From mboxrd@z Thu Jan 1 00:00:00 1970 From: "John W. Linville" Subject: Re: 802.11 headers Date: Thu, 17 Aug 2006 08:43:33 -0400 Message-ID: <20060817124327.GA17771@tuxdriver.com> References: <1155800899.3653.13.camel@ux156> <20060817143443.77ede2e5@griffin.suse.cz> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Johannes Berg , netdev@vger.kernel.org, Jouni Malinen , David Woodhouse Return-path: Received: from ra.tuxdriver.com ([70.61.120.52]:24078 "EHLO ra.tuxdriver.com") by vger.kernel.org with ESMTP id S932468AbWHQMow (ORCPT ); Thu, 17 Aug 2006 08:44:52 -0400 To: Jiri Benc Content-Disposition: inline In-Reply-To: <20060817143443.77ede2e5@griffin.suse.cz> Sender: netdev-owner@vger.kernel.org List-Id: netdev.vger.kernel.org On Thu, Aug 17, 2006 at 02:34:43PM +0200, Jiri Benc wrote: > On Thu, 17 Aug 2006 09:48:19 +0200, Johannes Berg wrote: > > Or put another way, my suggestions: > > - put all in-kernel API (for drivers) into net/d80211.h > > - put stuff defined in the IEEE 802.11 specification series into > > linux/ieee80211.h, make user-visible (headers-y in Kbuild!). This > > serves just as a reference. It's not really necessary that the kernel > > ships this, but we need it in the kernel and probably won't be > > pulling it from some other source. Besides, I wouldn't know where > > else to maintain it. > > - radiotap simply moves to linux/ and gets to be userland-visible. Of > > course, d80211 should actually start using it too ;) > > - all the other userland stuff seems to be configuration related and > > hence moves over to linux/nl80211.h > > > > Comments? > > I like this proposal. I second that. John -- John W. Linville linville@tuxdriver.com