From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from he.sipsolutions.net ([78.46.109.217]:42564 "EHLO sipsolutions.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754167Ab1BPI7d (ORCPT ); Wed, 16 Feb 2011 03:59:33 -0500 Subject: Re: [PATCH 4/5] iwlwifi: auto-tune tx queue size to minimize latency From: Johannes Berg To: "John W. Linville" Cc: Nathaniel Smith , linux-wireless@vger.kernel.org, wey-yi.w.guy@intel.com, ilw@linux.intel.com In-Reply-To: <20110215150317.GA2557@tuxdriver.com> References: <1297619803-2832-1-git-send-email-njs@pobox.com> <1297619803-2832-5-git-send-email-njs@pobox.com> <1297685871.3785.39.camel@jlt3.sipsolutions.net> <1297771986.3935.9.camel@jlt3.sipsolutions.net> <20110215150317.GA2557@tuxdriver.com> Content-Type: text/plain; charset="UTF-8" Date: Wed, 16 Feb 2011 09:59:25 +0100 Message-ID: <1297846765.4533.1.camel@jlt3.sipsolutions.net> Mime-Version: 1.0 Sender: linux-wireless-owner@vger.kernel.org List-ID: On Tue, 2011-02-15 at 10:03 -0500, John W. Linville wrote: > On Tue, Feb 15, 2011 at 01:13:06PM +0100, Johannes Berg wrote: > > > mac80211 has the advantage that TX status is reported back, so it could > > pretty accurately know how many frames are queued up -- more generically > > this cannot be done. > > > > But since it's not perfectly accurate right now, it may not be the best > > idea to do this at this time. > > Let's not make the perfect the enemy of the good. An imperfect > implementation at the mac80211 layer is likely better than a somewhat > better implementation for one driver and no implementation for > the rest. Well, no, I think the fact that it's not perfect probably means that the bandwidth/packet estimate will drift if any packets are dropped etc. > Now, someone help me understand how .11n aggregation complicates this... :-) I'll try ... later :-) johannes