From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-qk0-x230.google.com (mail-qk0-x230.google.com [IPv6:2607:f8b0:400d:c09::230]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by lists.bufferbloat.net (Postfix) with ESMTPS id E18113B260 for ; Sat, 19 Nov 2016 12:18:12 -0500 (EST) Received: by mail-qk0-x230.google.com with SMTP id q130so303670354qke.1 for ; Sat, 19 Nov 2016 09:18:12 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:in-reply-to:references:from:date:message-id:subject:to :cc:content-transfer-encoding; bh=FLDdEWIBDi+muHw53rEK8sIWs6oA9FxK0t87YI33xOw=; b=dzLpmVWKVrGp/gpmJk+k9z+yxZ8/mjXdRBXdVsXP1nKDboueibRwR8+YM2nK+x3mIC scR8MMuqJI93cHyYVEwFWBPDcejul96Bh7MN9N/CrGmew4aWZsEGAnApKGXCLPmUL4MJ SX5jBCN7teFCSnv14E/nAKwD6mPag3d45YDwgm6ZuFTGHKSBc2u30LaJQRKOxY1vIAwW j1yC2vjRhYNfnqnpCxwiLHIgWz5/S+EEU3jUycJBtLK+IDGCPriaxEwkjPZE3MkBV1mU Lp80qeILUfzBAT6iKMcY4mVbLw06naTNhOhW73xBht+OOQnApAkWqdaYKdf/x/ymG2BJ lvIA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20130820; h=x-gm-message-state:mime-version:in-reply-to:references:from:date :message-id:subject:to:cc:content-transfer-encoding; bh=FLDdEWIBDi+muHw53rEK8sIWs6oA9FxK0t87YI33xOw=; b=hHt9y+5W0hd/uh3qHXtbFEYM2nXiHpbJby9gWfhkuWdMehy5n/YKz2WsQsYWBIyYxz myvE4pGKom9kYC9cX/ENpKGrXdR7Bw5C0TUJOJx1zYSlpRirA3nYLiG5MpfF8dWrF2zT ZRWbSbI05pxKOvC//EsGhv1vF1ZaF7uHPQOQEXDlHV3ta/uR2vca5eqLrGJQo97jlGks RZjoIK1HebUmSVQiqHk0g5q00wYtiYVcFhQ8So39ebDUXdFsAGrtu1dMZC+i4ISWxTIV crDngWy3w6zMxQEASz1ZrzCqcbNVfOw3w7FTbnRxkckH/lkpKZj3EIAtlDhz+NL/nWnN TFiA== X-Gm-Message-State: AKaTC01eRXbgvMd5ETtYWU9sJD46IXyXGbXCm84HhjwsfvPYeRLYzdM0PM5+XHiiAI/UydbCpGAHlu5r+6+6Fg== X-Received: by 10.55.136.134 with SMTP id k128mr7416223qkd.196.1479575892388; Sat, 19 Nov 2016 09:18:12 -0800 (PST) MIME-Version: 1.0 Received: by 10.12.137.198 with HTTP; Sat, 19 Nov 2016 09:18:11 -0800 (PST) In-Reply-To: <48ff9504-958c-7c32-9da8-e227ce701851@gmail.com> References: <87zikyyspy.fsf@toke.dk> <871sy8ztz2.fsf@toke.dk> <87r368ybnz.fsf@toke.dk> <48ff9504-958c-7c32-9da8-e227ce701851@gmail.com> From: Dave Taht Date: Sat, 19 Nov 2016 09:18:11 -0800 Message-ID: To: Noah Causin Cc: make-wifi-fast@lists.bufferbloat.net Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Subject: Re: [Make-wifi-fast] Updated LEDE build with airtime patches X-BeenThere: make-wifi-fast@lists.bufferbloat.net X-Mailman-Version: 2.1.20 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 19 Nov 2016 17:18:13 -0000 Thx noah! I just tried this, and aside from lacking babeld by default (I need this to dynamically reroute the testbed), your QCA build works on the uap-lite. On the ath10k side, I get ~300mbit download rates from the osx box, (vs 150-200 on the archer) via the uap-lite, with about 12ms latency (in line with expectations). (in both test series I was using vht80). However, I do not see codel induced drops on that side: 9a:d9# cat aqm tid ac backlog-bytes backlog-packets new-flows drops marks overlimit collisions tx-bytes tx-packets 0 2 0 0 204580 0 0 0 0 1463032915 954453 And latencies get way higher if I use vht40, so I kind of suspect we still have multiple issues here. as for the first issue: dropping at the ag71xx? NAPI bug (as pointed out by eric dumazet)? Not enough patches? As for the latter, I was concerned about the NAPI patches recently added in for the ath10k, but don't know if those are in the 4.4 backport to lede. I'd rather like to try a napi budget of like 16, rather than 64 on router platforms that can't achieve a gbit in the first place. ... There is a great deal more work to do on testing more stations (I usually have to shut down the yurtlab in the winter, and cluttering up the apt with dozens of wifi stations is a matter of ongoing negotiation with gf), channels, rates, etc - I'm merely happy to have a smoke test succeed enough to be able to encourage more to try this stuff - and my first priority is nailing the ath9k to perfection..... the snapon build server has long been standing idle, if you want to use that for future builds. ... Testing the 2.4ghz (ath9k HT20) side of the uap, I get similar results as to the archer, and sane codel drops, etc, with about 40Mbits of throughput and sane latency (given how crowded the air is at the apt I'm surprised it's even this good) /sys/kernel/debug/ieee80211/phy1/netdev:wlan1/stations/48:d7:05:c0:9a:d9 # cat aqm tid ac backlog-bytes backlog-packets new-flows drops marks overlimit collisions tx-bytes tx-packets 0 2 168740 110 11230 644 0 0 0 239846710 156783 On Fri, Nov 18, 2016 at 7:00 PM, Noah Causin wrote: > Hi Dave, > > I built firmware for your Ubiquiti AC Lite from Toke's updated repository= . > > https://drive.google.com/drive/folders/0B86IEgnrkAk0YkxMalhlUUsyOFU?usp= =3Dsharing > > There are two versions. > > I made one using LEDE's stock ath10k firmware and one using Candelatech's > ath10k firmware. > > Packages: > > luci-ssl-openssl kmod-sched-cake luci-app-sqm luci-app-ntpc nano ethtool > ebtables-firewall ebtables-netfilter openssh-sftp-server dnsmasq wpad > basic-iputils (ping, ping6, etc...) > iperf iperf3 netperf tcpdump > > I changed the Cake and sqm-scripts Makefiles to point to their latest sta= ble > git versions. > > I changed the ath10k-firmware and ath10k-ct Makefiles to point to updated > versions. > > LEDE stock ath10k firmware: > > From > firmware-5.bin_10.2.4.70.54 > To > firmware-5.bin_10.2.4.70.58 > > Candelatech ath10k firmware: > > From > firmware-2-ct-full-community.bin-18.rc1-lede > To > firmware-2-ct-full-community.bin-18.rc2-lede > > Noah Causin > > _______________________________________________ > Make-wifi-fast mailing list > Make-wifi-fast@lists.bufferbloat.net > https://lists.bufferbloat.net/listinfo/make-wifi-fast --=20 Dave T=C3=A4ht Let's go make home routers and wifi faster! With better software! http://blog.cerowrt.org