From mboxrd@z Thu Jan 1 00:00:00 1970 From: "John W. Linville" Subject: Re: Please pull 'upstream' branch of wireless-2.6 Date: Mon, 7 May 2007 19:38:11 -0400 Message-ID: <20070507233811.GB4245@tuxdriver.com> References: <20070507175121.GF5125@tuxdriver.com> <463FB1A0.3070608@garzik.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: linux-wireless@vger.kernel.org, netdev@vger.kernel.org To: Jeff Garzik Return-path: Received: from ra.tuxdriver.com ([70.61.120.52]:4689 "EHLO ra.tuxdriver.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S967313AbXEHAC1 (ORCPT ); Mon, 7 May 2007 20:02:27 -0400 Content-Disposition: inline In-Reply-To: <463FB1A0.3070608@garzik.org> Sender: netdev-owner@vger.kernel.org List-Id: netdev.vger.kernel.org On Mon, May 07, 2007 at 07:09:20PM -0400, Jeff Garzik wrote: > John W. Linville wrote: > >Ivo van Doorn (1): > > Add 93cx6 eeprom library > >Michael Wu (1): > > Add rtl8187 wireless driver I presume these are the two parts you question. (Just checking...) > The normal development process is: > > * commit new code to your repository > * that goes into -mm for public testing and review > * merge window opens > * the code that has seen public testing and review goes upstream > > The general idea is everything you want in 2.6.22 should be prepared and > in -mm BEFORE 2.6.21 is released, and the 2.6.22 merge window opens. > > It's quite normal for fixes and minor changes to trickle in after the > push that follows the merge window opening. But new drivers that have > not been through this process do not fall under "fixes and minor changes." I agree whole-heartedly. The whole mac80211 driver package (including rtl8187) has been in -mm (and rawhide) for most of the 2.6.21 development cycle. The rtl8187 driver in particular has been relatively stable for the past couple of months ("3 files changed, 13 insertions(+), 4 deletions(-)" since 7 March 2007), and it seems to work well. > I also did not see any response to my "better as a single file driver" > suggestion. Yes, I see that now. Do you consider this a merge requirement? My main concern pre-merge would be accidentally destabilizing the code while stitching the files together. Other than that, the suggestion seems worthwhile. However, I think Michael plans to expand the driver to support rtl8180 and rtl8185. This may factor into why he chose to organize the code the way he has. > So, I'll pull if you remove the two late additions. Any chance that you find the comments above persuasive? :-) John -- John W. Linville linville@tuxdriver.com