From owner-freebsd-stable@FreeBSD.ORG  Mon Nov 10 09:32:08 2003
Return-Path: <owner-freebsd-stable@FreeBSD.ORG>
Delivered-To: freebsd-stable@freebsd.org
Received: by hub.freebsd.org (Postfix, from userid 916)
	id 8A5A116A4CF; Mon, 10 Nov 2003 09:32:08 -0800 (PST)
Date: Mon, 10 Nov 2003 09:32:08 -0800
From: Prafulla Deuskar <pdeuskar@freebsd.org>
To: Doug White <dwhite@gumbysoft.com>
Message-ID: <20031110093208.A74727@hub.freebsd.org>
References: <20031107172535.A35573@hub.freebsd.org>
	<20031107174754.R27719@carver.gumbysoft.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
In-Reply-To: <20031107174754.R27719@carver.gumbysoft.com>;
	from dwhite@gumbysoft.com on Fri, Nov 07, 2003 at 05:52:46PM -0800
X-Operating-System: FreeBSD 4.9-PRERELEASE on an i386
cc: freebsd-stable@freebsd.org
Subject: Re: Panic on -stable with 10 Gigabit driver
X-BeenThere: freebsd-stable@freebsd.org
X-Mailman-Version: 2.1.1
Precedence: list
List-Id: Production branch of FreeBSD source code
	<freebsd-stable.freebsd.org>
List-Unsubscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-stable>,
	<mailto:freebsd-stable-request@freebsd.org?subject=unsubscribe>
List-Archive: <http://lists.freebsd.org/pipermail/freebsd-stable>
List-Post: <mailto:freebsd-stable@freebsd.org>
List-Help: <mailto:freebsd-stable-request@freebsd.org?subject=help>
List-Subscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-stable>,
	<mailto:freebsd-stable-request@freebsd.org?subject=subscribe>
X-List-Received-Date: Mon, 10 Nov 2003 17:32:08 -0000

> > We are having a reproducible panic with our 10 Gigabit driver on -stable
> > under stress. (4.8-RELEASE, 4.9-RELEASE) We have not been able to
> > reproduce the panic on 5.1-RELEASE with same load.
> 
> What revision of uipc_mbuf.c generated this crashdump?  I'm going to guess
> 4.8 RELEASE since it points to an actual line of code.
> 
> I'm going to take a guess and shoot for cluster corruption of some sort.
> The offending line is
> 
> MCLGET(mp, how);

Yup. I am aware that it is probably MCLGET which is causing the problem.
Well I am not sure why it happens other than speculating that it is caused by
stress created by 10Gbe cards.

It doesn't happen if we just stream UDP traffic to the receiver. 

So what is the solution for this issue other than maintaining our own free list?

Thanks,
Prafulla

> 
> Which would explode if the free list is wrecked perhaps.  If you can, load
> up the crashdump and poke around the freelist and see if something got
> spammed perhaps.
>

<-- crash dump output removed -->