Return-path: Received: from smtp-out.google.com ([216.239.44.51]:59684 "EHLO smtp-out.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752335Ab0JNP7b (ORCPT ); Thu, 14 Oct 2010 11:59:31 -0400 Received: from kpbe12.cbf.corp.google.com (kpbe12.cbf.corp.google.com [172.25.105.76]) by smtp-out.google.com with ESMTP id o9EFxUtR008434 for ; Thu, 14 Oct 2010 08:59:30 -0700 Received: from iwn37 (iwn37.prod.google.com [10.241.68.101]) by kpbe12.cbf.corp.google.com with ESMTP id o9EFxTXw018266 for ; Thu, 14 Oct 2010 08:59:29 -0700 Received: by iwn37 with SMTP id 37so7591956iwn.15 for ; Thu, 14 Oct 2010 08:59:29 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: <1287067869.4641.440.camel@jlt3.sipsolutions.net> References: <20101013224851.B8E5D20465@glenhelen.mtv.corp.google.com> <1287064265.4641.435.camel@jlt3.sipsolutions.net> <1287066750.4641.439.camel@jlt3.sipsolutions.net> <1287067869.4641.440.camel@jlt3.sipsolutions.net> Date: Thu, 14 Oct 2010 08:59:28 -0700 Message-ID: Subject: Re: [PATCH 3/6] Add a function for evaluating rate changes for possible notification From: Paul Stewart To: Johannes Berg Cc: linux-wireless@vger.kernel.org Content-Type: text/plain; charset=ISO-8859-1 Sender: linux-wireless-owner@vger.kernel.org List-ID: On Thu, Oct 14, 2010 at 7:51 AM, Johannes Berg wrote: > On Thu, 2010-10-14 at 16:32 +0200, Johannes Berg wrote: > >> I wonder if "rate" is really the best interface for this. Since it's a >> threshold event, does it matter? Maybe there should instead be events >> for MCS changes and legacy rate changes so we can avoid all those >> calculations? > > In fact, since you'd only care about changes across a threshold etc, not > the absolute rate, you might be able to build the system without ever > doing the expensive rate calculation from MCS. What if I change the calculate() so that it does a cached table lookup? Since MCS numbers aren't linear or monotonic with regard to bitrate, those numbers can't be compared directly to set thresholds. I think it'd be a mistake to push it to userspace to interpret MCS and alert on every MCS change since that would just tradeoff calculation expense for nl80211 socket traffic, since without an understanding of the MCS ranking, it'd would not be possible to set thresholds. -- Paul