Return-path: Received: from mail.candelatech.com ([208.74.158.172]:42593 "EHLO ns3.lanforge.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754197Ab0I1PRJ (ORCPT ); Tue, 28 Sep 2010 11:17:09 -0400 Message-ID: <4CA206F0.40607@candelatech.com> Date: Tue, 28 Sep 2010 08:17:04 -0700 From: Ben Greear MIME-Version: 1.0 To: Johannes Berg CC: linux-wireless@vger.kernel.org Subject: Re: [PATCH 2/2] mac80211: Allow scanning single channel if other VIF is associated. References: <1285632457-27539-1-git-send-email-greearb@candelatech.com> <1285632457-27539-2-git-send-email-greearb@candelatech.com> <1285658564.3885.1.camel@jlt3.sipsolutions.net> <4CA1F2A4.2090005@candelatech.com> <1285682224.3885.18.camel@jlt3.sipsolutions.net> In-Reply-To: <1285682224.3885.18.camel@jlt3.sipsolutions.net> Content-Type: text/plain; charset=UTF-8; format=flowed Sender: linux-wireless-owner@vger.kernel.org List-ID: On 09/28/2010 06:57 AM, Johannes Berg wrote: > On Tue, 2010-09-28 at 06:50 -0700, Ben Greear wrote: > >> At best, it would be a race for wpa_s to determine if any other interface >> on it's interface's hardware is associated, so I don't see a good way to >> do this in user-space. >> >> When at least one is associated, I don't want the NIC to go offchannel at all, >> at least not for wpa_s scanning. Anything else is going to interrupt other >> stations's traffic, and for no good reason that I can see, since the scanning >> interface must associate on the same channel as the rest of the stations >> anyway. >> >> If you have any suggestions for how to accomplish this, please let me know.. >> otherwise, I can just carry this patch in my tree. > > You can run all networks in a single wpa_s instance, I believe, and then > it'd know about all this, right? But then, if you want to add an additional interface, you have to restart everything. I was thinking that maybe I could keep the logic in wpa_s, but instead of all the special casing, just locate the current associated channel for the phy in question and populate the scan-req with that single channel. Then, I think I wouldn't have to muck with much of the scan logic..just a bit of code on the entry point to select the proper channel. Does that sound like a possible solution to you? Thanks, Ben > > johannes -- Ben Greear Candela Technologies Inc http://www.candelatech.com