From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from smtp96.iad3a.emailsrvr.com (smtp96.iad3a.emailsrvr.com [173.203.187.96]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by lists.bufferbloat.net (Postfix) with ESMTPS id 228F63B29E for ; Thu, 19 May 2022 15:06:51 -0400 (EDT) Received: from app53.wa-webapps.iad3a (relay-webapps.rsapps.net [172.27.255.140]) by smtp37.relay.iad3a.emailsrvr.com (SMTP Server) with ESMTP id ADC203D2C; Thu, 19 May 2022 15:06:50 -0400 (EDT) Received: from deepplum.com (localhost.localdomain [127.0.0.1]) by app53.wa-webapps.iad3a (Postfix) with ESMTP id 99403E032B; Thu, 19 May 2022 15:06:50 -0400 (EDT) Received: by apps.rackspace.com (Authenticated sender: dpreed@deepplum.com, from: dpreed@deepplum.com) with HTTP; Thu, 19 May 2022 15:06:50 -0400 (EDT) X-Auth-ID: dpreed@deepplum.com Date: Thu, 19 May 2022 15:06:50 -0400 (EDT) From: "David P. Reed" To: "Dave Taht" Cc: "Matt Taggart" , "cerowrt-devel" MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_20220519150650000000_56369" Importance: Normal X-Priority: 3 (Normal) X-Type: html In-Reply-To: References: <86499cf6-50eb-3521-4665-ea3ad1d4a48c@lackof.org> X-Client-IP: 209.6.168.128 Message-ID: <1652987210.624723955@apps.rackspace.com> X-Mailer: webmail/19.0.13-RC X-Classification-ID: b440fe9c-a398-4554-b3c3-9224cc15d2d0-1-1 Subject: Re: [Cerowrt-devel] Minirouter with pi compute module 4 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: , X-List-Received-Date: Thu, 19 May 2022 19:06:51 -0000 ------=_20220519150650000000_56369 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable =0ADave -=0AI'm certain that I have non-x86 devices that can forward more t= han a gbit/sec in both directions. If only because I have a very nice syste= m based on LS2160A ARM implementation that I use to forward 10 GigE traffic= in both directions. Though the packets are not tiny. It's not cheap, of co= urse. I am running Fedora server on it in my lab. The LS2160A carrier comes= from a company called SolidRun.=0A =0AI do understand that many of the low= end ARM boards (and maybe even the Pi4 Compute Module) have driver/bus wea= knesses. Have you tried to understand what the issues are? Could be Linux = kernel and driver issues specific to the GigE hardware. Though I'd assume t= hat USB 3.1 would not inherently get in the way.=0A =0AI currently use a Ce= leron board with 2 1 GigE interfaces for my 1Gig cable connection, running = Linux configured according to my preferences. As you say, x86 boards and PC= Ie ethernet interfaces tend to be fine with two 1 GigE ports.=0A =0ASo mayb= e you want to be a bit more specific about what you mean "device on the mar= ket"?=0A =0AOn Thursday, May 19, 2022 2:46pm, "Dave Taht" said:=0A=0A=0A=0A> I am sadly re-discovering there is not a single dev= ice on the market=0A> outside the x86 universe that can actually forward a = gbit in both=0A> directions at the same time.=0A> =0A> =0A> On Thu, May 19,= 2022 at 1:36 PM Matt Taggart wrote:=0A> >=0A> > This loo= ks like an interesting router candidate=0A> >=0A> >=0A> https://www.seeedst= udio.com/Dual-GbE-Carrier-Board-with-4GB-RAM-32GB-eMMC-RPi-CM4-Case-p-5029.= html=0A> >=0A> > Description says:=0A> > * one NIC is Broadcom BCM54210PE (= from the CM4)=0A> > * the other is "Microchip's LAN7800" behind usb3=0A> > = * 2 additional usb3 ports=0A> > * the usb3 uses the CM4's PCIe 2.0 x1 (500M= B/s)=0A> > * wifi/BLE is the CM4's onboard, I think "Cypress CYW43455"?=0A>= >=0A> > It sort of reminds me of the Espressobin device from a few years b= ack,=0A> > but much faster and the pi has a much larger installed base, bet= ter=0A> > support, etc.=0A> >=0A> > --=0A> > Matt Taggart=0A> > matt@lackof= .org=0A> > _______________________________________________=0A> > Cerowrt-de= vel mailing list=0A> > Cerowrt-devel@lists.bufferbloat.net=0A> > https://li= sts.bufferbloat.net/listinfo/cerowrt-devel=0A> =0A> =0A> =0A> --=0A> FQ Wor= ld Domination pending: https://blog.cerowrt.org/post/state_of_fq_codel/=0A>= Dave T=C3=A4ht CEO, TekLibre, LLC=0A> ____________________________________= ___________=0A> Cerowrt-devel mailing list=0A> Cerowrt-devel@lists.bufferbl= oat.net=0A> https://lists.bufferbloat.net/listinfo/cerowrt-devel=0A> ------=_20220519150650000000_56369 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable

Dave -

=0A

I'm certain that I have non-x86 devices that can forward more = than a gbit/sec in both directions. If only because I have a very nice syst= em based on LS2160A ARM implementation that I use to forward 10 GigE traffi= c in both directions. Though the packets are not tiny. It's not cheap, of c= ourse. I am running Fedora server on it in my lab. The LS2160A carrier come= s from a company called SolidRun.

=0A

 

=0A<= p style=3D"margin:0;padding:0;font-family: arial; font-size: 10pt; overflow= -wrap: break-word;">I do understand that many of the low end ARM boards (an= d maybe even the Pi4 Compute Module) have driver/bus weaknesses.  Have= you tried to understand what the issues are? Could be Linux kernel and dri= ver issues specific to the GigE hardware. Though I'd assume that USB 3.1 wo= uld not inherently get in the way.

=0A

 

=0A=

I currently use a Celeron board with 2 1 GigE interfac= es for my 1Gig cable connection, running Linux configured according to my p= references. As you say, x86 boards and PCIe ethernet interfaces tend to be = fine with two 1 GigE ports.

=0A

 

=0A

So maybe you want to be a bit more specific about what you me= an "device on the market"?

=0A

 

=0A

On Thursday, May 19, 2022 2:46pm, "Dave Taht" <dave.taht@gm= ail.com> said:

=0A
=0A

> I am sadly re-discovering there is not a single devi= ce on the market
> outside the x86 universe that can actually forwa= rd a gbit in both
> directions at the same time.
>
&g= t;
> On Thu, May 19, 2022 at 1:36 PM Matt Taggart <matt@lackof.= org> wrote:
> >
> > This looks like an interesting= router candidate
> >
> >
> https://www.seeed= studio.com/Dual-GbE-Carrier-Board-with-4GB-RAM-32GB-eMMC-RPi-CM4-Case-p-502= 9.html
> >
> > Description says:
> > * one= NIC is Broadcom BCM54210PE (from the CM4)
> > * the other is "M= icrochip's LAN7800" behind usb3
> > * 2 additional usb3 ports> > * the usb3 uses the CM4's PCIe 2.0 x1 (500MB/s)
> >= * wifi/BLE is the CM4's onboard, I think "Cypress CYW43455"?
> >= ;
> > It sort of reminds me of the Espressobin device from a few= years back,
> > but much faster and the pi has a much larger in= stalled base, better
> > support, etc.
> >
> = > --
> > Matt Taggart
> > matt@lackof.org
>= ; > _______________________________________________
> > Cerow= rt-devel mailing list
> > Cerowrt-devel@lists.bufferbloat.net> > https://lists.bufferbloat.net/listinfo/cerowrt-devel
>=
>
>
> --
> FQ World Domination pending:= https://blog.cerowrt.org/post/state_of_fq_codel/
> Dave T=C3=A4ht = CEO, TekLibre, LLC
> ______________________________________________= _
> Cerowrt-devel mailing list
> Cerowrt-devel@lists.buffer= bloat.net
> https://lists.bufferbloat.net/listinfo/cerowrt-devel>

=0A
 ------=_20220519150650000000_56369--