From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-ie0-x22b.google.com (mail-ie0-x22b.google.com [IPv6:2607:f8b0:4001:c03::22b]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (Client CN "smtp.gmail.com", Issuer "Google Internet Authority" (verified OK)) by huchra.bufferbloat.net (Postfix) with ESMTPS id B9FA921F1A7 for ; Thu, 11 Jul 2013 13:18:13 -0700 (PDT) Received: by mail-ie0-f171.google.com with SMTP id qd12so18646117ieb.2 for ; Thu, 11 Jul 2013 13:18:12 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type:content-transfer-encoding; bh=6kaL7xvI/5QQlPdvgBe+tF7OQZPL3Dm7nMmhCuELnkg=; b=cBSkaPRq/3ZWpK1YCGza9UAPEuFc+frklK5ERjP14YMZV5UP/171g4Xm0Y+EsMmLfX 6tgBldkQE/kr0EUmNYNISfrcpyQ7N4ZKMwtnlDrT1FWlIpIIUeiVcbC89ZKYtz6Wcjh2 W4n8DMbBE2EN8bQLNjJ4mdeqnBADt3kkWLUswiehQ+hedIBkTkApiYC450QhkNXLDo2u wdPBGrPhoSFD3l3rNdR59dJYhG43+sJTNMQmHDALaTZ/9bYmfkuU1hAPB5LwjxzMer61 qNGIdNDO9SdFva5XivM/K6kHB+7i9nTvZqnTQ+Sk8YsYtYI3aScXb6HqQIl6Qs087EEb c7HQ== MIME-Version: 1.0 X-Received: by 10.50.67.111 with SMTP id m15mr13790391igt.54.1373573892815; Thu, 11 Jul 2013 13:18:12 -0700 (PDT) Received: by 10.64.98.162 with HTTP; Thu, 11 Jul 2013 13:18:12 -0700 (PDT) In-Reply-To: <51DF0B8B.3050000@gmail.com> References: <51DEE063.5040307@laptop.org> <51DEEC00.801@laptop.org> <51DF0B8B.3050000@gmail.com> Date: Thu, 11 Jul 2013 13:18:12 -0700 Message-ID: From: Dave Taht To: Robert Bradley Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Cc: cerowrt-devel@lists.bufferbloat.net Subject: Re: [Cerowrt-devel] Turning off the probe blocker X-BeenThere: cerowrt-devel@lists.bufferbloat.net X-Mailman-Version: 2.1.13 Precedence: list List-Id: Development issues regarding the cerowrt test router project List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 11 Jul 2013 20:18:13 -0000 On Thu, Jul 11, 2013 at 12:46 PM, Robert Bradley wrote: > On 11/07/13 19:41, Dave Taht wrote: >> >> Nice stats for an oft-loaded fq_codel based wifi mesh network with not >> a lot of fixes (besides disabling 802.11e at key points so far)! >> >> I have seen much, much, much, worse from most other meshes. > > > I noticed yesterday that disabling WMM/802.11e in CeroWRT limited me to 5= 4 > Mb/s. Hmm. That shouldn't be the case... >The reduced rates are probably not an issue for the mesh network, > especially with 11g nodes present. Well, the mesh is all 5ghz presently. The 2Ghz nodes are dedicated to users (presently) and yes, 11g is everpresent... >Would it be better though to force the > traffic into one queue (BE, presumably) and keep the higher data rates? That's what I did. I just squashed the TOS/TCLASS values to BE at various chokepoints with iptables. It made an enormous difference, even when the main codepoints were BE and CS1. Maximizing for better aggregation and smartly managing the one queue worked tons better than keeping the diffserv markings "sane". In the long run, I think, I'm going to disable diffserv in the wifi driver itself and try to do something saner in a fq_codel derivative. Someday. A truly stupendous amount of traffic entering from the internet was (mis)marked CS1, so all traffic is squashed to BE there too, although on obviously torrent traffic the ingress rate limiter is doing the right thing.... at the netfiltter conference there was some discussion about doing "pairing" in conttrack between output flows (hopefully appropriately marked) and input flows (usually mismarked) in order to carry the intent around the internal net In fact, I just figured out why 172.20.142.10 was showing the latency spikes it was, it wasn't mashing its traffic, I'd forgot to make the masher be the default in the build. I am not huge on smashing the codepoint (I could explicitly set the packet priority field to 256 instead) #!/bin/sh ipt() { iptables $* ip6tables $* } #iptables doesn't support an inverted match #iptables -t mangle -A PREROUTING -m dscp ! --dscp-class BE -j DSCP --set-d= scp 0 ipt -t mangle -N FIX_TOS ipt -t mangle -A FIX_TOS -m dscp --dscp-class BE -j ACCEPT ipt -t mangle -A FIX_TOS -j DSCP --set-dscp 0 ipt -t mangle -A POSTROUTING -o wlan0 -j FIX_TOS > > -- > Robert Bradley > > > _______________________________________________ > Cerowrt-devel mailing list > Cerowrt-devel@lists.bufferbloat.net > https://lists.bufferbloat.net/listinfo/cerowrt-devel --=20 Dave T=E4ht Fixing bufferbloat with cerowrt: http://www.teklibre.com/cerowrt/subscribe.= html