Return-path: Received: from mail-ww0-f46.google.com ([74.125.82.46]:46888 "EHLO mail-ww0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752966Ab0BHJXw (ORCPT ); Mon, 8 Feb 2010 04:23:52 -0500 Received: by wwi18 with SMTP id 18so1625914wwi.19 for ; Mon, 08 Feb 2010 01:23:51 -0800 (PST) To: Andres Salomon Cc: linux-wireless@vger.kernel.org, Johannes Berg , akpm@linux-foundation.org Subject: Re: [PATCH 2/2] mac80211: give warning if building w/out rate ctrl algorithm References: <20100207214837.027dfd3d@mycelium.queued.net> From: Kalle Valo Date: Mon, 08 Feb 2010 11:23:47 +0200 In-Reply-To: <20100207214837.027dfd3d@mycelium.queued.net> (Andres Salomon's message of "Sun\, 7 Feb 2010 21\:48\:37 -0500") Message-ID: <87vde8gjx8.fsf@purkki.valot.fi> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: linux-wireless-owner@vger.kernel.org List-ID: Andres Salomon writes: > I discovered that if EMBEDDED=y, one can accidentally build a > mac80211 stack w/ no rate control algorithm. When RC_MINISTREL and > RC_PID are both disabled, the RC_DEFAULT string (which rate.c uses > as the fallback algorithm) will be "". That'll cause the > rate_control_alloc to fail, which will in turn cause > ieee80211_register_hw to fail. IOW, no driver will load. There are devices that don't need rate control algorithm, for example wl1271 (see IEEE80211_HW_HAS_RATE_CONTROL). A better solution is instead to fix rate_control_alloc() to handle this, or something similar. -- Kalle Valo