From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from he.sipsolutions.net ([78.46.109.217]:49746 "EHLO sipsolutions.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757737Ab2C1JS4 (ORCPT ); Wed, 28 Mar 2012 05:18:56 -0400 Message-Id: <20120328091346.442543219@sipsolutions.net> (sfid-20120328_111858_724256_2CFB2788) Date: Wed, 28 Mar 2012 11:13:46 +0200 From: Johannes Berg To: linux-wireless@vger.kernel.org Cc: Michal Kazior Subject: [RFC 0/2] mac80211 queue redesign Mime-Version: 1.0 Sender: linux-wireless-owner@vger.kernel.org List-ID: Here's the queue redesign we've been talking about. The monitor patch is needed because when a driver actually uses the new queue feature there isn't any queue to TX frames on for purely virtual interfaces. This only affects monitor (AP_VLAN are tied to AP) and only when injecting frames that don't have a TA matching another interface (as we bind the frames to that interface then). That, however is probably the biggest use case (now that monitor-for-AP is dying). As a consequence, drivers wanting to support "pure" injection and using the new queue features need to support the pure monitor interface feature as well. Also, the pure monitor feature will help implement monitor mode on devices where that is mutually exclusive with other modes (like iwlwifi.) The monitor feature I've tested in hwsim, it works. The queue feature isn't tested at all, I haven't even done any kind of regression tests for drivers not using it. Thanks goes out to everybody I discussed this with (on IRC or elsewhere), and also Michal for reviewing a first version of this patch :-) johannes