From owner-freebsd-stable Wed Dec 26 15:45:11 2001 Delivered-To: freebsd-stable@freebsd.org Received: from gscamnlm03.wr.usgs.gov (gscamnlm03.wr.usgs.gov [130.118.4.113]) by hub.freebsd.org (Postfix) with ESMTP id 43DDA37B416; Wed, 26 Dec 2001 15:44:58 -0800 (PST) To: ulf@Alameda.net Cc: Allen Landsidel , freebsd-stable@FreeBSD.ORG, owner-freebsd-stable@FreeBSD.ORG, "Peter/Los Angeles, CA" , Sam Drinkard , sthaug@nethelp.no Subject: Re: 4.5 PRERELEASE - Call for testing MIME-Version: 1.0 X-Mailer: Lotus Notes Release 5.0.8 June 18, 2001 Message-ID: From: "Robert L Sowders" Date: Wed, 26 Dec 2001 15:44:50 -0800 X-MIMETrack: Serialize by Router on gscamnlm03/SERVER/USGS/DOI(Release 5.0.8 |June 18, 2001) at 12/26/2001 03:44:58 PM, Serialize complete at 12/26/2001 03:44:58 PM Content-Type: multipart/alternative; boundary="=_alternative 0082731E88256B2E_=" Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG This is a multipart message in MIME format. --=_alternative 0082731E88256B2E_= Content-Type: text/plain; charset="us-ascii" This should be moved to chat, hardware, hackers, you pick it. Regardless of what the standard says, it has been my experience that if you have a connection that fails to autoneg to full duplex, then forcing the freebsd nic to 100 full will result in the expected behavior of a 100 full connection. During the negotiation phase the nic says I can only do 100 full and the switch responds with 100 full. Recent experience with HP-Procurve and NBase Mega switches confirmed that they continually refused to autoneg 100 full until the nics were forced to 100 full. Subsequently turning the nics back to autoneg produced 100 full connections. Why this occurred is beyond me, but after three days of testing that was the fix. It may be that this is not the experience of the other readers of this thread. That's fine, all I'm saying is, if you're having trouble with full duplex then try forcing the nic to full and see what happens, it works for me. This really should be moved off stable now. Ulf Zimmermann Sent by: owner-freebsd-stable@FreeBSD.ORG 12/26/2001 02:29 PM Please respond to ulf To: "Peter/Los Angeles, CA" cc: Sam Drinkard , Allen Landsidel , sthaug@nethelp.no, freebsd-stable@FreeBSD.ORG Subject: Re: 4.5 PRERELEASE - Call for testing On Wed, Dec 26, 2001 at 02:22:15PM -0800, Peter/Los Angeles, CA wrote: > Please correct me if I'm wrong, but isn't it kind of strange that > auto-sensing/auto-negotiating must be enabled on both sides for the feature > to work a bit strange? > > At home, I have a Netgear FS116, a 16-Port unmanaged switch. It is > auto-sensing/full-duplex 10/100Mbits/sec switch. Therefore, we cannot > control how it will behave. > > On the other hand, I have network cards on my computer which I can set to > full/half/auto/10/100, whatever combination I like, and yet, the switch will > continue to work. > > What I'm getting at is that just because one end is not set to > auto-negotiate/auto-sense that there will be no communication at all. Say, > that one end is set manually, and the other end is automatic. The automatic > end will set itself to the parameters of the one that is manually set. This > is how my network works. Thus, I don't believe that both ends, need to be > set the same way in order to work in this scenario. > > The automatic will automatically negotiate/auto-sense to whatever can't be > changed, as a result, they work. Or is this even what you folks are arguing > about? I do not know if this is written in the standard, but I have seen devices where if you turn off AutoNeg, they won't respond at all to the packets. But I have also seens devices which will still respond, but just to their fixed setting. I personal prefer it that if I turn off autoneg, it won't do anything, because that gives you a fixed point. Any production system I always set switch and host to full 100. -- Regards, Ulf. --------------------------------------------------------------------- Ulf Zimmermann, 1525 Pacific Ave., Alameda, CA-94501, #: 510-865-0204 You can find my resume at: http://seven.Alameda.net/~ulf/resume.html To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message --=_alternative 0082731E88256B2E_= Content-Type: text/html; charset="us-ascii"
This should be moved to chat, hardware, hackers, you pick it.

Regardless of what the standard says, it has been my experience that if you have a connection that fails to autoneg to full duplex, then forcing the freebsd nic to 100 full will result in the expected behavior of a 100 full connection.  During the negotiation phase the nic says I can only do 100 full and the switch responds with 100 full.

Recent experience with HP-Procurve and NBase Mega switches confirmed that they continually refused to autoneg 100 full until the nics were forced to 100 full.  Subsequently turning the nics back to autoneg produced 100 full connections.  Why this occurred is beyond me, but after three days of testing that was the fix.

It may be that this is not the experience of the other readers of this thread.  That's fine, all I'm saying is, if you're having trouble with full duplex then try forcing the nic to full and see what happens, it works for me.

This really should be moved off stable now.



Ulf Zimmermann <ulf@Alameda.net>
Sent by: owner-freebsd-stable@FreeBSD.ORG

12/26/2001 02:29 PM
Please respond to ulf

       
        To:        "Peter/Los Angeles, CA" <peter@haloflightleader.net>
        cc:        Sam Drinkard <sam@wa4phy.net>, Allen Landsidel <all@biosys.net>, sthaug@nethelp.no, freebsd-stable@FreeBSD.ORG
        Subject:        Re: 4.5 PRERELEASE - Call for testing


On Wed, Dec 26, 2001 at 02:22:15PM -0800, Peter/Los Angeles, CA wrote:
> Please correct me if I'm wrong, but isn't it kind of strange that
> auto-sensing/auto-negotiating must be enabled on both sides for the feature
> to work a bit strange?
>
> At home, I have a Netgear FS116, a 16-Port unmanaged switch.  It is
> auto-sensing/full-duplex 10/100Mbits/sec switch.  Therefore, we cannot
> control how it will behave.
>
> On the other hand, I have network cards on my computer which I can set to
> full/half/auto/10/100, whatever combination I like, and yet, the switch will
> continue to work.
>
> What I'm getting at is that just because one end is not set to
> auto-negotiate/auto-sense that there will be no communication at all.  Say,
> that one end is set manually, and the other end is automatic.  The automatic
> end will set itself to the parameters of the one that is manually set.  This
> is how my network works.  Thus, I don't believe that both ends, need to be
> set the same way in order to work in this scenario.
>
> The automatic will automatically negotiate/auto-sense to whatever can't be
> changed, as a result, they work.  Or is this even what you folks are arguing
> about?

I do not know if this is written in the standard, but I have seen devices
where if you turn off AutoNeg, they won't respond at all to the packets.
But I have also seens devices which will still respond, but just to their
fixed setting. I personal prefer it that if I turn off autoneg, it won't
do anything, because that gives you a fixed point. Any production system
I always set switch and host to full 100.

--
Regards, Ulf.

---------------------------------------------------------------------
Ulf Zimmermann, 1525 Pacific Ave., Alameda, CA-94501, #: 510-865-0204
You can find my resume at: http://seven.Alameda.net/~ulf/resume.html

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message

--=_alternative 0082731E88256B2E_=-- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message