Return-path: Received: from mfe1.polimi.it ([131.175.12.23]:34385 "EHLO polimi.it" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1751412AbXLUVAv (ORCPT ); Fri, 21 Dec 2007 16:00:51 -0500 Date: Fri, 21 Dec 2007 21:57:58 +0100 From: Stefano Brivio To: Johannes Berg Cc: linux-wireless , sam , Michael Wu , Michael Buesch Subject: Re: [PATCH] mac80211: better rate control algorithm selection Message-ID: <20071221215758.4f962a39@morte> (sfid-20071221_210054_825614_75E999FE) In-Reply-To: <1198269816.16241.82.camel@johannes.berg> References: <1198253375.16241.76.camel@johannes.berg> <20071221210017.36f34fe3@morte> <1198269816.16241.82.camel@johannes.berg> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Sender: linux-wireless-owner@vger.kernel.org List-ID: On Fri, 21 Dec 2007 21:43:36 +0100 Johannes Berg wrote: > > What about avoiding to force any algorithm to y, and instead always build a > > dummy (a few lines of code which would just set the rate to the lowest > > available) rate control algorithm into mac80211? > > Why? I think that rc80211-simple is broken. A user may just want not to have a real rate control algorithm (i.e. no need for rates above the lowest one or certainty of perfect signal). Currently, that user would choose rc80211-simple, I guess, and well, this is almost fine, as he could manually set a rate. But let's say it's an embedded device, and small footprint is a must. Why would he need to use any RC algorithm (even other than rc80211-simple, as your Kconfig changes allow for this) then? mac80211 currently fails if no RC algorithms are available. So, I'd say, let's fix mac80211, so that it can work without any RC algorithm. And the default in this case could just be to set the lowest rate. That's what I called a dummy RC algorithm. -- Ciao Stefano