Return-path: Received: from relay2.ukrpost.ua ([195.5.6.22]:59697 "EHLO relay2.ukrpost.ua" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755260Ab1BIOVL (ORCPT ); Wed, 9 Feb 2011 09:21:11 -0500 Received: from mail.academy.zt.ua ([82.207.120.245]) by relay2.ukrpost.ua with esmtp (Exim 4.72) (envelope-from ) id 1PnALx-0007P9-4C for linux-wireless@vger.kernel.org; Wed, 09 Feb 2011 15:45:05 +0200 Received: from [10.0.2.42] by mail.academy.zt.ua (Cipher SSLv3:RC4-MD5:128) (MDaemon PRO v11.0.3) with ESMTP id md50000019954.msg for ; Wed, 09 Feb 2011 15:43:22 +0200 Subject: SSB AI support code From: George Kashperko To: linux-wireless@vger.kernel.org Content-Type: text/plain Date: Wed, 09 Feb 2011 15:36:30 +0200 Message-Id: <1297258590.17400.37.camel@dev.znau.edu.ua> Mime-Version: 1.0 Sender: linux-wireless-owner@vger.kernel.org List-ID: Hi guys. Starting this thread in order to discuss SSB AI support. Recently there were discussion on this matter but unfortunately missed it therefore starting new thread. Recently was playing around making new kernels work on ASUS RT-N16 (bcm4716-based with AI style bus). Started with commits by Bernhard Loos early 2010, and finally managed to get things work. Currently with code modified/added mine bcm4716 have working usb ehci/ohci and ethernet. Well, its pretty much everything the box have, except wifi - currently i'm half-way done with porting bcm80211 nphy code changes to b43 (Hopefully recent commits by Rafal Milecki will save me from porting more, but haven't tried it yet). All changes made were tested on SB-based hardware (two old bcm5354-based boxes). Unfortunately have no access to other BCM hw I could test things on, but I'm pretty much sure old hardware should not be affected by code changes while newer ones have some chance to get supported. Changes made are based on ASUS GPL sourcecodes for N16, bcm80211 staging code and bcm-v4.sipsolutions.net researches. I will post patches here as soon as get this message back in my emailer to keep things in single thread. Please note these changes are how things work for me. It's not commit request or anything alike. Most of the changes are easy, while some (well, actually just one - 5th patch with core state/control management propagation) should be well thought of before applying. Actually things with 5th patch could be done in completely backwards way - keep existing TMSLOW/TMSHIGH defines intact with flags in two high bytes and do ">> 16" things in AI code instead of "<< 16" in SB code but again - its just the way things work for me currently, not the way I could insist or something on. This is partial subset of changes to get newer BCM4716s' to work with linux. If anyone is interested the whole set of patches will be accessible in few hours via http - just left few things to manage with them.