From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756056Ab3KVRyh (ORCPT ); Fri, 22 Nov 2013 12:54:37 -0500 Received: from qmta10.emeryville.ca.mail.comcast.net ([76.96.30.17]:49405 "EHLO qmta10.emeryville.ca.mail.comcast.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755517Ab3KVRye (ORCPT ); Fri, 22 Nov 2013 12:54:34 -0500 From: Shuah Khan To: anton@enomsg.org, dwmw2@infradead.org, rostedt@goodmis.org, fweisbec@gmail.com, mingo@redhat.com, rjw@rjwysocki.net, gregkh@linuxfoundation.org Cc: Shuah Khan , linux-kernel@vger.kernel.org, shuahkhan@gmail.com Subject: [PATCH v3 0/2] power_supply: Fix Oops from NULL pointer dereference from wakeup_source_activate Date: Fri, 22 Nov 2013 10:54:27 -0700 Message-Id: X-Mailer: git-send-email 1.8.3.2 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org power_supply_register() calls device_init_wakeup() to register a wakeup source before initializing dev_name. As a result, device_wakeup_enable() end up registering wakeup source with a null name when wakeup_source_register() gets called with dev_name(dev) which is null at the time. device_wakeup_enable() uses dev_name(dev) as the wakeup source name. When it gets called with a device with its name not yet set, ws structure with ws->name = NULL gets created. When kernel is booted with wakeup_source_activate enabled, it will panic when the trace point code tries to dereference ws->name. Registering a a wakeup source without a name should be possible. This patch series fixes power_supply_register() to initialize the device name prior to calling device_init_wakeup() and fixes tracing infrastructure to be more robust in handling null strings in __assign_string() and __string(). With this change null string is handled gracefully and replacing it with "(null)" when trace is generated. These two patches are not dependent, I left them as a series since the original discussion started the fixes grouped in a series. power_supply patch - no changes since series patch v1, other than including Acked-by from Greg and Anton. Also added stable tag. Second patch now is the fix to tracepoint infrastructure routines __assign_string() and __string(). This will address the problem at the tracing infrastructure level which is better than fixing individual tracepoint code. These patches can be applied independently with no ill effects. I included the trace for the wakeup_source_activate trace output in the change logs for both patches. Shuah Khan (2): power_supply: Fix Oops from NULL pointer dereference from wakeup_source_activate tracing: Fix Oops from NULL pointer dereference from __assign_str drivers/power/power_supply_core.c | 12 ++++++------ include/trace/ftrace.h | 7 +++++-- 2 files changed, 11 insertions(+), 8 deletions(-) -- 1.8.3.2