Return-path: Received: from mail-wy0-f174.google.com ([74.125.82.174]:34290 "EHLO mail-wy0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754339Ab1HOO4d (ORCPT ); Mon, 15 Aug 2011 10:56:33 -0400 Received: by wyg24 with SMTP id 24so3542388wyg.19 for ; Mon, 15 Aug 2011 07:56:32 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: <1313234481.2052.7.camel@localhost.localdomain> References: <1313234481.2052.7.camel@localhost.localdomain> Date: Mon, 15 Aug 2011 20:26:32 +0530 Message-ID: (sfid-20110815_165643_997523_D6A8765F) Subject: Re: [ath9k-devel] Packet loss regression with Atheros AR9280 From: Mohammed Shafi To: Jan Sundman Cc: robert@ocallahan.org, ath9k-devel@venema.h4ckr.net, linux-wireless@vger.kernel.org Content-Type: text/plain; charset=ISO-8859-1 Sender: linux-wireless-owner@vger.kernel.org List-ID: On Sat, Aug 13, 2011 at 4:51 PM, Jan Sundman wrote: > Hi, > > I'm not sure whether you have found a solution to this issue yet but I > thought that I should share my findings. I'm running Arch linux on my > laptop which have the AR9280 and I saw massive packet loss as well. is it an in built card? > > I noticed however that this _only_ occurred when the laptop was on > battery power. So it seems to be due to some power save functionality > for the card. if its an in built card, then this looks like a platform issue, if so please use http://www.kernel.org/pub/linux/kernel/people/mcgrof/scripts/cpudmalatency.c http://johannes.sipsolutions.net/files/netlatency.c.txt > > "iwconfig wlan0 power off" seems to fix it so I added this to my startup > script and this fixes it. This is of course not a optimal solution but > may point you in the right direction on fixing the problem. > > Br, > > /Jan > >>On 12 August 2011 10:27, Robert O'Callahan > > wrote: > >>> Anything I can get building with this kernel, I guess. Any particular >>> version I should try? > >>I'd suggest going back a month at a time >>(http://www.orbit-lab.org/kernel/compat-wireless-2.6/2011/), >try each >>until you find one that works, then report >"compat-wireless X works, >>compat-wireless Y doesn't". That'll be a good way to >test multiple >>copies of the development wireless tree. > >>If someone Fedora-clued knows which version(s) of the >wireless code >>are involved with Robert's good and bad behaviour >reports, that'll >>also help. > >>Anything to kick-start a git bisect :-) > >>Ciao, > > >>Adrian > > _______________________________________________ > ath9k-devel mailing list > ath9k-devel@lists.ath9k.org > https://lists.ath9k.org/mailman/listinfo/ath9k-devel > -- shafi