Return-path: Received: from s3.sipsolutions.net ([144.76.43.152]:48049 "EHLO sipsolutions.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751199AbaC1IhW (ORCPT ); Fri, 28 Mar 2014 04:37:22 -0400 Message-ID: <1395995839.4175.5.camel@jlt4.sipsolutions.net> (sfid-20140328_093730_006847_962ED9FD) Subject: Re: [PATCH] mac80211: protect AP VLAN list with local->mtx From: Johannes Berg To: Michal Kazior Cc: linux-wireless Date: Fri, 28 Mar 2014 09:37:19 +0100 In-Reply-To: (sfid-20140320_081256_908709_730DC511) References: <1394021648-22605-1-git-send-email-michal.kazior@tieto.com> <1395238008.4142.15.camel@jlt4.sipsolutions.net> (sfid-20140320_081256_908709_730DC511) Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 Sender: linux-wireless-owner@vger.kernel.org List-ID: On Thu, 2014-03-20 at 08:12 +0100, Michal Kazior wrote: > On 19 March 2014 15:06, Johannes Berg wrote: > > On Wed, 2014-03-05 at 13:14 +0100, Michal Kazior wrote: > >> It was impossible to change chanctx of master AP > >> for AP VLANs because the copy function requires > >> RTNL which can't be simply taken in mac80211 code > >> due to possible deadlocks. > >> > >> This is required for future chanctx reservation > >> that re-bind vifs to new chanctx. This requires > >> safe AP VLAN iteration without RTNL. > >> > >> Now VLANs can be iterated while holding either > >> RTNL or local->mtx because the list is modified > >> while holding both of these locks. > > > > No objection really, but maybe it would make more sense to use > > iflist_mtx? > > I used local->mtx because it seemed easier at the time (the lock is > already used on all related codepaths). > > Using local->iflist_mtx would add another mutex to for > csa/reservation. I think it shouldn't be hard to do it though. Should > I re-spin? (this will probably need a re-spin of Luca's reservation > patchset and my RFC). No, it's fine, I was just curious. johannes