From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from mail-qy0-f174.google.com ([209.85.216.174]:53647 "EHLO mail-qy0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754952Ab1ERHtL convert rfc822-to-8bit (ORCPT ); Wed, 18 May 2011 03:49:11 -0400 Received: by qyk7 with SMTP id 7so2546528qyk.19 for ; Wed, 18 May 2011 00:49:10 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: References: <1305677203-16660-1-git-send-email-zajec5@gmail.com> <1305677203-16660-2-git-send-email-zajec5@gmail.com> Date: Wed, 18 May 2011 09:49:10 +0200 Message-ID: (sfid-20110518_094914_637910_26974681) Subject: Re: [PATCH 2/9] b43: add bus device abstraction layer From: =?UTF-8?B?UmFmYcWCIE1pxYJlY2tp?= To: Julian Calaby Cc: linux-wireless@vger.kernel.org, "John W. Linville" , b43-dev@lists.infradead.org Content-Type: text/plain; charset=UTF-8 Sender: linux-wireless-owner@vger.kernel.org List-ID: W dniu 18 maja 2011 09:21 użytkownik Julian Calaby napisał: > 2011/5/18 Rafał Miłecki : >> W dniu 18 maja 2011 02:28 użytkownik Julian Calaby >> napisał: >>> Rafał, >>> >>> A quick question: >>> >>> 2011/5/18 Rafał Miłecki : >>>> >>>> Signed-off-by: Rafał Miłecki >>>> --- >>>>  drivers/net/wireless/b43/Makefile |    1 + >>>>  drivers/net/wireless/b43/b43.h    |    4 +++- >>>>  drivers/net/wireless/b43/bus.c    |   36 ++++++++++++++++++++++++++++++++++++ >>> >>> Would it make more sense to have this be called ssb.c as it contains >>> all the ssb specific functions, >> >> It's only ssb specific for now. It will contain BCMA code later. >> >> >>> that way you can then have an brcma.c >>> file to contain the functions specific to that bus? >> >> I'll put BCMA specific code in bus.c. >> Right now bus.c contains 100 LOC* and I believe its SSB part is >> complete. All the ops functions are one liners. It's so small and >> simple file I don't see sense to splitting it and having more mess in >> list of files instead. > > As I see it, having two sets of mostly identical  wrapper functions in > a file seems incorrect to me. Especially as once the abstraction is > complete it would technically be correct to build b43 without SSB > support - it's much cleaner to not compile a file than have a massive > #ifdef block in a common file. > > Anyway, it's only a minor thing. Massive? It's *one* ifdef for one bus type in this file. >> A one quick question: >> Why didn't you respond in "[RFC][PATCH] b43: add bus abstraction >> layer" on 2011-04-08? Or more recent "[RFC ONLY 2/5] b43: add bus >> device abstraction later" posted on 2011-05-09? > > While I try to read every patch that passes through the linux-wireless > mailing list, I only skim them, and tend to miss some details. The > thing that prompted this comment was the SSB comment at the start of > the SSB specific wrappers - something I probably didn't read the last > two times the patch came up on the list. OK, I ask because it's much easier to discuss such a things before you got 20 patches. That's why I posted very early RFC. -- Rafał