From mboxrd@z Thu Jan  1 00:00:00 1970
Return-path: <linux-wireless-owner@vger.kernel.org>
Received: from he.sipsolutions.net ([78.46.109.217]:49192 "EHLO
	sipsolutions.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1753528Ab2LLQUm (ORCPT
	<rfc822;linux-wireless@vger.kernel.org>);
	Wed, 12 Dec 2012 11:20:42 -0500
Message-ID: <1355329248.9708.16.camel@jlt4.sipsolutions.net> (sfid-20121212_172045_095829_9396B373)
Subject: Re: [PATCHv2 00/21] cw1200: ST-Ericsson cw1200 wlan device driver
From: Johannes Berg <johannes@sipsolutions.net>
To: Solomon Peachy <pizza@shaftnet.org>
Cc: Christian Lamparter <chunkeey@googlemail.com>,
	"John W. Linville" <linville@tuxdriver.com>,
	Steev Klimaszewski <steev@genesi-usa.com>,
	Janusz Dziedzic <janusz.dziedzic@gmail.com>,
	Dmitry Tarnyagin <abi.dmitryt@gmail.com>,
	linux-wireless@vger.kernel.org
Date: Wed, 12 Dec 2012 17:20:48 +0100
In-Reply-To: <20121212161845.GA21461@shaftnet.org>
References: <1330720003-15866-1-git-send-email-dmitry.tarnyagin@stericsson.com>
	 <20120706201839.GE1879@tuxdriver.com> <20121211232311.GA9080@shaftnet.org>
	 <201212121613.12750.chunkeey@googlemail.com>
	 <20121212161845.GA21461@shaftnet.org>
Content-Type: text/plain; charset="UTF-8"
Mime-Version: 1.0
Sender: linux-wireless-owner@vger.kernel.org
List-ID: <linux-wireless.vger.kernel.org>

On Wed, 2012-12-12 at 11:18 -0500, Solomon Peachy wrote:
> On Wed, Dec 12, 2012 at 04:13:12PM +0100, Christian Lamparter wrote:
> > Mac80211 already reserves at least 18 bytes tailroom [which can be
> > allocated by a simple skb_put]. In fact p54 uses this tailroom to
> > pass the MIC (8 bytes) + ICV (4 bytes) to the firmware. Do you need
> > more than 18 bytes?
> 
> I just went back and had a look at the code -- There were two related 
> workarounds: One made room for the MIC, but as you said, that hasn't 
> been necessary for a while so the code was disabled.
> 
> The other workaround is still in place, however -- cw1200 also needs a 
> gap in the frame payload for the IV, even though it's generated by the 
> firmware. I worked around this by telling setting the 
> IEEE80211_KEY_FLAG_GENERATE_IV flag on the mac80211 keys, and then 
> zeroing that part of the payload when the frame went out to the harware.
> 
> I figured it was a less costly approach than memmoving the whole payload 
> over by sizeof(ICV).
> 
> Is there a better way to handle this now?

There's IEEE80211_KEY_FLAG_PUT_IV_SPACE now which does exactly that.

johannes