From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from sabertooth02.qualcomm.com ([65.197.215.38]:22816 "EHLO sabertooth02.qualcomm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751340AbaFDQPi (ORCPT ); Wed, 4 Jun 2014 12:15:38 -0400 Date: Wed, 4 Jun 2014 21:46:21 +0530 From: Rajkumar Manoharan To: Johannes Berg CC: , Subject: Re: [PATCH 00/13] ath9k: Channel context support Message-ID: <20140604161619.GA14852@qca.qualcomm.com> (sfid-20140604_181542_431423_2F0FD53C) References: <1401889599-412-1-git-send-email-rmanohar@qti.qualcomm.com> <1401890446.6079.31.camel@jlt4.sipsolutions.net> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" In-Reply-To: <1401890446.6079.31.camel@jlt4.sipsolutions.net> Sender: linux-wireless-owner@vger.kernel.org List-ID: On Wed, Jun 04, 2014 at 04:00:46PM +0200, Johannes Berg wrote: > On Wed, 2014-06-04 at 19:16 +0530, Rajkumar Manoharan wrote: > > > These patches adds a basic infrastruture for multi channel > > context. Currently channel context usage is restricted by > > modparam "use_chanctx" as it is not yet fully implemented. > > This isn't really a review - just a note. You should probably also tell > mac80211 that you use multiple HW queues (at least per chanctx) so you > avoid a situation where stopping one channel context leads to stopping > the entire traffic for that AC even on other contexts, otherwise you > could unfairly stop the other interface? > Completely agree. IEEE80211_HW_QUEUE_CONTROL support is still in-progress. > I guess I'm also surprised by the implementation just switching from a > worker, I would have thought you'd switch from a HW timer interrupt and > use different hardware queues, but hey :) > These are just a preparatory changes for enabling MCC. The actual scheduler is based on HW timers that will kickstart the worker thread. Will post all of the changes after completing regression test cycle. -Rajkumar