From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-wm0-x229.google.com (mail-wm0-x229.google.com [IPv6:2a00:1450:400c:c09::229]) (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 21DE23BA8E for ; Thu, 6 Sep 2018 13:51:32 -0400 (EDT) Received: by mail-wm0-x229.google.com with SMTP id 207-v6so12070396wme.5 for ; Thu, 06 Sep 2018 10:51:32 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=heistp.net; s=google; h=from:message-id:mime-version:subject:date:in-reply-to:cc:to :references; bh=HTsORDu91rEwHw1UgAihnLvnWta8LACEzoYTi9LK8So=; b=CyEYurwL5ewOh01B2HHbqsAKhrI7emQsrGE5ncT6Fup+M9JYM1PDQKO/E7R34rAOgx eYTc8bmLkAPqZODQXPiSh5PCre2Ge/qWhlMAb0lLbxKud/uvq3gKLdUXJdwMxgGgv6YG CPmFoYUWkiMbif3QDWIBgazczGSnvfVhG8q0G3FbggyXXsng0KObYW5XUuicnfCUkZHE yz5R1/NfRXEz7wj0YLcZlRZJcTNegZcZo6M6D3eOCFIIY6WF0L/F4e7Dcqxmp7KXIF26 8nGTpUZA0JqiBD2osoqm2QwQdO3vQTGO2RqPvYxc8N8ngWNWHMEDhIOY3m/1vKmdQSub fBOA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:message-id:mime-version:subject:date :in-reply-to:cc:to:references; bh=HTsORDu91rEwHw1UgAihnLvnWta8LACEzoYTi9LK8So=; b=mvsmXrRJF01c/QVq3DYVg0hJQhEEBsSrP1U+MEso3hRUzVzfAK0zxGaXM9y1Rg1sIc mkc4KKnKI6TFqBsC9p1HsrGrGEJIioyeaMVhxN4AJ0I5p5z4t52c5DW7fkhRZ4IvPQBa pndF7tKScJyLjoBP2tOdmXC0znKta6NYe2TFXNDYtBTLyyWUikdcrPsrjBVy7cTsvEhU 4/uobcQ5t7LfIJdpCBZVvszeMLLEewUqz+KEy0WE61GJFIZ5H0P3RBNKuI1bnJ9m8cin X7RdnMSrSdYjP7pgVeqK2yHOcrokP6yGiieySKIYvc/RFSuACUMOOzuIuRY1dxqbR+X+ 0jIA== X-Gm-Message-State: APzg51D5/EQIaisfEbJqoXGyr4VXhM/V3dSuXmRTqqTCHGsqSM2Yw2oE MgYWijrF+CdVtud3CgqXSG6vNQ== X-Google-Smtp-Source: ANB0VdY7fCWRSIvHoD+1xxYqkZTeodeqaniPjqYFm+yJz/yB3o8LcqSBxOr9TgBT4GUe2gb1ezphcQ== X-Received: by 2002:a1c:3411:: with SMTP id b17-v6mr2757021wma.85.1536256291306; Thu, 06 Sep 2018 10:51:31 -0700 (PDT) Received: from tron.luk.heistp.net (h-1169.lbcfree.net. [185.193.85.130]) by smtp.gmail.com with ESMTPSA id k5-v6sm13165785wrm.96.2018.09.06.10.51.30 (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Thu, 06 Sep 2018 10:51:30 -0700 (PDT) From: Pete Heist Message-Id: <1D2079D5-3BA8-4968-AD5A-990AFA3A7698@heistp.net> Content-Type: multipart/alternative; boundary="Apple-Mail=_61612EBE-255F-4707-867D-740D681C69AA" Mime-Version: 1.0 (Mac OS X Mail 11.5 \(3445.9.1\)) In-Reply-To: Cc: Cake List , cerowrt-devel@lists.bufferbloat.net To: Dave Taht References: X-Mailer: Apple Mail (2.3445.9.1) X-Mailman-Approved-At: Mon, 30 Mar 2020 07:21:14 -0400 Subject: Re: [Cerowrt-devel] [Cake] apu2 sqm/htb issue + a minor win for speeding up fq_codel itself X-BeenThere: cerowrt-devel@lists.bufferbloat.net X-Mailman-Version: 2.1.20 Precedence: list List-Id: Development issues regarding the cerowrt test router project List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Date: Thu, 06 Sep 2018 17:51:32 -0000 X-Original-Date: Thu, 6 Sep 2018 19:51:29 +0200 X-List-Received-Date: Thu, 06 Sep 2018 17:51:32 -0000 --Apple-Mail=_61612EBE-255F-4707-867D-740D681C69AA Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=utf-8 Cool, well I for one would like to see the APU be able to handle higher = speeds, for FreeNet=E2=80=99s backhaul, at least. Although frankly, = I=E2=80=99ve not definitively witnessed any significant bloat in their = backhaul yet with production traffic. A good number of their routers are still ALIX = (https://www.pcengines.ch/alix2d2.htm = ), all of which are on an upgrade = list. These don=E2=80=99t do hfsc + sfq on kernel 2.6.26 much beyond = about 70 Mbit. Not a problem to focus on=E2=80=A6 :) > On Sep 4, 2018, at 11:16 PM, Dave Taht wrote: >=20 > my guess is that burst and cburst should scale roughly as a function > of the bytes that can fit into 1ms. > On Tue, Sep 4, 2018 at 2:14 PM Dave Taht wrote: >>=20 >> making htb's cburst and burst parameters 64k gets the APU2 up to >> where it can shape 900mbits. 3 ksoftirq handlers start getting cpu >> time, and we end up 54% idle to achiefe that. >>=20 >> I should really go around running my own old code. I was deeply >> involved in sqm when we still had to run at sub 200mbit levels. since >> then it's been >> mostly tbf (burst 64k) + fq_codel or cake, and me ignoring various = bug >> reports about it not scaling well enough at higher rates. >> On Tue, Sep 4, 2018 at 12:59 PM Dave Taht = wrote: >>>=20 >>> less than scientifically (via monitoring top) - on the apu2 >>>=20 >>> 100Mbit sqm (htb + fq_codel) >>>=20 >>> fq_codel_mainline | fq_codel_fast >>> idle 78.8 | 83.5 | >>> si 20 | 16.1 | >>>=20 >>> Yea! But: >>>=20 >>> 900Mbit sqm (htb + fq_codel) >>>=20 >>> fq_codel_mainline | fq_codel_fast >>> idle 74.4 | 74.4 | >>> si 25 | 25.1 | >>>=20 >>> Here: completely bottlenecked on ksoftirqd - and I only get 340Mbits >>> out of the 900mbit setting. quantum 96k and burst of 15000. Haven't >>> fiddled with higher values yet... --Apple-Mail=_61612EBE-255F-4707-867D-740D681C69AA Content-Transfer-Encoding: quoted-printable Content-Type: text/html; charset=utf-8 Cool,= well I for one would like to see the APU be able to handle higher = speeds, for FreeNet=E2=80=99s backhaul, at least. Although frankly, = I=E2=80=99ve not definitively witnessed any significant bloat in their = backhaul yet with production traffic.

A good number of their routers are = still ALIX (https://www.pcengines.ch/alix2d2.htm), all of which are = on an upgrade list. These don=E2=80=99t do hfsc + sfq on kernel 2.6.26 = much beyond about 70 Mbit. Not a problem to focus on=E2=80=A6 :)

On Sep 4, 2018, at 11:16 PM, Dave Taht <dave.taht@gmail.com>= wrote:

my guess is that burst and cburst should scale roughly as a = function
of the bytes that can fit into 1ms.
On Tue, Sep 4, 2018 at 2:14 PM Dave Taht <dave.taht@gmail.com>= wrote:

making htb's cburst and burst parameters 64k gets the APU2 =  up to
where it can shape 900mbits. 3 ksoftirq = handlers start getting cpu
time, and we end up 54% idle to = achiefe that.

I should really go around = running my own old code. I was deeply
involved in sqm when = we still had to run at sub 200mbit levels. since
then it's = been
mostly tbf (burst 64k) + fq_codel or cake, and me = ignoring various bug
reports about it not scaling well = enough at higher rates.
On Tue, Sep 4, 2018 at 12:59 PM = Dave Taht <dave.taht@gmail.com> wrote:

less than scientifically (via = monitoring top) - on the apu2

100Mbit sqm = (htb + fq_codel)

fq_codel_mainline | = fq_codel_fast
idle 78.8 =             &n= bsp;  | 83.5 |
si    20 =             &n= bsp;     | 16.1 |

Yea! But:

900Mbit sqm (htb + = fq_codel)

fq_codel_mainline | = fq_codel_fast
idle 74.4 =             &n= bsp;  | 74.4 |
si    25 =             &n= bsp;     | 25.1 |

Here: completely bottlenecked on ksoftirqd - and I only get = 340Mbits
out of the 900mbit setting. quantum 96k and burst = of 15000. Haven't
fiddled with higher values yet...

= --Apple-Mail=_61612EBE-255F-4707-867D-740D681C69AA--