From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from smtp171.iad.emailsrvr.com (smtp171.iad.emailsrvr.com [207.97.245.171]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) by huchra.bufferbloat.net (Postfix) with ESMTPS id AE79C21F107; Fri, 21 Dec 2012 11:34:05 -0800 (PST) Received: from localhost (localhost.localdomain [127.0.0.1]) by smtp27.relay.iad1a.emailsrvr.com (SMTP Server) with ESMTP id AE19A118DA8; Fri, 21 Dec 2012 14:34:04 -0500 (EST) X-Virus-Scanned: OK Received: from legacy12.wa-web.iad1a (legacy12.wa-web.iad1a.rsapps.net [192.168.4.98]) by smtp27.relay.iad1a.emailsrvr.com (SMTP Server) with ESMTP id 80D7F118DA5; Fri, 21 Dec 2012 14:34:04 -0500 (EST) Received: from reed.com (localhost.localdomain [127.0.0.1]) by legacy12.wa-web.iad1a (Postfix) with ESMTP id 6BB05F78001; Fri, 21 Dec 2012 14:34:04 -0500 (EST) Received: by apps.rackspace.com (Authenticated sender: dpreed@reed.com, from: dpreed@reed.com) with HTTP; Fri, 21 Dec 2012 14:34:04 -0500 (EST) Date: Fri, 21 Dec 2012 14:34:04 -0500 (EST) Subject: =?utf-8?Q?Re=3A_=5BCerowrt-devel=5D_hardware_hacking_on_fq=5Fcodel_in_FPG?= =?utf-8?Q?A_form_at_10GigE?= From: dpreed@reed.com To: "Jim Gettys" MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_20121221143404000000_46376" Importance: Normal X-Priority: 3 (Normal) X-Type: html In-Reply-To: References: <20121220103213.4ABAD800037@ip-64-139-1-69.sjc.megapath.net> <1356011592.2372508@apps.rackspace.com> <1356112139.35814863@apps.rackspace.com> <1356114301.310221396@apps.rackspace.com> Message-ID: <1356118444.43673943@apps.rackspace.com> X-Mailer: webmail7.0 Cc: "codel@lists.bufferbloat.net" , bloat-devel , "cerowrt-devel@lists.bufferbloat.net" , blogic@openwrt.org X-BeenThere: bloat-devel@lists.bufferbloat.net X-Mailman-Version: 2.1.13 Precedence: list List-Id: "Developers working on AQM, device drivers, and networking stacks" List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 21 Dec 2012 19:34:06 -0000 ------=_20121221143404000000_46376 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable =0AUnderstand about the chip vendors. It is the case, though, that system = vendors often have NDA's with the chip vendors ... and they are the immedi= ate customer of the chip vendors, so have some leverage to get their help r= esolving the issues.=0A =0AI think the first step is doing the right measur= ements, to show that the chip is a bottleneck (as you are doing with Lantiq= ).=0A =0A-----Original Message-----=0AFrom: "Jim Gettys" =0ASent: Friday, December 21, 2012 1:45pm=0ATo: "David P Reed" =0ACc: "Dave Taht" , "bloat-devel" , "cerowrt-devel@lists.bufferbloat.net" , "codel@lists.bufferbloat.net" , blogic@openwrt.org=0ASubject: Re: [Cerowrt-devel] hardware hac= king on fq_codel in FPGA form at 10GigE=0A=0A=0A=0A=0A=0A=0A=0A=0AOn Fri, D= ec 21, 2012 at 1:25 PM, <[mailto:dpreed@reed.com] dpreed@reed.com> wrote:= =0A=0AI'm sure the HSPA+ modem is a binary blob. For regulatory reasons, i= f nothing else...=0A =0AWhen I said open, I did not mean the ultimate ideal= . Hardware devices with firmware are often sold as black boxes.=0A =0AFixi= ng the bloat in HSPA+ and LTE systems will require organizations like Erics= son, ALU and Cisco and Huawei to get their acts together and fix their own = bugs.=0AActually, this needs probably/certainly needs help from the chip ve= ndors as well as the system vendors. The system vendors (such as ALU) ofte= n get their chips and drivers from those chip vendors.=0A=0AI haven't looke= d into the HSPA+ or LTE systems; I am looking into DSL. In the DSL concret= e example, Broadcom controls the bulk of the DSL chip market, and (at least= on the client home router end) provides a binary blob Linux driver based o= n an antique Linux release (2.6.26, IIRC). The hardware itself did not look= like it had bad buffering, but the drivers....=0AI've been working with bl= ogic (John Crispin) on debloating the Lantiq driver, which we are trying to= test; since it is entirely open source. In the Lantiq DSL driver case, th= e hardware sprouted exactly the same sort of transmit/receive rings found i= n ethernet drivers, with the corresponding bufferbloat due to device driver= s being totally stupid. Unfortunately, BQL is ethernet only, and right now= doesn't work properly at low bandwidths.=0AWhen John cut the transmit ring= down to 2 packets, the latency dropped (at the bandwidth he was testing at= , .5Mbps) to around 60ms running fq_codel, about what you would expect give= n the remaining buffers; we have to automate the control of the transmit ri= ng properly (since we probably do need some help when running VDSL at 100Mb= ps).=0AI suspect/expect we'll find similar "features" when we dig into LTE = and HSPA+, though that is not certain.=0ANow, back to my saga of my recalci= trant water well, after which I can go back to testing DSL the next trip to= New Jersey (and flush the toilets and have heat; I have geothermal heat...= ).=0ASigh,=0A- Jim=0A=0A=0A =0A-----Original Message-----=0AFrom: "Jim Gett= ys" <[mailto:jg@freedesktop.org] jg@freedesktop.org>=0ASent: Friday, Decemb= er 21, 2012 1:14pm=0ATo: "David P Reed" <[mailto:dpreed@reed.com] dpreed@re= ed.com>=0A Cc: "Dave Taht" <[mailto:dave.taht@gmail.com] dave.taht@gmail.co= m>, "bloat-devel" <[mailto:bloat-devel@lists.bufferbloat.net] bloat-devel@l= ists.bufferbloat.net>, "[mailto:cerowrt-devel@lists.bufferbloat.net] cerowr= t-devel@lists.bufferbloat.net" <[mailto:cerowrt-devel@lists.bufferbloat.net= ] cerowrt-devel@lists.bufferbloat.net>, "[mailto:codel@lists.bufferbloat.ne= t] codel@lists.bufferbloat.net" <[mailto:codel@lists.bufferbloat.net] codel= @lists.bufferbloat.net>=0A Subject: Re: [Cerowrt-devel] hardware hacking on= fq_codel in FPGA form at 10GigE=0A=0A=0A=0A=0A=0A=0A=0A=0AOn Fri, Dec 21, = 2012 at 12:48 PM, <[mailto:dpreed@reed.com] dpreed@reed.com> wrote:=0A=0AA= nd if you want an open tablet the ODROID U2 there is pretty nice for <$900 = all in, including an HSPA+ modem w/GPS. I can only imagine what one can d= o with this....=0AOften the modem bits are binary blobs. Any information o= n this front?=0A- Jim=0A=0A =0A-----Original Message-----=0AFrom: "Dave Tah= t" <[mailto:dave.taht@gmail.com] dave.taht@gmail.com>=0ASent: Friday, Decem= ber 21, 2012 7:34am=0A To: [mailto:dpreed@reed.com] dpreed@reed.com=0A Cc: = "Hal Murray" <[mailto:hmurray@megapathdsl.net] hmurray@megapathdsl.net>, "b= loat-devel" <[mailto:bloat-devel@lists.bufferbloat.net] bloat-devel@lists.b= ufferbloat.net>, [mailto:codel@lists.bufferbloat.net] codel@lists.bufferblo= at.net, [mailto:cerowrt-devel@lists.bufferbloat.net] cerowrt-devel@lists.bu= fferbloat.net=0A Subject: Re: [Cerowrt-devel] hardware hacking on fq_codel = in FPGA form at 10GigE=0A=0A=0A=0A=0AIt really is astounding what is going = on in the arm world:=0A=0A[http://www.hardkernel.com/renewal_2011/products/= prdt_info.php?g_code=3DG135341370451] http://www.hardkernel.com/renewal_201= 1/products/prdt_info.php?g_code=3DG135341370451=0A=0A=0A> The wndr3700v4 is= out, and appears to be a good hardware upgrade from=0A> the 3800 series, b= ut it's not supported by openwrt yet.=0A>=0A > I took a look at their GPL s= ource distribution. And yea! it's openwrt.=0A > And boo! it's ancient openw= rt, for example dnsmasq is 2.39 (current is=0A> 2.64), and their kernel is = 2.6.31.=0A>=0A> I think the cpu and ethernet chips tho look a lot better: A= theros=0A > AR9344+ AR9580(5GHz)+AR9344(2.4GHz). It's my hope these do ipv6= =0A > better.=0A>=0A> I have looked at similar products from buffalo and ot= hers, would like=0A> external antennas and a good switch this time around. = Thoughts?=0A >=0A> --=0A> Dave T=C3=A4ht=0A>=0A> Fixing bufferbloat with ce= rowrt: [http://www.teklibre.com/cerowrt/subscribe.html] http://www.teklibre= .com/cerowrt/subscribe.html=0A=0A=0A=0A=0A-- =0ADave T=C3=A4ht=0A=0AFixing = bufferbloat with cerowrt: [http://www.teklibre.com/cerowrt/subscribe.html] = http://www.teklibre.com/cerowrt/subscribe.html=0A=0A_______________________= ________________________=0A Bloat-devel mailing list=0A[mailto:Bloat-devel@= lists.bufferbloat.net] Bloat-devel@lists.bufferbloat.net=0A[https://lists.b= ufferbloat.net/listinfo/bloat-devel] https://lists.bufferbloat.net/listinfo= /bloat-devel=0A=0A=0A ------=_20121221143404000000_46376 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable

= Understand about the chip vendors.  It is the case, though, that syste= m vendors often have NDA's with the chip vendors ...  and they are the= immediate customer of the chip vendors, so have some leverage to get their= help resolving the issues.

=0A

 =0A

I think the first step is doing the r= ight measurements, to show that the chip is a bottleneck (as you are doing = with Lantiq).

=0A

 

=0A

-----Original Message-----
From: "Jim Gettys"= <jg@freedesktop.org>
Sent: Friday, December 21, 2012 1:45pm
To: "David P Reed" <dpreed@reed.com>
Cc: "Dave Taht" <dave.= taht@gmail.com>, "bloat-devel" <bloat-devel@lists.bufferbloat.net>= , "cerowrt-devel@lists.bufferbloat.net" <cerowrt-devel@lists.bufferbloat= .net>, "codel@lists.bufferbloat.net" <codel@lists.bufferbloat.net>= , blogic@openwrt.org
Subject: Re: [Cerowrt-devel] hardware hacking on = fq_codel in FPGA form at 10GigE

=0A
=0A

=0A


=0A
On Fri, Dec 21, 2012 at 1:25 PM, <dpreed@re= ed.com> wrote:
=0A
<= span style=3D"font-family: times new roman;">=0A

I'm sure the HSPA+ modem is a binary blob. = ; For regulatory reasons, if nothing else...

=0A

 

=0A

When I said open, I did not mean the ultimate ideal= .  Hardware devices with firmware are often sold as black boxes.

= =0A

 

=0A

Fixing the bloat in HSPA= + and LTE systems will require organizations like Ericsson, ALU and Cisco a= nd Huawei to get their acts together and fix their own bugs.

=0A<= /blockquote>=0A
Actually, this needs probably/certainly needs help from= the chip vendors as well as the system vendors.  The system vendors (= such as ALU) often get their chips and drivers from those chip vendors.=0A
=0A
I haven't looked into the HSPA+ or LTE systems; I a= m looking into DSL.  In the DSL concrete example, Broadcom controls th= e bulk of the DSL chip market, and (at least on the client home router end)= provides a binary blob Linux driver based on an antique Linux release (2.6= .26, IIRC). The hardware itself did not look like it had bad buffering, but= the drivers....
=0A
I've been working with blogic (John Crispin) = on debloating the Lantiq driver, which we are trying to test; since it is e= ntirely open source.  In the Lantiq DSL driver case, the hardware spro= uted exactly the same sort of transmit/receive rings found in ethernet driv= ers, with the corresponding bufferbloat due to device drivers being totally= stupid.  Unfortunately, BQL is ethernet only, and right now doesn't w= ork properly at low bandwidths.
=0A
When John cut the transmit rin= g down to 2 packets, the latency dropped (at the bandwidth he was testing a= t, .5Mbps) to around 60ms running fq_codel, about what you would expect giv= en the remaining buffers; we have to automate the control of the transmit r= ing properly (since we probably do need some help when running VDSL at 100M= bps).
=0A
I suspect/expect we'll find similar "features" when we d= ig into LTE and HSPA+, though that is not certain.
=0A
Now, back t= o my saga of my recalcitrant water well, after which I can go back to testi= ng DSL the next trip to New Jersey (and flush the toilets and have heat; I = have geothermal heat...).
=0A
Sigh,
=0A
- Jim
=0A=0A
=0A
=0A

 

=0A

-----Original Message-----
From: "Jim Gettys" <jg@freedesktop.org>Sent: Friday, December 21, 2012 1:14pm
To: "David P Reed" <dpreed@reed.com>
Cc: "Dave Taht" <dave.taht@gmail.com>, "bloat-devel" <bloat-devel@lists.bufferblo= at.net>, "cerowrt-devel@lists.bufferbloat.net" <cerowrt-devel@lis= ts.bufferbloat.net>, "codel@lists.bufferbloat.net" <codel@lists.bufferbloat.net>
Subject: Re: [Cerowrt-devel] hardware hacking on fq_codel in F= PGA form at 10GigE

=0A
=0A

=0A

=0A
On Fri, De= c 21, 2012 at 12:48 PM, <dpreed@reed.com> wrote:
=0A=0A

And if you wa= nt an open tablet the ODROID U2 there is pretty nice for <$900 all in, i= ncluding an HSPA+ modem w/GPS.   I can only imagine what one can = do with this....

=0A
=0A
Often the modem bits are= binary blobs.  Any information on this front?
=0A
- Jim=0A
=0A
=0A

 

=0A

----= -Original Message-----
From: "Dave Taht" <dave.taht@gmail.com>
Sent: Frid= ay, December 21, 2012 7:34am
To: dpreed@reed.com
Cc: "Hal Murray" <hmurray@megapathdsl.net>, "bloat-devel" <bloat-devel@lists.bufferbloat.net>, codel@lists.bufferbl= oat.net, cerowrt-devel@lists.bufferbloat.net
Subject: Re: [Cer= owrt-devel] hardware hacking on fq_codel in FPGA form at 10GigE

=

=0A
=0A
=0A
=0A

It really is astounding what is going on in the arm world:

http://www.hardkernel.com/re= newal_2011/products/prdt_info.php?g_code=3DG135341370451

=0A
=0A
> The wndr3700v4 is out, and appears to be a good hardware upgrade fr= om
> the 3800 series, but it's not supported by openwrt yet.
&= gt;
> I took a look at their GPL source distribution. And yea! it'= s openwrt.
> And boo! it's ancient openwrt, for example dnsmasq is= 2.39 (current is
> 2.64), and their kernel is 2.6.31.
>> I think the cpu and ethernet chips tho look a lot better: Atheros > AR9344+ AR9580(5GHz)+AR9344(2.4GHz). It's my hope these do ipv6 > better.
>
> I have looked at similar products fro= m buffalo and others, would like
> external antennas and a good swi= tch this time around. Thoughts?
>
> --
> Dave T=C3= =A4ht
>
> Fixing bufferbloat with cerowrt: http://www.tek= libre.com/cerowrt/subscribe.html=0A
=0A



--
Dave= T=C3=A4ht

Fixing bufferbloat with cerowrt: http://www.teklibr= e.com/cerowrt/subscribe.html=0A
=0A
=0A
=0A
__________= _____________________________________
Bloat-devel mailing list
<= a href=3D"mailto:Bloat-devel@lists.bufferbloat.net" target=3D"_blank">Bloat= -devel@lists.bufferbloat.net
https://lists.bufferbloat.net/l= istinfo/bloat-devel=0A
=0A

=0A
=0A
=0A=0A
=0A
=0A
=0A=0A=0A=0A=0A ------=_20121221143404000000_46376--