marc.info/?l=linux-kernel&m=131919035525533 Yong Zhang (4): lockdep: lock_set_subclass() fix lockdep: Let register_lock_class() can be called with/without graph_lock lockdep: split lockdep_init_map() lockdep: fix race condition in __lock_set_class() include/linux/lockdep.h | 2 +- kernel/lockdep.c | 50 ++++++++++++++++++++++++++++++++-------------- 2 files changed, 36 insertions(+), 16 deletions(-) -- 1.7.5.4