Return-path: Received: from mx1.redhat.com ([209.132.183.28]:56333 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756599Ab3EGIlQ (ORCPT ); Tue, 7 May 2013 04:41:16 -0400 Date: Tue, 7 May 2013 10:42:42 +0200 From: Stanislaw Gruszka To: Johannes Berg Cc: Jake Edge , linux-wireless@vger.kernel.org, lkml Subject: Re: Bisected 3.9 regression for iwl4965 connection problem to 1672c0e3 Message-ID: <20130507084241.GA1581@redhat.com> (sfid-20130507_104140_020874_85B4C0B8) References: <20130505143803.7e46e4c6@chukar.edge2.net> <20130506123805.GA1602@redhat.com> <20130506083759.556dac76@chukar.edge2.net> <20130506153044.GB1602@redhat.com> <1367854279.8434.13.camel@jlt4.sipsolutions.net> <1367855046.8434.16.camel@jlt4.sipsolutions.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <1367855046.8434.16.camel@jlt4.sipsolutions.net> Sender: linux-wireless-owner@vger.kernel.org List-ID: On Mon, May 06, 2013 at 05:44:06PM +0200, Johannes Berg wrote: > On Mon, 2013-05-06 at 17:31 +0200, Johannes Berg wrote: > > > > But if so, I would also see > > > the breakage on my setup, but I don't - it works quite well here. > > > > Are you testing on a passive channel? Try with a large beacon interval. > > I think most likely what happens is that it's on a passive channel, and > the firmware drops the TX packet with a bad status. Before the patch, > we'd just wait sitting on the channel for HZ/5 (200ms) before trying > again, with the patch we immediately retransmit the packet, which will > fail again and again until the firmware received a beacon. > > If you look at iwlwifi/dvm/, it has some passive_no_rx workaround for > this, which I don't see in iwlegacy. Can you explain why it is named passive_no_rx instead passive_no_tx ? > I think the best way to solve this would be to do such a thing in > iwlegacy as well, but until then and for stable maybe we should > introduce another HW flag to restore the previous mac80211 behaviour? I'm not sure if I like to add passive_no_rx to iwlegacy. Stopping queues and waiting for beacon looks sticky, what happen if beacon will not be received? Perhaps I will just remove IEEE80211_HW_REPORTS_TX_ACK_STATUS from 4965, it's simpler workaround ? Stanislaw