From owner-freebsd-sparc64@FreeBSD.ORG Thu Oct 14 13:19:30 2004 Return-Path: Delivered-To: freebsd-sparc64@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id CB3D416A4CE for ; Thu, 14 Oct 2004 13:19:30 +0000 (GMT) Received: from electra.cse.Buffalo.EDU (electra.cse.Buffalo.EDU [128.205.32.2]) by mx1.FreeBSD.org (Postfix) with ESMTP id 6885F43D62 for ; Thu, 14 Oct 2004 13:19:30 +0000 (GMT) (envelope-from kensmith@cse.Buffalo.EDU) Received: from electra.cse.Buffalo.EDU (kensmith@localhost [127.0.0.1]) i9EDJTIo004799; Thu, 14 Oct 2004 09:19:29 -0400 (EDT) Received: (from kensmith@localhost) by electra.cse.Buffalo.EDU (8.12.10/8.12.9/Submit) id i9EDJTLe004798; Thu, 14 Oct 2004 09:19:29 -0400 (EDT) Date: Thu, 14 Oct 2004 09:19:29 -0400 From: Ken Smith To: Herve Boulouis Message-ID: <20041014131929.GD3528@electra.cse.Buffalo.EDU> References: <20041014114647.A69222@ra.aabs> <20041014125330.GB3528@electra.cse.Buffalo.EDU> <20041014150624.B69222@ra.aabs> Mime-Version: 1.0 Content-Type: text/plain; charset=unknown-8bit Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: <20041014150624.B69222@ra.aabs> User-Agent: Mutt/1.4.1i cc: Ken Smith cc: freebsd-sparc64@freebsd.org Subject: Re: Strange timing problems with BETA7 X-BeenThere: freebsd-sparc64@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Porting FreeBSD to the Sparc List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 14 Oct 2004 13:19:30 -0000 On Thu, Oct 14, 2004 at 03:06:24PM +0200, Herve Boulouis wrote: > Le 14/10/2004 à 08:53, Ken Smith a écrit: > > > > initializing counter-timer > > Timecounter "counter-timer" frequency 1000000 Hz quality 100 > > Timecounters tick every 10.000 msec > > > > I'm worried that you might not have 'counter-timer'. > > Here is the relevant part of my dmesg : > > FreeBSD 5.3-BETA7 #5: Tue Oct 5 11:50:48 CEST 2004 > root@netra:/usr/obj/usr/src/sys/NETRA > WARNING: MPSAFE network stack disabled, expect reduced performance. > Timecounter "tick" frequency 296036546 Hz quality 0 > real memory = 805306368 (768 MB) > avail memory = 772554752 (736 MB) > cpu0: Sun Microsystems UltraSparc-II Processor (296.04 MHz CPU) > cpu1: Sun Microsystems UltraSparc-II Processor (296.04 MHz CPU) > FreeBSD/SMP: Multiprocessor System Detected: 2 CPUs > nexus0: > pcib0: on nexus0 > pcib0: Psycho, impl 0, version 4, ign 0x7c0, bus B > pcib0: [FAST] > pcib0: [GIANT-LOCKED] > pcib0: [FAST] > pcib0: [GIANT-LOCKED] > pcib0: [FAST] > initializing counter-timer > Timecounter "counter-timer" frequency 1000000 Hz quality 0 > ... > Timecounters tick every 10.000 msec > > Do not hesitate to ask if you need info or want me to test patches, > this box is not in production. > Do you follow along with cvsup, or by doing full installs from ISO's? This looks like the configuration before the final set of timer related patches, I think the last chunk went in late last week. If you cvsup now you should get everything. Your case may actually be one of the ones that this set of patches was supposed to help - if the system chooses to use your CPU's timer as its primary timecounter it might cause problems because we don't even attempt to keep the clocks on the individual CPUs in sync. The last set of patches sets it up so that on MP systems the "quality" of the CPU's timers gets demoted so that other counters should wind up being chosen as the primary timecounter (in this case "counter-timer" should always be chosen). -- Ken Smith - From there to here, from here to | kensmith@cse.buffalo.edu there, funny things are everywhere. | - Theodore Geisel |