Return-path: Received: from mail.w1.fi ([212.71.239.96]:39069 "EHLO li674-96.members.linode.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752954AbbCaO7O (ORCPT ); Tue, 31 Mar 2015 10:59:14 -0400 Date: Tue, 31 Mar 2015 17:59:10 +0300 From: Jouni Malinen To: Avinash Patil Cc: Johannes Berg , "linux-wireless@vger.kernel.org" , Amitkumar Karwar , Cathy Luo , Xinming Hu , Li Long Subject: Re: [PATCH 1/3] mwifiex: add cfg80211 set_default_mgmt_key handler Message-ID: <20150331145910.GA8244@w1.fi> (sfid-20150331_165918_698232_3BA79BB5) References: <1423764954-30412-1-git-send-email-patila@marvell.com> <1423850815.2936.4.camel@sipsolutions.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: Sender: linux-wireless-owner@vger.kernel.org List-ID: On Tue, Mar 31, 2015 at 07:33:39AM -0700, Avinash Patil wrote: > We feel set_default_key_mgmt handler is required for devices which supports 2 pair of GTKs. Why would a device not support multiple GTKs/IGTKs? It does not sound possible to implement RSN correctly without such support.. AP side could kind of try to work with only a single GTK/IGTK, but non-AP STA would not handle GTK/IGTK rekeying. In any case, IEEE Std 802.11-2012 seems to be pretty clear on the assumption being that the device (including AP-only cases) supports multiple GTK/IGTK. The authenticator state machines use key index values 1 and 2 and swap between these when doing rekeying. -- Jouni Malinen PGP id EFC895FA