From owner-freebsd-smp Sun Sep 30 7:39:34 2001 Delivered-To: freebsd-smp@freebsd.org Received: from luke.immure.com (luke.immure.com [207.8.42.74]) by hub.freebsd.org (Postfix) with ESMTP id 231F537B40C for ; Sun, 30 Sep 2001 07:39:31 -0700 (PDT) Received: (from bob@localhost) by luke.immure.com (8.11.2/8.11.5) id f8UEdLh29743; Sun, 30 Sep 2001 09:39:21 -0500 (CDT) (envelope-from bob) Date: Sun, 30 Sep 2001 09:39:21 -0500 From: Bob Willcox To: Hassan Halta Cc: freebsd-smp@FreeBSD.ORG Subject: Re: Zero % values for CPU and WCPU Message-ID: <20010930093921.A28955@luke.immure.com> Reply-To: Bob Willcox References: <20010928012505.X52188-100000@quark.cs.earlham.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <20010928012505.X52188-100000@quark.cs.earlham.edu>; from hassan@cs.earlham.edu on Fri, Sep 28, 2001 at 01:29:24AM -0500 Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org I have this problem on one of my older systems here. It's a dual 400 MHz Xeon using an ASUS XG-DLS mb, running a 4.4-RC version of FreeBSD. This board has always (since I originally assembled it over 2.5 years ago now) done this. I am (still) running the following patch that I got from Tor Egge that seems to "fix" it, though: Index: clock.c =================================================================== RCS file: /usr/cvs/FreeBSD/src/sys/i386/isa/clock.c,v retrieving revision 1.149.2.3 diff -c -r1.149.2.3 clock.c *** clock.c 2001/04/18 23:17:41 1.149.2.3 --- clock.c 2001/09/30 14:36:16 *************** *** 202,207 **** --- 202,211 ---- static void clkintr(struct clockframe frame) { + #if 1 /* Patch from Tor Egge to keep rtc running */ + while (rtcin(RTC_INTR) & RTCIR_PERIOD) + statclock(&frame); + #endif if (timecounter->tc_get_timecount == i8254_get_timecount) { disable_intr(); if (i8254_ticked) This works for me (I've not noticed any ill effects from it) but YMMV. Bob On Fri, Sep 28, 2001 at 01:29:24AM -0500, Hassan Halta wrote: > Hi all, > > I have a weird problem with top after having SMP enabled in the Kernel. > Everything seems to be working just fine, except for stuff like top starts > giving me weird errors with zero % of the CPU usage. I tried to hammer the > CPU with some benchmarks, and see if that makes it get away from zero, but > there was no hope. I even recomplied top, and things seem to be the same. > Just a quick note, when I add "options NCPU=x" it says that NPU is an > unknown command, which is totally weird. > > I am running FreeBSD 4.3 Stable. It is been a problem since I had FreeBSD > 4.2 and I hoped that it would go away with 4.3, but I guess not. > > I hope that I can get some help with this! > > Thanks a lot, > Hassan > > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-smp" in the body of the message -- Bob Willcox Putt's Law: bob@vieo.com Technology is dominated by two types of people: Austin, TX Those who understand what they do not manage. Those who manage what they do not understand. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Tue Oct 2 8:15:48 2001 Delivered-To: freebsd-smp@freebsd.org Received: from scout.networkphysics.com (fw.networkphysics.com [205.158.104.176]) by hub.freebsd.org (Postfix) with ESMTP id 0B74D37B407 for ; Tue, 2 Oct 2001 08:15:40 -0700 (PDT) Received: from gto.networkphysics.com (gto.networkphysics.com [10.1.0.20]) by scout.networkphysics.com (8.11.5/8.11.5) with ESMTP id f9207L557669 for ; Mon, 1 Oct 2001 17:07:22 -0700 (PDT) From: Tai Jin Received: from vigor.fractal.networkphysics.com (xo-gw [205.158.104.176]) by gto.networkphysics.com (8.9.3/8.9.3) with ESMTP id RAA91121 for ; Mon, 1 Oct 2001 17:07:21 -0700 (PDT) Received: (from tjin@localhost) by vigor.fractal.networkphysics.com (8.11.1/8.11.3) id f9206f258608 for freebsd-smp@freebsd.org; Mon, 1 Oct 2001 17:06:41 -0700 (PDT) Date: Mon, 1 Oct 2001 17:06:41 -0700 (PDT) Message-Id: <200110020006.f9206f258608@vigor.fractal.networkphysics.com> To: freebsd-smp@freebsd.org Subject: SMP profiling? Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org Hi, I'm trying to profile an SMP kernel, but the system freezes within two seconds of turning it on. So I tried to find some documentation and came across this comment on the /~fsmp/SMP/TODO.html page: * Fix kernel profiling to be SMP aware and design and implement tools for profiling lock usage and congestion. Does this mean that profiling on SMP kernels does not currently work? I am running a 4.2Release based kernel. thanks...tai To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Oct 3 5:24:42 2001 Delivered-To: freebsd-smp@freebsd.org Received: from seine.ma.ic.ac.uk (seine.ma.ic.ac.uk [155.198.190.53]) by hub.freebsd.org (Postfix) with ESMTP id 0835037B406 for ; Wed, 3 Oct 2001 05:24:34 -0700 (PDT) Received: (from pruess@localhost) by seine.ma.ic.ac.uk (8.11.6/8.11.6) id f93COVX06988; Wed, 3 Oct 2001 13:24:31 +0100 (BST) (envelope-from pruess) From: Gunnar Pruessner Message-Id: <200110031224.f93COVX06988@seine.ma.ic.ac.uk> Subject: significant SMP slowdown To: freebsd-smp@freebsd.org Date: Wed, 3 Oct 2001 13:24:31 +0100 (BST) Cc: Gunnar Pruessner X-Mailer: ELM [version 2.4ME+ PL94b (25)] MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org Hi! Since upgrading from FBSD 3.0.0-19990112-SNAP to FBSD 4.4-RELEASE my system is running horribly slow as soon as I turn on SMP, i.e. the GENERIC kernel runs just fine, but using a kernel which supports SMP (i.e. GENERIC with uncommented SMP and APIC_IO, and I[346]86_CPU commented out) slows down the entire system significantly. The system was running fine for almost two years, but with 4.4 the system even freezed once and another time during compilation of a new kernel an idling emacs crashed with SIGABRT (6) and gensetdefs with SIGSEGV (11). Now it runs with the GENERIC kernel smoothly and without any problems. Please find attached dmesg from the new 4.4-RELEASE and from the old 3.0.0-19990112-SNAP. Please see also a similar message from Chris Faulhaber: http://www.FreeBSD.org/cgi/getmsg.cgi?fetch=33169+38797+/usr/local/www/db/text/2001/freebsd-smp/20010930.freebsd-smp Any suggestions? Best wishes, Gunnar #### dmesg Copyright (c) 1992-2001 The FreeBSD Project. Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994 The Regents of the University of California. All rights reserved. FreeBSD 4.4-RELEASE #0: Sat Sep 29 17:43:16 BST 2001 root@seine.ma.ic.ac.uk:/usr/src/sys/compile/GENERIC_CUSTOM Timecounter "i8254" frequency 1193182 Hz CPU: Pentium/P55C (199.43-MHz 586-class CPU) Origin = "GenuineIntel" Id = 0x543 Stepping = 3 Features=0x8003bf real memory = 201326592 (196608K bytes) avail memory = 191213568 (186732K bytes) Programming 24 pins in IOAPIC #0 IOAPIC #0 intpin 2 -> irq 0 FreeBSD/SMP: Multiprocessor motherboard cpu0 (BSP): apic id: 0, version: 0x00030010, at 0xfee00000 cpu1 (AP): apic id: 1, version: 0x00030010, at 0xfee00000 io0 (APIC): apic id: 2, version: 0x00170011, at 0xfec00000 Preloaded elf kernel "kernel" at 0xc0491000. Intel Pentium detected, installing workaround for F00F bug md0: Malloc disk npx0: on motherboard npx0: INT 16 interface pcib0: on motherboard IOAPIC #0 intpin 19 -> irq 2 IOAPIC #0 intpin 18 -> irq 16 IOAPIC #0 intpin 16 -> irq 18 pci0: on pcib0 isab0: at device 7.0 on pci0 isa0: on isab0 atapci0: port 0xf000-0xf00f at device 7.1 on pci0 ata0: at 0x1f0 irq 14 on atapci0 de0: port 0x6400-0x647f mem 0xe0906000-0xe090607f irq 2 at device 17.0 on pci0 de0: 21041 [10Mb/s] pass 1.1 de0: address 00:80:ad:1e:70:41 fxp0: port 0x6500-0x653f mem 0xe0800000-0xe08fffff,0xe0905000-0xe0905fff irq 16 at device 18.0 on pci0 fxp0: Ethernet address 00:d0:b7:2c:12:30 inphy0: on miibus0 inphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto ahc0: port 0x6000-0x60ff mem 0xe0904000-0xe0904fff irq 17 at device 19.0 on pci0 aic7880: Ultra Wide Channel A, SCSI Id=7, 16/255 SCBs pci0: at 20.0 irq 18 orm0: