From owner-freebsd-net@FreeBSD.ORG Thu May 18 22:42:24 2006 Return-Path: X-Original-To: freebsd-net@freebsd.org Delivered-To: freebsd-net@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id CB44016A404 for ; Thu, 18 May 2006 22:42:24 +0000 (UTC) (envelope-from jfvogel@gmail.com) Received: from py-out-1112.google.com (py-out-1112.google.com [64.233.166.178]) by mx1.FreeBSD.org (Postfix) with ESMTP id 6429243D46 for ; Thu, 18 May 2006 22:42:24 +0000 (GMT) (envelope-from jfvogel@gmail.com) Received: by py-out-1112.google.com with SMTP id f28so697332pyf for ; Thu, 18 May 2006 15:42:23 -0700 (PDT) DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:to:subject:mime-version:content-type:content-transfer-encoding:content-disposition; b=agWeBOUuRNLiKZQya1uqjWEBkSPMbG1we0wA9svy3ygSYOV3FHDGvOPHv4Q/V+r76jqpjtMi7xtRqh8eVNx40QaZptdfEF/BclrdwCBCfv0YxrkOPEwmIAk4vcGxdYbv5W3alitjFn4JcriWchFFlqzkVuwGjh2qXIBejiOP9hE= Received: by 10.35.36.13 with SMTP id o13mr1108982pyj; Thu, 18 May 2006 15:42:23 -0700 (PDT) Received: by 10.35.29.15 with HTTP; Thu, 18 May 2006 15:42:23 -0700 (PDT) Message-ID: <2a41acea0605181542q2459ebdr5d5815b242ad8fed@mail.gmail.com> Date: Thu, 18 May 2006 15:42:23 -0700 From: "Jack Vogel" To: freebsd-net@freebsd.org MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: quoted-printable Content-Disposition: inline Subject: Problem with iterative load/unload of em 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: Thu, 18 May 2006 22:42:24 -0000 Our internal test group has run into a problem, I have witnessed it, but not had time to pursue it. I was wondering if this has been previously observed, and if anyone has any thoughts. What they do is run a script that runs 100 passes at loading, bringing up and configuring the driver, then bringing down and unloading. It does this as fast as the script will run. At some random point in the loop it will cause a crash, I do not have a stack trace in front of me now to include. I am pretty sure the stack varies from instance to instance. This has been observed on both Intel versions of the em driver and community based versions, and both on 5.4, 6.0, and 6.1. If sleeps are inserted into the script the problem will not happen, so its timing sensitive. Its sort of an artificial test, but this apparently didnt happen on 4.11, also reported to not happen with a Broadcom. I've looked at the detach path and so far nothing real obvious jumps out at me. I am too busy to look at this right now, but any ideas or suggestions are welcome. Jack