Return-path: Received: from mail-lf0-f52.google.com ([209.85.215.52]:38662 "EHLO mail-lf0-f52.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752318AbdH1IjL (ORCPT ); Mon, 28 Aug 2017 04:39:11 -0400 Date: Mon, 28 Aug 2017 10:39:12 +0200 From: Johan Hovold To: Samuel Ortiz , "David S. Miller" Cc: linux-wireless@vger.kernel.org, netdev@vger.kernel.org, Dan Carpenter , Johan Hovold , stable , Greg Kroah-Hartman , linux-kernel@vger.kernel.org, Andrew Morton , Ben Hutchings Subject: Re: [PATCH] NFC: fix device-allocation error return Message-ID: <20170828083912.GB20634@localhost> (sfid-20170828_103931_657940_251A8F5E) References: <20170709110858.20331-1-johan@kernel.org> <20170722133228.GE2729@localhost> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <20170722133228.GE2729@localhost> Sender: linux-wireless-owner@vger.kernel.org List-ID: Samuel or David, On Sat, Jul 22, 2017 at 03:32:28PM +0200, Johan Hovold wrote: > On Sun, Jul 09, 2017 at 01:08:58PM +0200, Johan Hovold wrote: > > A recent change fixing NFC device allocation itself introduced an > > error-handling bug by returning an error pointer in case device-id > > allocation failed. This is clearly broken as the callers still expected > > NULL to be returned on errors as detected by Dan's static checker. > > > > Fix this up by returning NULL in the event that we've run out of memory > > when allocating a new device id. > > > > Note that the offending commit is marked for stable (3.8) so this fix > > needs to be backported along with it. > > > > Fixes: 20777bc57c34 ("NFC: fix broken device allocation") > > Cc: stable # 3.8 > > Reported-by: Dan Carpenter > > Signed-off-by: Johan Hovold > Could you apply this follow-up fix so that it can be backported along > with the offending commit (which was just added to the stable queues)? > > We would only hit this error path if an ida allocation fails due to OOM; > so while this is not critical, it would still be nice to get it fixed. Another reminder about this one; can you apply it so we can get it into 4.14-rc1? Note that the offending commit has now been backported to the stable trees and we really want this trivial follow-up fix to be backported as well. Let me know if you want me to resend the patch. Thanks, Johan