From owner-freebsd-current@FreeBSD.ORG Thu Mar 3 08:19:25 2005 Return-Path: Delivered-To: freebsd-current@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 0328516A4CE for ; Thu, 3 Mar 2005 08:19:25 +0000 (GMT) Received: from comp.chem.msu.su (comp.chem.msu.su [158.250.32.97]) by mx1.FreeBSD.org (Postfix) with ESMTP id ECFFE43D31 for ; Thu, 3 Mar 2005 08:19:23 +0000 (GMT) (envelope-from yar@comp.chem.msu.su) Received: from comp.chem.msu.su (localhost [127.0.0.1]) by comp.chem.msu.su (8.12.9p2/8.12.9) with ESMTP id j238JFnZ094567; Thu, 3 Mar 2005 11:19:15 +0300 (MSK) (envelope-from yar@comp.chem.msu.su) Received: (from yar@localhost) by comp.chem.msu.su (8.12.9p2/8.12.9/Submit) id j238JEPR094562; Thu, 3 Mar 2005 11:19:14 +0300 (MSK) (envelope-from yar) Date: Thu, 3 Mar 2005 11:19:13 +0300 From: Yar Tikhiy To: Dag-Erling Sm?rgrav Message-ID: <20050303081913.GI92040@comp.chem.msu.su> References: <20050228145210.GG13933@comp.chem.msu.su> <86ekezl5ig.fsf@xps.des.no> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <86ekezl5ig.fsf@xps.des.no> User-Agent: Mutt/1.5.6i X-Mailman-Approved-At: Thu, 03 Mar 2005 12:50:01 +0000 cc: freebsd-current@freebsd.org Subject: Re: CURRENT inside MS Virtual Server 2005 X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 03 Mar 2005 08:19:25 -0000 On Tue, Mar 01, 2005 at 05:01:27PM +0100, Dag-Erling Sm?rgrav wrote: > Yar Tikhiy writes: > > 4.x and 5.x run smoothly in this virtual environment, but CURRENT > > spends 20-30% of CPU time in the "swi4: clock sio" thread. > > Moreover, CPU idle time data look really weird. Neither tweaking > > kern.timecounter.hardware nor disabling ACPI seems to affect that. > > try adding 'kern.hz=100' to /boot/loader.conf. This fixed idle time accounting, but the clock thread still is eating a lot of CPU time. It's not a big deal since its CPU time drops as soon as a real task starts, but the whole picture looks a bit weird. Do you have an idea why the clock thread behaves this way? Thanks! %%% yar@current:~$sysctl kern.clockrate kern.clockrate: { hz = 100, tick = 10000, profhz = 1024, stathz = 128 } yar@current:~$top -S ... CPU states: 0.0% user, 0.0% nice, 2.7% system, 18.9% interrupt, 78.4% idle Mem: 6012K Active, 4860K Inact, 10M Wired, 7920K Buf, 100M Free Swap: 231M Total, 231M Free PID USERNAME PRI NICE SIZE RES STATE TIME WCPU CPU COMMAND 11 root 171 52 0K 8K RUN 6:13 74.17% 74.17% idle 27 root -32 -151 0K 8K *Giant 0:08 17.97% 17.97% swi4: clock sio ... yar@current:~$gzip < /dev/zero > /dev/null & yar@current:~$top -S ... CPU states: 69.1% user, 0.0% nice, 27.2% system, 3.7% interrupt, 0.0% idle Mem: 6668K Active, 5800K Inact, 11M Wired, 8816K Buf, 98M Free Swap: 231M Total, 231M Free PID USERNAME PRI NICE SIZE RES STATE TIME WCPU CPU COMMAND 473 yar 132 0 1532K 1016K RUN 7:00 92.97% 92.97% gzip 27 root -32 -151 0K 8K WAIT 0:18 2.98% 2.98% swi4: clock sio ... %%% -- Yar