Return-path: Received: from purkki.adurom.net ([80.68.90.206]:36173 "EHLO purkki.adurom.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751507Ab2JMFiq (ORCPT ); Sat, 13 Oct 2012 01:38:46 -0400 From: Kalle Valo To: "John W. Linville" Cc: Greg KH , Pontus Fuchs , linux-wireless@vger.kernel.org, hch@lst.de, s.L-H@gmx.de, mcgrof@qca.qualcomm.com Subject: Re: [PATCH 0/4] Driver for the ar5523 chipset References: <1349985702-21322-1-git-send-email-pontus.fuchs@gmail.com> <20121011203112.GB3317@kroah.com> <87haq017wi.fsf@purkki.adurom.net> <20121012143355.GA25235@tuxdriver.com> Date: Sat, 13 Oct 2012 08:38:42 +0300 In-Reply-To: <20121012143355.GA25235@tuxdriver.com> (John W. Linville's message of "Fri, 12 Oct 2012 10:33:56 -0400") Message-ID: <87a9vr0w4d.fsf@purkki.adurom.net> (sfid-20121013_073859_524465_A5E0B17B) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: linux-wireless-owner@vger.kernel.org List-ID: "John W. Linville" writes: >> I did a really quick 5 min review of this driver and IMHO this should go >> directly to drivers/net/wireless/ (or maybe to drivers/net/wireless/ath, >> Luis?), I don't see any reason why this should go to staging. The code >> looks clean and is only 2 kLOC. But I'll review this better during the >> weekend and also run some build tests. > > I have no particular objection to that. FWIW, Pontus contacted me > and said something like "I've got this driver that sucks and needs > a lot of work, should it go to staging?" Perhaps he was being too > hard on himself? :-) Well, we are not exactly famous about giving positive feedback ;) So no wonder new people are extra cautious. But after a more thorough review I still think that this driver should go to drivers/net/wireless. For example, it has proper endian support, it's both gcc and sparse warning free and the code is a pleasure to read(!). Here are my comments: * use ALIGN() and IS_ALIGNED() macros instead of doing it manually * checkpatch complains about DOS line endings (but that might be due to mail server, not sure) * git-am complained about whitespace damage * defines inside structures is not common, kernel style is to have the defines on top of the structure definition -- Kalle Valo