Return-path: Received: from mail-iw0-f174.google.com ([209.85.214.174]:46925 "EHLO mail-iw0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754297Ab0KUVuv convert rfc822-to-8bit (ORCPT ); Sun, 21 Nov 2010 16:50:51 -0500 MIME-Version: 1.0 In-Reply-To: References: <20101121130236.GE23423@thunk.org> <20101121172906.GD3703@kroah.com> <20101121203124.1ba8212e@lxorguk.ukuu.org.uk> <20101121214407.GH23423@thunk.org> From: "Luis R. Rodriguez" Date: Sun, 21 Nov 2010 13:50:30 -0800 Message-ID: Subject: Re: Challenges with doing hardware bring up with Linux first To: =?UTF-8?Q?G=C3=A1bor_Stefanik?= Cc: "Ted Ts'o" , Alan Cox , Greg KH , linux-kernel@vger.kernel.org, linux-wireless , David Miller , "John W. Linville" , Stephen Hemminger , "Perez-Gonzalez, Inaky" , Charles Marker , Jouni Malinen , Kevin Hayes , Zhifeng Cai , Don Breslin , Doug Dahlby , Julia Lawall Content-Type: text/plain; charset=UTF-8 Sender: linux-wireless-owner@vger.kernel.org List-ID: 2010/11/21 Gábor Stefanik : > On Sun, Nov 21, 2010 at 10:44 PM, Ted Ts'o wrote: >> On Sun, Nov 21, 2010 at 08:31:24PM +0000, Alan Cox wrote: >>> >>> Which we know in practice they won't. They'll sit on fixes (often >>> security fixes) and tweak and add private copies of features. In turn the >>> Linux one could then only keep up by adding features itself - which would >>> have to be GPL to stop the same abuse continuing. >>> >>> It's a nice idea but the corporations exist to make money and adding >>> proprietary custom stack add-ons is clearly a good move on their part to >>> do that. >> >> Hence my recommendation that if someone is going to do the work to >> create a 802.11 layer that has shims that work on multiple operating >> systems, it be GPL with explicit exceptions to allow said layer to >> work on legacy operating systems like QNX, et. al.  That way it forces >> the hardware specific code to be released under the GPL --- if they >> want to take advantage of the "write onces, work on multiple operating >> systems" feature. >> >> If someone is going to go through all of this work to make it possible >> --- particularly if it's at a company such as Luis's employer, or any >> other wifi chipset provider --- why should it allow their competitors >> to do closed source drivers?  Better to structure the driver licensing >> such that (a) there is benefit for companies to make a Linux driver by >> using this common stack, and (b) but in exchange, it forces them to >> make a driver which is guaranteed to be usable by Linux by virtual of >> the fact that (1) the native interface is Linux's wireless stack, and >> (2) the license forces them to GPL their driver. > > By forcing the driver to be GPL, you automatically exclude Windows > from the list of platforms supported by such a cross-OS driver, as the > Windows NDIS headers are AFAIK under a GPL-incompatible license, so no > GPL driver can be written for Windows. I've actually have been told GPL drivers for windows are possible with some hard work. I have yet to investigate further on what that "hard work" means. But yes, that is a good example. Luis