From owner-freebsd-net@FreeBSD.ORG Sat Jan 2 19:04:42 2010 Return-Path: Delivered-To: freebsd-net@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 07802106566B for ; Sat, 2 Jan 2010 19:04:42 +0000 (UTC) (envelope-from barney_cordoba@yahoo.com) Received: from web63903.mail.re1.yahoo.com (web63903.mail.re1.yahoo.com [69.147.97.118]) by mx1.freebsd.org (Postfix) with SMTP id B7C158FC15 for ; Sat, 2 Jan 2010 19:04:41 +0000 (UTC) Received: (qmail 59701 invoked by uid 60001); 2 Jan 2010 19:04:33 -0000 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=yahoo.com; s=s1024; t=1262459073; bh=ON7N/p15nvTP4SJpJjTABn05Z7ZHcpfkQwvwrUENaIc=; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Subject:To:Cc:In-Reply-To:MIME-Version:Content-Type:Content-Transfer-Encoding; b=sCdwYuMn5Xz/VcMatDCX7Id2meFElIiG841u+PNK1w5x3sv+s4LXFUKKmST9wnWikFribqpK4cnu85b/9ILE+eAujtZqd7SDwTUwJly1UYQLPXHd/GGDME9LCc9Bbx9OmQo4E+/lT0kERGfY6pO1k2KT3ESW5YtdHB/NmueC41A= DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Subject:To:Cc:In-Reply-To:MIME-Version:Content-Type:Content-Transfer-Encoding; b=kGeGDVZHNd0LphLGFp8w5g2lb9CIowVHjgJEOQ+v6Ah9YgzHKkp0ZjGxJY10pWo7r3gqldQ/czIMhyaNDZiFlrPUSs4TJeKD4xXDm1ikq53fhkWRGf2uan75FtiVWQzIw91+Bpp94qejlteGdRd5rEO8LjxD1V7nAtTKZ9e+3PM=; Message-ID: <422185.59153.qm@web63903.mail.re1.yahoo.com> X-YMail-OSG: tefQcxMVM1kNA_i6A9lc52dYcldYqYc4pzJrnROtsgyJBHuVNNT8hYmBt_348WDVEc87.SreK8wFRlybFWx65ivMoh7Um3p0a9pYpGTDYSS0cbsKrmc_ZA.FRGBLIqQJD1O5Nyu.Tt6YEHOq4dPCdD4mbfaoM.gHcuP5Yd1WcI0XJEfMzohlBi2Lrymp2NyoPX1VIxZjhjB35.byjb3O2ktNaIBmK7iRO1.oZo51mcHlpEhxkzpks1HZiDpVZTQ6m6wjjfeQzwWWzXO4tJBamPY5d314rl8RLZ3KaR_s7wPyZSERtg-- Received: from [98.203.21.152] by web63903.mail.re1.yahoo.com via HTTP; Sat, 02 Jan 2010 11:04:33 PST X-Mailer: YahooMailClassic/9.0.20 YahooMailWebService/0.8.100.260964 Date: Sat, 2 Jan 2010 11:04:33 -0800 (PST) From: Barney Cordoba To: George Sanders , Jack Vogel In-Reply-To: <2a41acea0912111851x598a4f5bsccf013e0c8107743@mail.gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable Cc: freebsd-net@freebsd.org Subject: Re: 10gigE link with FreeBSD hosts ? X-BeenThere: freebsd-net@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Networking and TCP/IP with FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 02 Jan 2010 19:04:42 -0000 =0A=0A--- On Fri, 12/11/09, Jack Vogel wrote:=0A=0A> Fr= om: Jack Vogel =0A> Subject: Re: 10gigE link with FreeBS= D hosts ?=0A> To: "George Sanders" =0A> Cc: freebsd-n= et@freebsd.org=0A> Date: Friday, December 11, 2009, 9:51 PM=0A> Can't get 1= 0 gigabit thru a PCI-X=0A> slot, it doesnt have the bandwidth, you=0A> need= PCI Express, in=0A> fact you will find that lots of factors will come into= play=0A> and might keep=0A> you from achieving=0A> optimal thruput.=0A> = =0A> For a two port adapter you will need an 8x PCIE slot,=0A> preferably G= en 2, and=0A> if all three systems=0A> don't have that its gonna slow every= one down.=0A> =0A> Further, once you involve packet forwarding it gets even= =0A> more demanding.=0A> =0A> Trying to do it 'on the cheap' and you might = get 3 or 4 Gb,=0A> maybe even worse=0A> depending.=0A> =0A> Takes quality t= o get real performance, sorry :)=0A> =0A> Jack=0A> =0A> =0A> On Fri, Dec 11= , 2009 at 2:30 PM, George Sanders wrote:=0A> =0A> >= =0A> >=0A> > I am seeing Sun Microsystems branded 10gigE adaptors=0A> (pci-= x) on ebay=0A> > nowadays for $500.=0A> >=0A> > So ... let's say I took som= e plain-jane, modern-ish PC=0A> (intel core, or c2d,=0A> > pci-x, etc.) and= set up this topology:=0A> >=0A> >=0A> > a <-- cable --> b=A0 <-- cable -->= c=0A> >=0A> > and on each of (a) and (c) put one 10gigE adaptor, and=0A> o= n (b) put two=0A> > 10gigE adaptors, and set up bridging on (b)=0A> >=0A> >= Would it be reasonable to then expect to transfer data=0A> from (a) to (c)= at=0A> > a, roughly, 10 gigabit/s rate ?=0A> >=0A> > If not, what is the l= imiting factor ?=0A> >=0A> > The idea is that I would use a commodity PC + = multiple=0A> pci-X slots as a=0A> > "poor mans" 10gigE switch ... I can't a= fford an actual=0A> 10gigE switch, but I=0A> > don't need that many ports a= nyway.=0A> >=0A> > Any comments or field reports would be appreciated.=0A> = >=0A> >=0A=0ASupermicro makes some cheap cards that are under $500. Part# i= s=0AAOC-STG-i2 and they work with FreeBSD. You should be able to find=0Athe= m for around $450, so if you just need a 2 port switch its fairly=0Ainexpen= sive.=0A=0AThe problem is that the FreeBSD ixgbe driver can't do close to 1= 0Gb/s, =0Aso its not a very good switch. =0A=0ABarney=0A=0A=0A From owner-freebsd-net@FreeBSD.ORG Sat Jan 2 19:30:04 2010 Return-Path: Delivered-To: freebsd-net@hub.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 0C0671065670 for ; Sat, 2 Jan 2010 19:30:04 +0000 (UTC) (envelope-from gnats@FreeBSD.org) Received: from freefall.freebsd.org (freefall.freebsd.org [IPv6:2001:4f8:fff6::28]) by mx1.freebsd.org (Postfix) with ESMTP id D60658FC08 for ; Sat, 2 Jan 2010 19:30:03 +0000 (UTC) Received: from freefall.freebsd.org (localhost [127.0.0.1]) by freefall.freebsd.org (8.14.3/8.14.3) with ESMTP id o02JU3Xa011773 for ; Sat, 2 Jan 2010 19:30:03 GMT (envelope-from gnats@freefall.freebsd.org) Received: (from gnats@localhost) by freefall.freebsd.org (8.14.3/8.14.3/Submit) id o02JU3qL011770; Sat, 2 Jan 2010 19:30:03 GMT (envelope-from gnats) Date: Sat, 2 Jan 2010 19:30:03 GMT Message-Id: <201001021930.o02JU3qL011770@freefall.freebsd.org> To: freebsd-net@FreeBSD.org From: Dalibor Gudzic Cc: Subject: Re: kern/140796: [ath] [panic] privileged instruction fault X-BeenThere: freebsd-net@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list Reply-To: Dalibor Gudzic List-Id: Networking and TCP/IP with FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 02 Jan 2010 19:30:04 -0000 The following reply was made to PR kern/140796; it has been noted by GNATS. From: Dalibor Gudzic To: bug-followup@freebsd.org Cc: Subject: Re: kern/140796: [ath] [panic] privileged instruction fault Date: Sat, 2 Jan 2010 20:29:06 +0100 On Wed, Dec 30, 2009 at 4:59 PM, Rui Paulo wrote: > Are you sure this card works properly? Can you try with a different card? > -- > Rui Paulo > Sorry for not replying sooner - holidays. ;) Anyway, the card works properly with Windows XP, and Fedora 10, as I stated in earlier PR: http://www.freebsd.org/cgi/query-pr.cgi?pr=kern/116444 P.S. Damn, forgot to cc the list. From owner-freebsd-net@FreeBSD.ORG Sat Jan 2 19:42:30 2010 Return-Path: Delivered-To: freebsd-net@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id C56EB106568B for ; Sat, 2 Jan 2010 19:42:30 +0000 (UTC) (envelope-from barney_cordoba@yahoo.com) Received: from web63905.mail.re1.yahoo.com (web63905.mail.re1.yahoo.com [69.147.97.120]) by mx1.freebsd.org (Postfix) with SMTP id 729238FC15 for ; Sat, 2 Jan 2010 19:42:30 +0000 (UTC) Received: (qmail 48986 invoked by uid 60001); 2 Jan 2010 19:42:25 -0000 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=yahoo.com; s=s1024; t=1262461344; bh=9D7o4Laqmxel9bfREa/KZplSVP36bFS3d7ctQLT1JYI=; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Subject:To:Cc:MIME-Version:Content-Type; b=NnXWXgwqR8Bqr0IqZbsQoO4HZEkdKfKiwCjSW4496Xs+/c+NMbCIXi5uTDY/ltx/xJsO48ADwMh3P5ulEjCKW2EgtKgNoHF5AP55s4owmmMw30fWnR6vD+7ZkCiapXExsI0Rrg5BaeLAWw6YextWUpX86dO+BSRxvE/yA1Xm8fM= DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Subject:To:Cc:MIME-Version:Content-Type; b=kq1Y9wU5qZTWJU/2qF2hjPgwrgeSPXu/frqpI+syQee+LhU94XTTjtFvgXQeQKkKhkTX67Qb/U/IG1dAxUOqtu8j8UHEl/uv/MXIE60g9BkYISLvMVvwOtk+4vqBtQUrUTso3mhP72YTwK373QmlhI4KUg43cv+xW5GyrPGWf+A=; Message-ID: <969646.48748.qm@web63905.mail.re1.yahoo.com> X-YMail-OSG: bM.DR0MVM1lf.BXF_nCcaPRv8befPguMBC.FwumyZ3CZn6NffeS.Hh7fTBNVTlGtWrdGaWMFzukRlPY2batK7Efsb_hW9ARR3cflQVInpiPheWcB6WBUyYtJbxx9qTcJ4qXtGJySxN2K4g.RJM7WhIqSds8YnF1IppF8WmByqBdJudqaQzmJIY7uQ_zW28NWPizcM06MneGV5MqJSkmRTvIToxkTSRiF7Ami29zU8ZwRNtqkddNUWA.ZwsmC Received: from [98.203.21.152] by web63905.mail.re1.yahoo.com via HTTP; Sat, 02 Jan 2010 11:42:24 PST X-Mailer: YahooMailClassic/9.0.20 YahooMailWebService/0.8.100.260964 Date: Sat, 2 Jan 2010 11:42:24 -0800 (PST) From: Barney Cordoba To: freebsd-net@freebsd.org MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: jfvogel@gmail.com Subject: igb interrupt moderation X-BeenThere: freebsd-net@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Networking and TCP/IP with FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 02 Jan 2010 19:42:30 -0000 Jack, I'm trying to get some clarification on differences I'm finding between the 82575 and 82576 parts with respect to interrupt moderation. The spec I have for the 82576 (82576_Datasheet_v2p1.pdf) indicates that the ITR algorithm is different than the one used (I don't have one of the secret copies of the 82575 spec). The algorithm shown is interrupts/sec = 1/(2 * 10-6sec x interval) (page 295, Section 7.3.4) which is clearly wrong from practice. I have an 82576 (device id 10C9) if I use the 125d setting in the example get just under 32000 interrupts per second. Clearly your code doesnt implement this, nor do you have different settings for the 82575 and 82576 parts. So I assume that the same formula for the em parts hold for the igb parts, and that the datasheet is wrong? There does seem to be a slight difference. The setting that gets 1000 ints/second on the 82575 generates about 1020 on the 82576. Not a big deal but I wonder why there's a difference? Is the reference clock for these something that may not be fixed and could vary from board to board? Note that both devices are on the same MB. Also, it seems that settings to EITR over 32767 wrap on the 82576 (for example writing 32768 to EITR is the same as writing a 1). So the minimum setting on the 82576 is around 125 ints/second. The 82575 can accept values up the 65535 before wrapping. The 82576 document doesn't have a map of the register that I can find, so Im curious as to whether these observations are something I can assume is true across all parts and motherboards/cards, or is there some implementation variance that will cause these to only apply to the ones I happen to be testing? Thanks, Barney