From owner-freebsd-hackers@FreeBSD.ORG Thu Sep 18 01:45:56 2003 Return-Path: Delivered-To: freebsd-hackers@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id E5D6216A4B3 for ; Thu, 18 Sep 2003 01:45:56 -0700 (PDT) Received: from cirb503493.alcatel.com.au (c211-28-27-130.belrs2.nsw.optusnet.com.au [211.28.27.130]) by mx1.FreeBSD.org (Postfix) with ESMTP id 95CB943F3F for ; Thu, 18 Sep 2003 01:45:55 -0700 (PDT) (envelope-from PeterJeremy@optushome.com.au) Received: from cirb503493.alcatel.com.au (localhost.alcatel.com.au [127.0.0.1])h8I8jegh019265; Thu, 18 Sep 2003 18:45:40 +1000 (EST) (envelope-from jeremyp@cirb503493.alcatel.com.au) Received: (from jeremyp@localhost) by cirb503493.alcatel.com.au (8.12.8/8.12.8/Submit) id h8I8jctD019264; Thu, 18 Sep 2003 18:45:38 +1000 (EST) Date: Thu, 18 Sep 2003 18:45:38 +1000 From: Peter Jeremy To: Andrew Gallatin Message-ID: <20030918084538.GB19197@cirb503493.alcatel.com.au> References: <16232.56412.745069.248974@grasshopper.cs.duke.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <16232.56412.745069.248974@grasshopper.cs.duke.edu> User-Agent: Mutt/1.4.1i cc: freebsd-hackers@freebsd.org Subject: Re: PCI interrupts passing DMA X-BeenThere: freebsd-hackers@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Technical Discussions relating to FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 18 Sep 2003 08:45:57 -0000 On Wed, Sep 17, 2003 at 06:12:44PM -0400, Andrew Gallatin wrote: >My question is: What the heck could the SMP kernel be doing which >causes the DMA to "complete" faster? My guess is that this is a coherency issue rather than a timing issue. The SMP kernels are far more careful about ensuring consistency between CPUs and this could be fixing/masking the problem That said, any decent disk controller uses DMA and if the UP kernel really did have PCI-bus coherency problems (or massive delays), a significant part of the user-base would be screaming. The other possibility: Is your PCI card doing something wierd? Are you bending one of the PCI specs a bit far? Peter