Return-path: Received: from ebb05.tieto.com ([131.207.168.36]:47920 "EHLO ebb05.tieto.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754071Ab2FTGOX (ORCPT ); Wed, 20 Jun 2012 02:14:23 -0400 From: Michal Kazior To: CC: Subject: [RFCv2] respect channels in iface combinations Date: Wed, 20 Jun 2012 08:14:06 +0200 Message-ID: <1340172859-18146-1-git-send-email-michal.kazior@tieto.com> (sfid-20120620_081432_065624_67F018DD) MIME-Version: 1.0 Content-Type: text/plain Sender: linux-wireless-owner@vger.kernel.org List-ID: Hi, Here goes an updated patchset with multi-channel work on cfg80211. Thanks go to Johannes and Eliad for review and pointing out my mistakes. I've addressed the issue with non-fixed channel IBSS. Channels are now cosidered either shared or exclusive. An exclusive channel always takes a channel resource (when accouting interface combinations and num_different_channels). Exclusive channel could also be called hopping but I thought the first one fits better as it explicitly states code logic decisions. I've also refactored and fixed interface combination checking. There are now functions like cfg80211_can_use_chan, cfg80211_can_use_iftype_chan. Monitor channel is now always defined. When there are only monitors running we set a channel (first we can find in supported bands structures). This way the corner case (of a monitor not having a channel and .get_channel returning NULL) mentioned by Johannes is fixed. This is based on wireless-next tree. -- Pozdrawiam / Best regards, Michal Kazior.