From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from mail.candelatech.com ([208.74.158.172]:35860 "EHLO ns3.lanforge.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S933800Ab3BTOYi (ORCPT ); Wed, 20 Feb 2013 09:24:38 -0500 Message-ID: <5124DCA0.6030106@candelatech.com> (sfid-20130220_152440_526849_CEE5290A) Date: Wed, 20 Feb 2013 06:24:32 -0800 From: Ben Greear MIME-Version: 1.0 To: Johannes Berg CC: linux-wireless@vger.kernel.org Subject: Re: [PATCH] mac80211: Clean up work-queues on disassociation. References: <1361326267-16847-1-git-send-email-greearb@candelatech.com> (sfid-20130220_031220_995074_BF5628B4) <1361354394.8629.11.camel@jlt4.sipsolutions.net> <5124D90E.7070307@candelatech.com> <1361369637.8629.20.camel@jlt4.sipsolutions.net> In-Reply-To: <1361369637.8629.20.camel@jlt4.sipsolutions.net> Content-Type: text/plain; charset=UTF-8; format=flowed Sender: linux-wireless-owner@vger.kernel.org List-ID: On 02/20/2013 06:13 AM, Johannes Berg wrote: > On Wed, 2013-02-20 at 06:09 -0800, Ben Greear wrote: > >>>> @@ -2668,6 +2681,10 @@ static void ieee80211_sta_connection_lost(struct ieee80211_sub_if_data *sdata, >>>> false, frame_buf); >>>> mutex_unlock(&ifmgd->mtx); >>>> >>>> + /* Have to do this outside the ifmgd->mtx lock. */ >>>> + cancel_work_sync(&ifmgd->monitor_work); >>>> + cancel_work_sync(&ifmgd->beacon_connection_loss_work); >>> >>> OTOH, you do this many many times, and that doesn't seem necessary... >>> >>> If the work structs run when we disconnected, that's ok, they just >>> musn't run after we destroy the interface, so I think it'd be much >>> better to just put the two lines into ieee80211_mgd_stop() instead of >>> all the other places. >> >> Ok, that sounds promising to me. For that matter, could we just call the sta_quiesce >> method in mgd_stop? It would be nice to have all of the final timer and work-queue cleanup >> in a single place. > > That might be interesting, but I think we'd have to clear the > timer_pending thing afterwards? Are you talking about this code? if (del_timer_sync(&ifmgd->timer)) set_bit(TMR_RUNNING_TIMER, &ifmgd->timers_running); if (del_timer_sync(&ifmgd->chswitch_timer)) set_bit(TMR_RUNNING_CHANSW, &ifmgd->timers_running); If so, maybe make a helper method that does all of the first part of the sta_quiesce() logic, call that from both mgd_stop and the new sta_quiesce() method, and then have sta_quiesce look like: sta_cleanup_timers_and_work(); if (del_timer_sync(&ifmgd->timer)) set_bit(TMR_RUNNING_TIMER, &ifmgd->timers_running); if (del_timer_sync(&ifmgd->chswitch_timer)) set_bit(TMR_RUNNING_CHANSW, &ifmgd->timers_running); ? > >> And, maybe add some checks in the work-callback methods to bail early if the station >> isn't connected? > > They already do :) Ok, thanks. Ben -- Ben Greear Candela Technologies Inc http://www.candelatech.com