From mboxrd@z Thu Jan 1 00:00:00 1970 From: Nikolay Aleksandrov Subject: Re: [PATCH net-next 0/2] fix bonding neighbour setup handling Date: Mon, 05 Aug 2013 15:49:08 +0200 Message-ID: <51FFAD54.7090501@redhat.com> References: <1375463259-12033-1-git-send-email-vfalico@redhat.com> <20130802.154539.237539535236462726.davem@davemloft.net> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: vfalico@redhat.com, netdev@vger.kernel.org, fubar@us.ibm.com, andy@greyhouse.net, ebiederm@xmission.com, joe@perches.com To: David Miller Return-path: Received: from mx1.redhat.com ([209.132.183.28]:57527 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753255Ab3HENwr (ORCPT ); Mon, 5 Aug 2013 09:52:47 -0400 In-Reply-To: <20130802.154539.237539535236462726.davem@davemloft.net> Sender: netdev-owner@vger.kernel.org List-ID: On 08/03/2013 12:45 AM, David Miller wrote: > > Thanks for the detailed explanation of the situation, it made your patches > trivial to review. > > Applied, thanks. Hi all, Vaeceslav thanks for fixing this. Since the cat is out of the bag about this bug, as Vaeceslav discovered it independently and wasn't aware that there's a CVE number pending because it poses a security threat since the dereferenced first_slave pointer is taken from the struct vlan_dev_priv's ingress_priority map array which is user-controllable and any memory address can be dereferenced in that way, and taking after that first_slave->dev->netdev_ops and calling a function from the ops is making it even easier. Of course for that to happen the user must have CAP_NET_ADMIN. I've tested these patches and they apply cleanly on -net as well, so please queue them for -net and stable. Also apologies for the late reply but it wasn't up to me when to reveal this. Thank you, Nik