From owner-freebsd-smp Sun Aug 16 22:51:23 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id WAA21286 for freebsd-smp-outgoing; Sun, 16 Aug 1998 22:51:23 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from hermes.hrz.uni-bielefeld.de (hermes.hrz.uni-bielefeld.de [129.70.4.55]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id WAA21278 for ; Sun, 16 Aug 1998 22:51:17 -0700 (PDT) (envelope-from lkoeller@post.uni-bielefeld.de) Received: from mitch.hrz.uni-bielefeld.de (9134@mitch.hrz.uni-bielefeld.de [129.70.4.17]) by hermes.hrz.uni-bielefeld.de (8.8.6/8.8.6) with ESMTP id HAA24845; Mon, 17 Aug 1998 07:50:57 +0200 (METDST) Received: from localhost by mitch.hrz.uni-bielefeld.de with SMTP (8.8.6/16.2) id FAA27256; Mon, 17 Aug 1998 05:50:29 GMT Message-Id: <199808170550.FAA27256@mitch.hrz.uni-bielefeld.de> X-Mailer: exmh version 2.0.2 2/24/98 From: Lars =?iso-8859-1?Q?K=F6ller?= To: Chuck Robey Cc: freebsd-smp@FreeBSD.ORG Subject: Re: Per processor load? In-reply-to: chuckr's message of Thu, 13 Aug 1998 14:54:39 -0400. X-Face: eCcoCV}FjV*O{6>[1$XP/e%]TJhEw2MF33dFh)^HM7Gfd=[/(4+0a$~ > On Thu, 13 Aug 1998, Lars Köller wrote: > > > -------- > > > > Hi! > > > > I want to exptend my port of xperfmon++ with per processor load > > charts for 3.X. Is it possible to get the number of processors with a > > syscall? And more important, how to get the load per processor? The > > getloadavg function returns only the averaged load! > > For my 2 processor SMP board, sysctl gives: > > hw.machine: i386 > hw.model: Pentium Pro > hw.ncpu: 2 > > I'd guess you want the last one, right? O.k. Thanks! But thats only the first part of the problem! Where can I get processor specific load information? I need to draw a separate load graph for each CPU?? Regards Lars -- E-Mail: | Lars Köller Lars.Koeller@Uni-Bielefeld.DE | UNIX Sysadmin lkoeller@cc.FH-Lippe.DE | Computing Center PGP-key: | University of Bielefeld http://www.nic.surfnet.nl/pgp/pks-toplev.html | Germany ----------- FreeBSD, what else? ---- http://www.freebsd.org ------------- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Mon Aug 17 11:12:57 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id LAA14493 for freebsd-smp-outgoing; Mon, 17 Aug 1998 11:12:57 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from smtp04.primenet.com (smtp04.primenet.com [206.165.6.134]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id LAA14479 for ; Mon, 17 Aug 1998 11:12:54 -0700 (PDT) (envelope-from tlambert@usr08.primenet.com) Received: (from daemon@localhost) by smtp04.primenet.com (8.8.8/8.8.8) id LAA15377; Mon, 17 Aug 1998 11:12:14 -0700 (MST) Received: from usr08.primenet.com(206.165.6.208) via SMTP by smtp04.primenet.com, id smtpd015299; Mon Aug 17 11:12:04 1998 Received: (from tlambert@localhost) by usr08.primenet.com (8.8.5/8.8.5) id LAA06500; Mon, 17 Aug 1998 11:12:00 -0700 (MST) From: Terry Lambert Message-Id: <199808171812.LAA06500@usr08.primenet.com> Subject: Re: Per processor load? To: Lars.Koeller@post.uni-bielefeld.de (Lars =?iso-8859-1?Q?K=F6ller?=) Date: Mon, 17 Aug 1998 18:12:00 +0000 (GMT) Cc: chuckr@glue.umd.edu, freebsd-smp@FreeBSD.ORG In-Reply-To: <199808170550.FAA27256@mitch.hrz.uni-bielefeld.de> from "Lars =?iso-8859-1?Q?K=F6ller?=" at Aug 17, 98 07:50:28 am X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > But thats only the first part of the problem! Where can I get > processor specific load information? I need to draw a separate load > graph for each CPU?? System load is based on number of processees in the ready-to-run state. For a symmetric system, if the load is 1.0 or above, both CPU's should be actively working. I suppose that what you are asking for is a "processor not idle in the case of 1.0 >= load >= 0.0". To get this, you would have to insert counters into the per CPU idle loops, probably using the Appendix H cycle counter before and after the per CPU HLT instruction, subtracting the count at exit of the last HLT from both, and then subtracting the entry from the exit, and dividing to get an "idle ratio". Gathering this type of statistic could be actively harmful to CPU latency coming out of the HLT condition, and could be as high as 10% to 20% of the systems ability to do work. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Mon Aug 17 13:01:51 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id NAA05895 for freebsd-smp-outgoing; Mon, 17 Aug 1998 13:01:51 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from mail.ftf.dk (mail.ftf.dk [129.142.64.2]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id NAA05875 for ; Mon, 17 Aug 1998 13:01:44 -0700 (PDT) (envelope-from regnauld@deepo.prosa.dk) Received: from mail.prosa.dk ([192.168.100.254]) by mail.ftf.dk (8.8.8/8.8.8/gw-ftf-1.0) with ESMTP id WAA11808 for ; Mon, 17 Aug 1998 22:06:18 +0200 (CEST) (envelope-from regnauld@deepo.prosa.dk) Received: from deepo.prosa.dk (deepo.prosa.dk [192.168.100.10]) by mail.prosa.dk (8.8.8/8.8.5/prosa-1.1) with ESMTP id WAA27192 for ; Mon, 17 Aug 1998 22:09:20 +0200 (CEST) Received: (from regnauld@localhost) by deepo.prosa.dk (8.8.8/8.8.5/prosa-1.1) id WAA03440; Mon, 17 Aug 1998 22:00:46 +0200 (CEST) Message-ID: <19980817220046.44173@deepo.prosa.dk> Date: Mon, 17 Aug 1998 22:00:46 +0200 From: Philippe Regnauld To: freebsd-smp@FreeBSD.ORG Subject: PR440fx (Toshiba) + SMP ? Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.88e X-Operating-System: FreeBSD 2.2.6-RELEASE i386 Phone: +45 3336 4148 Address: Ahlefeldtsgade 16, 1359 Copenhagen K, Denmark Organization: PROSA Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org I've just replaced my GA-586DX with a PR440fx board with 2 P6-166, CSS audio and Intel 100Mb fxp. (the BIOS displays Toshiba at boot). - I've upgraded to the latest BIOS - I've recompiled my kernel with I686 (-current from last night) But FreeBSD insists on seeing only one CPU -- I've even tried hardwiring with NCPU=2, to no avail. I've also tried playing with the BIOS SMP settings (1.1 or 1.4) -- no luck... dmesg + mptable follows Copyright (c) 1992-1998 FreeBSD Inc. Copyright (c) 1982, 1986, 1989, 1991, 1993 The Regents of the University of California. All rights reserved. FreeBSD 3.0-CURRENT #0: Mon Aug 17 22:15:45 MET DST 1998 regnauld@tetard.glou.eu.org:/usr/src/sys/compile/TETARD.smp Timecounter "i8254" frequency 1193182 Hz cost 3588 ns CPU: Pentium Pro (686-class CPU) Origin = "GenuineIntel" Id = 0x619 Stepping=9 Features=0xfbff real memory = 67108864 (65536K bytes) avail memory = 62853120 (61380K bytes) FreeBSD/SMP: Multiprocessor motherboard cpu0 (BSP): apic id: 0, version: 0x00040011, at 0xfec08000 io0 (APIC): apic id: 13, version: 0x00170011, at 0xfec00000 Probing for devices on PCI bus 0: chip0: rev 0x02 on pci0.0.0 fxp0: rev 0x02 int a irq 18 on pci0.6. 0 fxp0: Ethernet address 00:a0:c9:49:58:d9 chip1: rev 0x01 on pci0.7.0 ahc0: rev 0x00 int a irq 17 on pci0.9. 0 ahc0: aic7880 Wide Channel, SCSI Id=7, 16/255 SCBs scbus0 at ahc0 bus 0 ahc0: target 0 Tagged Queuing Device sd0 at scbus0 target 0 lun 0 sd0: type 0 fixed SCSI 2 sd0: Direct-Access 4134MB (8467200 512 byte sectors) sd0: with 8205 cyls, 6 heads, and an average 171 sectors/track vga0: rev 0x00 int a irq 18 on pci0.17.0 Probing for devices on the ISA bus: sc0 at 0x60-0x6f irq 1 on motherboard sc0: VGA color <12 virtual consoles, flags=0x0> ed0 at 0x340-0x35f irq 9 on isa ed0: address 00:20:a9:00:4b:bd, type NE2000 (16 bit) sio0 at 0x3f8-0x3ff irq 4 on isa sio0: type 16550A sio1 at 0x2f8-0x2ff irq 3 on isa sio1: type 16550A psm0 not found at 0x60 lpt0 at 0x378-0x37f irq 7 on isa lpt0: Interrupt-driven port lp0: TCP/IP capable interface fdc0 at 0x3f0-0x3f7 irq 6 drq 2 on isa fdc0: FIFO enabled, 8 bytes threshold fd0: 1.44MB 3.5in npx0 on motherboard npx0: INT 16 interface css0 at 0x534 irq 5 drq 1 flags 0x100 on isa snd0: mpu0 not found at 0x330 APIC_IO: routing 8254 via 8259 on pin 0 =============================================================================== MPTable, version 2.0.15 ------------------------------------------------------------------------------- MP Floating Pointer Structure: location: BIOS physical address: 0x000f7ed0 signature: '_MP_' length: 16 bytes version: 1.4 checksum: 0x33 mode: Virtual Wire ------------------------------------------------------------------------------- MP Config Table Header: physical address: 0x000f7ee0 signature: 'PCMP' base table length: 216 version: 1.4 checksum: 0x62 OEM ID: 'INTEL ' Product ID: 'PR440FX ' OEM table pointer: 0x00000000 OEM table size: 0 entry count: 20 local APIC address: 0xfec08000 extended table length: 120 extended table checksum: 15 ------------------------------------------------------------------------------- MP Config Base Table Entries: -- Processors: APIC ID Version State Family Model Step Flags 0 0x11 BSP, usable 6 1 9 0xfbff -- Bus: Bus ID Type 0 PCI 18 ISA -- I/O APICs: APIC ID Version State Address 13 0x11 usable 0xfec00000 -- I/O Ints: Type Polarity Trigger Bus ID IRQ APIC ID PIN# ExtINT active-hi edge 18 0 13 0 INT active-hi edge 18 1 13 1 INT active-hi edge 18 3 13 3 INT active-hi edge 18 4 13 4 INT active-hi edge 18 5 13 5 INT active-hi edge 18 6 13 6 INT active-hi edge 18 7 13 7 INT active-hi edge 18 8 13 8 INT active-hi edge 18 9 13 9 INT active-hi edge 18 10 13 10 INT active-hi edge 18 12 13 12 INT active-hi edge 18 14 13 14 INT active-hi edge 18 15 13 15 INT active-lo level 0 17:A 13 18 INT active-lo level 0 9:A 13 17 INT active-lo level 0 6:A 13 18 -- Local Ints: Type Polarity Trigger Bus ID IRQ APIC ID PIN# ------------------------------------------------------------------------------- MP Config Extended Table Entries: -- bus ID: 0 address type: memory address address base: 0xd4000 address range: 0x4000 -- bus ID: 0 address type: memory address address base: 0xd8000 address range: 0x4000 -- bus ID: 0 address type: memory address address base: 0xdc000 address range: 0x4000 -- bus ID: 0 address type: memory address address base: 0xa0000 address range: 0x20000 -- bus ID: 0 address type: memory address address base: 0x4000000 address range: 0xfc000000 -- bus ID: 0 address type: I/O address address base: 0x0 address range: 0x10000 ------------------------------------------------------------------------------- # SMP kernel config file options: # Required: options SMP # Symmetric MultiProcessor Kernel options APIC_IO # Symmetric (APIC) I/O # Optional (built-in defaults will work in most cases): #options NCPU=1 # number of CPUs #options NBUS=2 # number of busses #options NAPIC=1 # number of IO APICs #options NINTR=24 # number of INTs =============================================================================== -- -[ Philippe Regnauld / sysadmin / regnauld@deepo.prosa.dk / +55.4N +11.3E ]- The Internet is busy. Please try again later. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Mon Aug 17 14:09:13 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id OAA20589 for freebsd-smp-outgoing; Mon, 17 Aug 1998 14:09:13 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from picnic.mat.net (picnic.mat.net [209.118.174.117]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id OAA20580 for ; Mon, 17 Aug 1998 14:09:06 -0700 (PDT) (envelope-from chuckr@glue.umd.edu) Received: from localhost (chuckr@localhost) by picnic.mat.net (8.9.1/8.8.5) with SMTP id PAA20580 for ; Mon, 17 Aug 1998 15:31:34 -0400 (EDT) Date: Mon, 17 Aug 1998 15:31:34 -0400 (EDT) From: Chuck Robey X-Sender: chuckr@picnic.mat.net To: FreeBSD-smp@FreeBSD.ORG Subject: shutting down ONE processor Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org I need to do some stuff with just one processor active. What's the right method of stopping one processor, then restarting it? Can it be done that way at all, or do I need to force one processor at boot (somehow) and then reboot? Thanks. ----------------------------+----------------------------------------------- Chuck Robey | Interests include any kind of voice or data chuckr@glue.umd.edu | communications topic, C programming, and Unix. 213 Lakeside Drive Apt T-1 | Greenbelt, MD 20770 | I run Journey2 and picnic (FreeBSD-current) (301) 220-2114 | and jaunt (NetBSD). ----------------------------+----------------------------------------------- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Mon Aug 17 14:15:26 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id OAA21702 for freebsd-smp-outgoing; Mon, 17 Aug 1998 14:15:26 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from smtp04.primenet.com (smtp04.primenet.com [206.165.6.134]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id OAA21680 for ; Mon, 17 Aug 1998 14:15:17 -0700 (PDT) (envelope-from tlambert@usr06.primenet.com) Received: (from daemon@localhost) by smtp04.primenet.com (8.8.8/8.8.8) id OAA22521; Mon, 17 Aug 1998 14:14:43 -0700 (MST) Received: from usr06.primenet.com(206.165.6.206) via SMTP by smtp04.primenet.com, id smtpd022502; Mon Aug 17 14:14:40 1998 Received: (from tlambert@localhost) by usr06.primenet.com (8.8.5/8.8.5) id OAA08860; Mon, 17 Aug 1998 14:14:39 -0700 (MST) From: Terry Lambert Message-Id: <199808172114.OAA08860@usr06.primenet.com> Subject: Re: PR440fx (Toshiba) + SMP ? To: regnauld@deepo.prosa.dk (Philippe Regnauld) Date: Mon, 17 Aug 1998 21:14:39 +0000 (GMT) Cc: freebsd-smp@FreeBSD.ORG In-Reply-To: <19980817220046.44173@deepo.prosa.dk> from "Philippe Regnauld" at Aug 17, 98 10:00:46 pm X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > But FreeBSD insists on seeing only one CPU -- I've even tried hardwiring > with NCPU=2, to no avail. I've also tried playing with the BIOS > SMP settings (1.1 or 1.4) -- no luck... Use 1.4. [ ... ] > MP Config Base Table Entries: > > -- > Processors: APIC ID Version State Family Model Step Flags > 0 0x11 BSP, usable 6 1 9 0xfbff > -- > Bus: Bus ID Type You have failed to tell your BIOS that you have a second CPU; therefore it has failed to configure the APIC ID on the second CPU, or otherwise identify it in the mptable so that FreeBSD can see it. A correctly configured BIOS will show two processors with different APIC ID's under the "Processors:" section in the mptable. Fix this, and your system should "just work", unless your clock doesn't go through your APIC, in which case you will need to twiddle a bit. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Mon Aug 17 14:41:24 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id OAA26734 for freebsd-smp-outgoing; Mon, 17 Aug 1998 14:41:24 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from feral.com (gw100.feral.com [192.67.166.129]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id OAA26696 for ; Mon, 17 Aug 1998 14:41:09 -0700 (PDT) (envelope-from mjacob@feral.com) Received: from localhost (mjacob@localhost) by feral.com (8.8.6/8.8.6) with SMTP id OAA02219; Mon, 17 Aug 1998 14:40:24 -0700 Date: Mon, 17 Aug 1998 14:40:23 -0700 (PDT) From: Matthew Jacob X-Sender: mjacob@feral-gw To: Philippe Regnauld cc: freebsd-smp@FreeBSD.ORG Subject: Re: PR440fx (Toshiba) + SMP ? In-Reply-To: <19980817220046.44173@deepo.prosa.dk> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Hmm. I have a PR440FX (intel) + 2 x 180Mhz PPros, which work for me... albeit I'm using the CAM branch right now: Copyright (c) 1992-1998 FreeBSD Inc. Copyright (c) 1982, 1986, 1989, 1991, 1993 The Regents of the University of California. All rights reserved. FreeBSD 3.0-CURRENT #10: Mon Aug 10 00:34:54 PDT 1998 mjacob@quarm.feral.com:/freebsd/cam/usr/src/sys/compile/SMP-QUARM Timecounter "i8254" frequency 1193182 Hz cost 3928 ns CPU: Pentium Pro (686-class CPU) Origin = "GenuineIntel" Id = 0x619 Stepping=9 Features=0xfbff real memory = 33554432 (32768K bytes) avail memory = 29958144 (29256K bytes) FreeBSD/SMP: Multiprocessor motherboard cpu0 (BSP): apic id: 0, version: 0x00040011, at 0xfec08000 cpu1 (AP): apic id: 12, version: 0x00040011, at 0xfec08000 io0 (APIC): apic id: 13, version: 0x00170011, at 0xfec00000 Probing for devices on PCI bus 0: chip0: rev 0x02 on pci0.0.0 fxp0: rev 0x02 int a irq 18 on pci0.6. 0 fxp0: Ethernet address 00:a0:c9:60:34:f5 chip1: rev 0x01 on pci0.7.0 ide_pci0: rev 0x00 on pci0.7.1 ahc0: rev 0x00 int a irq 17 on pci0.9.0 ahc0: aic7880 Wide Channel A, SCSI Id=7, 16/255 SCBs ***Qlogic ISP Driver, FreeBSD CAM Version 0.94, Core Version 1.1 isp0: rev 0x05 int a irq 17 on pci0.15.0 isp0: using I/O space register mapping isp0: Board Revision 1040B, loaded F/W Revision 2.10 Probing for devices on the ISA bus: sc0 at 0x60-0x6f irq 1 on motherboard sc0: VGA color <16 virtual consoles, flags=0x0> sio0 at 0x3f8-0x3ff irq 4 flags 0x10 on isa sio0: type 16550A sio1 at 0x2f8-0x2ff irq 3 on isa sio1: type 16550A fdc0 at 0x3f0-0x3f7 irq 6 drq 2 on isa fdc0: FIFO enabled, 8 bytes threshold fd0: 1.44MB 3.5in wdc0 at 0x1f0-0x1f7 irq 14 on isa wdc0: unit 0 (wd0): wd0: 2014MB (4124736 sectors), 4092 cyls, 16 heads, 63 S/T, 512 B/S npx0 on motherboard npx0: INT 16 interface APIC_IO: routing 8254 via 8259 on pin 0 isp0: Bus Reset SMP: AP CPU #1 Launched! sa0 at isp0 bus 0 target 4 lun 0 sa0: Removable Sequential Access SCSI2 device sa0: 5.0MB/s transfers (5.0MHz, offset 8) sa1 at isp0 bus 0 target 5 lun 0 sa1: Removable Sequential Access SCSI2 device sa1: 5.0MB/s transfers (5.0MHz, offset 8) da0 at ahc0 bus 0 target 1 lun 0 da0: Fixed Direct Access SCSI2 device da0: 20.0MB/s transfers (10.0MHz, offset 8, 16bit), Tagged Queueing Enabled da0: 2048MB (4194304 512 byte sectors: 255H 63S/T 261C) da1 at ahc0 bus 0 target 2 lun 0 da1: Fixed Direct Access SCSI2 device da1: 20.0MB/s transfers (10.0MHz, offset 8, 16bit), Tagged Queueing Enabled da1: 2048MB (4194304 512 byte sectors: 255H 63S/T 261C) da2 at ahc0 bus 0 target 3 lun 0 da2: Fixed Direct Access SCSI2 device da2: 20.0MB/s transfers (10.0MHz, offset 8, 16bit), Tagged Queueing Enabled da2: 2048MB (4194304 512 byte sectors: 255H 63S/T 261C) changing root device to wd0s1a ch0 at isp0 bus 0 target 6 lun 0 ch0: Removable Changer SCSI2 device ch0: 3.300MB/s transfers ch0: 11 slots, 2 drives, 1 picker, 0 portals To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Mon Aug 17 14:45:15 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id OAA27992 for freebsd-smp-outgoing; Mon, 17 Aug 1998 14:45:15 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from mail.ftf.dk (mail.ftf.dk [129.142.64.2]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id OAA27971 for ; Mon, 17 Aug 1998 14:45:10 -0700 (PDT) (envelope-from regnauld@deepo.prosa.dk) Received: from mail.prosa.dk ([192.168.100.254]) by mail.ftf.dk (8.8.8/8.8.8/gw-ftf-1.0) with ESMTP id XAA12961; Mon, 17 Aug 1998 23:49:42 +0200 (CEST) (envelope-from regnauld@deepo.prosa.dk) Received: from deepo.prosa.dk (deepo.prosa.dk [192.168.100.10]) by mail.prosa.dk (8.8.8/8.8.5/prosa-1.1) with ESMTP id XAA27294; Mon, 17 Aug 1998 23:52:45 +0200 (CEST) Received: (from regnauld@localhost) by deepo.prosa.dk (8.8.8/8.8.5/prosa-1.1) id XAA03900; Mon, 17 Aug 1998 23:44:11 +0200 (CEST) Message-ID: <19980817234411.28820@deepo.prosa.dk> Date: Mon, 17 Aug 1998 23:44:11 +0200 From: Philippe Regnauld To: Terry Lambert Cc: freebsd-smp@FreeBSD.ORG Subject: Re: PR440fx (Toshiba) + SMP ? References: <19980817220046.44173@deepo.prosa.dk> <199808172114.OAA08860@usr06.primenet.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.88e In-Reply-To: <199808172114.OAA08860@usr06.primenet.com>; from Terry Lambert on Mon, Aug 17, 1998 at 09:14:39PM +0000 X-Operating-System: FreeBSD 2.2.6-RELEASE i386 Phone: +45 3336 4148 Address: Ahlefeldtsgade 16, 1359 Copenhagen K, Denmark Organization: PROSA Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Terry Lambert writes: > > -- > > Processors: APIC ID Version State Family Model Step Flags > > 0 0x11 BSP, usable 6 1 9 0xfbff > > -- > > Bus: Bus ID Type > > You have failed to tell your BIOS that you have a second CPU; therefore it > has failed to configure the APIC ID on the second CPU, or otherwise identify > it in the mptable so that FreeBSD can see it. Well, I _cannot_ tell my BIOS to force detecting the other CPU -- this might be a stepping problem -- it was ordered from Computer123.com, and I don't remember them garanteeing the same stepping on both CPUs... > A correctly configured BIOS will show two processors with different APIC ID's > under the "Processors:" section in the mptable. I thought so... -- -[ Philippe Regnauld / sysadmin / regnauld@deepo.prosa.dk / +55.4N +11.3E ]- The Internet is busy. Please try again later. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Mon Aug 17 14:55:41 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id OAA00174 for freebsd-smp-outgoing; Mon, 17 Aug 1998 14:55:41 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from shadow.spel.com (elevator.cablenet-va.com [208.206.84.25]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id OAA00156 for ; Mon, 17 Aug 1998 14:55:27 -0700 (PDT) (envelope-from mturpin@shadow.spel.com) Received: from localhost (mturpin@localhost) by shadow.spel.com (8.8.5/8.8.5) with SMTP id RAA18164; Mon, 17 Aug 1998 17:56:45 -0400 (EDT) Date: Mon, 17 Aug 1998 17:56:44 -0400 (EDT) From: Mark turpin To: Philippe Regnauld cc: freebsd-smp@FreeBSD.ORG Subject: Re: PR440fx (Toshiba) + SMP ? In-Reply-To: <19980817220046.44173@deepo.prosa.dk> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org I have the same board and it works fine... I kept the BIOS the same as it was when I got it.. v1.00.01.DI? I think that's the number... Does it say 2 processors on the POST screen? On Mon, 17 Aug 1998, Philippe Regnauld wrote: > I've just replaced my GA-586DX with a PR440fx board with 2 P6-166, > CSS audio and Intel 100Mb fxp. > > (the BIOS displays Toshiba at boot). > > - I've upgraded to the latest BIOS > - I've recompiled my kernel with I686 (-current from last night) > > But FreeBSD insists on seeing only one CPU -- I've even tried hardwiring > with NCPU=2, to no avail. I've also tried playing with the BIOS > SMP settings (1.1 or 1.4) -- no luck... > > dmesg + mptable follows > > > Copyright (c) 1992-1998 FreeBSD Inc. > Copyright (c) 1982, 1986, 1989, 1991, 1993 > The Regents of the University of California. All rights reserved. > FreeBSD 3.0-CURRENT #0: Mon Aug 17 22:15:45 MET DST 1998 > regnauld@tetard.glou.eu.org:/usr/src/sys/compile/TETARD.smp > Timecounter "i8254" frequency 1193182 Hz cost 3588 ns > CPU: Pentium Pro (686-class CPU) > Origin = "GenuineIntel" Id = 0x619 Stepping=9 > Features=0xfbff > real memory = 67108864 (65536K bytes) > avail memory = 62853120 (61380K bytes) > FreeBSD/SMP: Multiprocessor motherboard > cpu0 (BSP): apic id: 0, version: 0x00040011, at 0xfec08000 > io0 (APIC): apic id: 13, version: 0x00170011, at 0xfec00000 > Probing for devices on PCI bus 0: > chip0: rev 0x02 on pci0.0.0 > fxp0: rev 0x02 int a irq 18 on pci0.6. > 0 > fxp0: Ethernet address 00:a0:c9:49:58:d9 > chip1: rev 0x01 on pci0.7.0 > ahc0: rev 0x00 int a irq 17 on pci0.9. > 0 > ahc0: aic7880 Wide Channel, SCSI Id=7, 16/255 SCBs > scbus0 at ahc0 bus 0 > ahc0: target 0 Tagged Queuing Device > sd0 at scbus0 target 0 lun 0 > sd0: type 0 fixed SCSI 2 > sd0: Direct-Access 4134MB (8467200 512 byte sectors) > sd0: with 8205 cyls, 6 heads, and an average 171 sectors/track > vga0: rev 0x00 int a irq 18 on pci0.17.0 > Probing for devices on the ISA bus: > sc0 at 0x60-0x6f irq 1 on motherboard > sc0: VGA color <12 virtual consoles, flags=0x0> > ed0 at 0x340-0x35f irq 9 on isa > ed0: address 00:20:a9:00:4b:bd, type NE2000 (16 bit) > sio0 at 0x3f8-0x3ff irq 4 on isa > sio0: type 16550A > sio1 at 0x2f8-0x2ff irq 3 on isa > sio1: type 16550A > psm0 not found at 0x60 > lpt0 at 0x378-0x37f irq 7 on isa > lpt0: Interrupt-driven port > lp0: TCP/IP capable interface > fdc0 at 0x3f0-0x3f7 irq 6 drq 2 on isa > fdc0: FIFO enabled, 8 bytes threshold > fd0: 1.44MB 3.5in > npx0 on motherboard > npx0: INT 16 interface > css0 at 0x534 irq 5 drq 1 flags 0x100 on isa > snd0: > mpu0 not found at 0x330 > APIC_IO: routing 8254 via 8259 on pin 0 > > > > =============================================================================== > > MPTable, version 2.0.15 > > ------------------------------------------------------------------------------- > > MP Floating Pointer Structure: > > location: BIOS > physical address: 0x000f7ed0 > signature: '_MP_' > length: 16 bytes > version: 1.4 > checksum: 0x33 > mode: Virtual Wire > > ------------------------------------------------------------------------------- > > MP Config Table Header: > > physical address: 0x000f7ee0 > signature: 'PCMP' > base table length: 216 > version: 1.4 > checksum: 0x62 > OEM ID: 'INTEL ' > Product ID: 'PR440FX ' > OEM table pointer: 0x00000000 > OEM table size: 0 > entry count: 20 > local APIC address: 0xfec08000 > extended table length: 120 > extended table checksum: 15 > > ------------------------------------------------------------------------------- > > MP Config Base Table Entries: > > -- > Processors: APIC ID Version State Family Model Step Flags > 0 0x11 BSP, usable 6 1 9 0xfbff > -- > Bus: Bus ID Type > 0 PCI > 18 ISA > -- > I/O APICs: APIC ID Version State Address > 13 0x11 usable 0xfec00000 > -- > I/O Ints: Type Polarity Trigger Bus ID IRQ APIC ID PIN# > > ExtINT active-hi edge 18 0 13 0 > INT active-hi edge 18 1 13 1 > INT active-hi edge 18 3 13 3 > INT active-hi edge 18 4 13 4 > INT active-hi edge 18 5 13 5 > INT active-hi edge 18 6 13 6 > INT active-hi edge 18 7 13 7 > INT active-hi edge 18 8 13 8 > INT active-hi edge 18 9 13 9 > INT active-hi edge 18 10 13 10 > INT active-hi edge 18 12 13 12 > INT active-hi edge 18 14 13 14 > INT active-hi edge 18 15 13 15 > INT active-lo level 0 17:A 13 18 > INT active-lo level 0 9:A 13 17 > INT active-lo level 0 6:A 13 18 > -- > Local Ints: Type Polarity Trigger Bus ID IRQ APIC ID PIN# > > ------------------------------------------------------------------------------- > > MP Config Extended Table Entries: > > -- > > bus ID: 0 address type: memory address > address base: 0xd4000 > address range: 0x4000 > -- > > bus ID: 0 address type: memory address > address base: 0xd8000 > address range: 0x4000 > -- > > bus ID: 0 address type: memory address > address base: 0xdc000 > address range: 0x4000 > -- > > bus ID: 0 address type: memory address > address base: 0xa0000 > address range: 0x20000 > -- > > bus ID: 0 address type: memory address > address base: 0x4000000 > address range: 0xfc000000 > -- > > bus ID: 0 address type: I/O address > address base: 0x0 > address range: 0x10000 > > ------------------------------------------------------------------------------- > > # SMP kernel config file options: > > > # Required: > options SMP # Symmetric MultiProcessor Kernel > options APIC_IO # Symmetric (APIC) I/O > > # Optional (built-in defaults will work in most cases): > #options NCPU=1 # number of CPUs > #options NBUS=2 # number of busses > #options NAPIC=1 # number of IO APICs > #options NINTR=24 # number of INTs > > =============================================================================== > > > -- > -[ Philippe Regnauld / sysadmin / regnauld@deepo.prosa.dk / +55.4N +11.3E ]- > > The Internet is busy. Please try again later. > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-smp" in the body of the message > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Mon Aug 17 16:29:26 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id QAA17998 for freebsd-smp-outgoing; Mon, 17 Aug 1998 16:29:26 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from mail.ftf.dk (mail.ftf.dk [129.142.64.2]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id QAA17992 for ; Mon, 17 Aug 1998 16:29:22 -0700 (PDT) (envelope-from regnauld@deepo.prosa.dk) Received: from mail.prosa.dk ([192.168.100.254]) by mail.ftf.dk (8.8.8/8.8.8/gw-ftf-1.0) with ESMTP id BAA13576 for ; Tue, 18 Aug 1998 01:33:54 +0200 (CEST) (envelope-from regnauld@deepo.prosa.dk) Received: from deepo.prosa.dk (deepo.prosa.dk [192.168.100.10]) by mail.prosa.dk (8.8.8/8.8.5/prosa-1.1) with ESMTP id BAA27398 for ; Tue, 18 Aug 1998 01:36:58 +0200 (CEST) Received: (from regnauld@localhost) by deepo.prosa.dk (8.8.8/8.8.5/prosa-1.1) id BAA04474; Tue, 18 Aug 1998 01:28:24 +0200 (CEST) Message-ID: <19980818012824.51792@deepo.prosa.dk> Date: Tue, 18 Aug 1998 01:28:24 +0200 From: Philippe Regnauld To: freebsd-smp@FreeBSD.ORG Subject: Re: PR440fx (Toshiba) + SMP ? References: <19980817220046.44173@deepo.prosa.dk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.88e In-Reply-To: <19980817220046.44173@deepo.prosa.dk>; from Philippe Regnauld on Mon, Aug 17, 1998 at 10:00:46PM +0200 X-Operating-System: FreeBSD 2.2.6-RELEASE i386 Phone: +45 3336 4148 Address: Ahlefeldtsgade 16, 1359 Copenhagen K, Denmark Organization: PROSA Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Philippe Regnauld writes: > I've just replaced my GA-586DX with a PR440fx board with 2 P6-166, > CSS audio and Intel 100Mb fxp. Hrmm... I need the conical hat -- sorry to everyone whose time I wasted: the second CPU was not socketed firmly :-} Works like a charm, now: 2x200 MHz and flying. PS: did anyone get the CSS to work properly with Luigi's code ? I only got it work with the css0 controller... Also, Netscape 4.05 locks up with 90%, something which didn't happen with P5-SMP... -- -[ Philippe Regnauld / sysadmin / regnauld@deepo.prosa.dk / +55.4N +11.3E ]- The Internet is busy. Please try again later. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Mon Aug 17 18:29:34 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id SAA07898 for freebsd-smp-outgoing; Mon, 17 Aug 1998 18:29:34 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from mail1.its.rpi.edu (mail1.its.rpi.edu [128.113.100.7]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id SAA07883 for ; Mon, 17 Aug 1998 18:29:31 -0700 (PDT) (envelope-from drosih@rpi.edu) Received: from [128.113.24.47] (ned.dynamic.rpi.edu [128.113.24.149]) by mail1.its.rpi.edu (8.8.8/8.8.6) with ESMTP id VAA43772; Mon, 17 Aug 1998 21:29:06 -0400 Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" X-Sender: drosih@pop1.rpi.edu Message-Id: In-Reply-To: <19980817220046.44173@deepo.prosa.dk> Date: Mon, 17 Aug 1998 21:33:00 -0400 To: Philippe Regnauld , freebsd-smp@FreeBSD.ORG From: Garance A Drosihn Subject: Re: PR440fx (Toshiba) + SMP ? Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org At 10:00 PM +0200 8/17/98, Philippe Regnauld wrote: > I've just replaced my GA-586DX with a PR440fx board with 2 P6-166, > CSS audio and Intel 100Mb fxp. > > (the BIOS displays Toshiba at boot). > > - I've upgraded to the latest BIOS > - I've recompiled my kernel with I686 (-current from last night) > > But FreeBSD insists on seeing only one CPU -- I've even tried > hardwiring with NCPU=2, to no avail. I've also tried playing > with the BIOS SMP settings (1.1 or 1.4) -- no luck... This is the behavior I got from my Toshiba, primarily because I am a bit new to Intel-ish hardware. I bought a PPro chip and tossed it in the second slot, and clipped the fan over it (after dabbing the CPU with some stuff that improves heat conductivity between the chip and the fan). Nothing would see the second CPU. Turns out I had put the CPU in wrong. I knew it was a "zero insertion force" socket, so I dropped the chip on the socket and sure enough it seemed to fit there with zero force. However, I had not lifted up the little lever on the side of the socket before dropping the chip in there, so the CPU wasn't really in the socket (even though it seemed OK). No wonder it was a little hard to hook the fan over that CPU and onto the socket!! So, I had to unhook the fan for the second CPU, lift the lever, let the second CPU really drop into the socket, and then put the lever back down. I hooked the fan back up, and it started working fine. I'm pretty sure there was nothing in the BIOS that I had to change to get the second CPU working. If you want to eliminate FreeBSD from the equation, you could check the BIOS to see if *it* sees the second CPU. Reboot the system, hit "F1" to get into the BIOS, and once the BIOS screen comes up then hit the right-arrow key to get into the "Advanced" BIOS screen. It should list "Processor 0 Type" and "Processor 1 Type". In my case those two lines say "Pentium(R) Pro CPU" and the "Processor Speed" line says 200 MHz. If the BIOS doesn't see the CPU, then it won't have the line about "Processor 1 Type". And if the BIOS doesn't see the CPU, then FreeBSD certainly won't. Also, at least on my machine (a Toshiba 6200M), you will *not* see the name "Toshiba" displayed at boot if you have updated to the latest BIOS from Intel for the PR440FX. When you get into the BIOS, does it report itself as version 1.00.08.DI0, or does it report 1.00.01D8 ? Upgrading the BIOS (at least on my machine) is somewhat tricky, as the only way to get the BIOS upgrade to work is to change the jumper on the motherboard to use the "emergency" BIOS upgrade procedure. Once I got the CPU in there correctly, it seems to be working fine under an SMP kernel under FreeBSD. I do have a problem if I try to rebuild Modula3 with an SMP kernel, but that's the only oddball thing I have noticed. --- Garance Alistair Drosehn = gad@eclipse.its.rpi.edu Senior Systems Programmer or drosih@rpi.edu Rensselaer Polytechnic Institute To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Mon Aug 17 22:40:13 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id WAA02930 for freebsd-smp-outgoing; Mon, 17 Aug 1998 22:40:13 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from hermes.hrz.uni-bielefeld.de (hermes.hrz.uni-bielefeld.de [129.70.4.55]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id WAA02922 for ; Mon, 17 Aug 1998 22:40:09 -0700 (PDT) (envelope-from lkoeller@post.uni-bielefeld.de) Received: from mitch.hrz.uni-bielefeld.de (9134@mitch.hrz.uni-bielefeld.de [129.70.4.17]) by hermes.hrz.uni-bielefeld.de (8.8.6/8.8.6) with ESMTP id HAA24693; Tue, 18 Aug 1998 07:39:55 +0200 (METDST) Received: from localhost by mitch.hrz.uni-bielefeld.de with SMTP (8.8.6/16.2) id FAA26168; Tue, 18 Aug 1998 05:39:31 GMT Message-Id: <199808180539.FAA26168@mitch.hrz.uni-bielefeld.de> X-Mailer: exmh version 2.0.2 2/24/98 From: Lars =?iso-8859-1?Q?K=F6ller?= To: Terry Lambert cc: chuckr@glue.umd.edu, freebsd-smp@FreeBSD.ORG Subject: Re: Per processor load? In-reply-to: tlambert's message of Mon, 17 Aug 1998 18:12:00 -0000. <199808171812.LAA06500@usr08.primenet.com> X-Face: eCcoCV}FjV*O{6>[1$XP/e%]TJhEw2MF33dFh)^HM7Gfd=[/(4+0a$~ > But thats only the first part of the problem! Where can I get > > processor specific load information? I need to draw a separate load > > graph for each CPU?? > > System load is based on number of processees in the ready-to-run > state. O.k. so far so good .... > For a symmetric system, if the load is 1.0 or above, both CPU's > should be actively working. > > I suppose that what you are asking for is a "processor not idle > in the case of 1.0 >= load >= 0.0". > > To get this, you would have to insert counters into the per CPU > idle loops, probably using the Appendix H cycle counter before and > after the per CPU HLT instruction, subtracting the count at exit > of the last HLT from both, and then subtracting the entry from the > exit, and dividing to get an "idle ratio". > > Gathering this type of statistic could be actively harmful to CPU > latency coming out of the HLT condition, and could be as high as 10% > to 20% of the systems ability to do work. The basic idea was to treat the CPU's as seperate systems each with it's own load. This is well known from HPUX, Linux, Solaris, ... They display the following in, e.g. top: System: share Tue Aug 18 07:30:58 1998 Load averages: 2.42, 2.29, 2.28 280 processes: 273 sleeping, 5 running, 2 zombies Cpu states: CPU LOAD USER NICE SYS IDLE BLOCK SWAIT INTR SSYS 0 2.62 0.4% 97.6% 2.0% 0.0% 0.0% 0.0% 0.0% 0.0% 1 2.22 0.8% 97.0% 2.2% 0.0% 0.0% 0.0% 0.0% 0.0% --- ---- ----- ----- ----- ----- ----- ----- ----- ----- avg 2.42 0.6% 97.2% 2.2% 0.0% 0.0% 0.0% 0.0% 0.0% Memory: 180344K (29336K) real, 256220K (66940K) virtual, 5160K free Page# 1/26 CPU TTY PID USERNAME PRI NI SIZE RES STATE TIME %WCPU %CPU COMMAND 0 ? 19703 mcfutz 251 25 632K 116K run 6:05 80.27 80.13 schlu 1 ? 19721 physik 251 25 632K 112K run 4:52 49.42 49.34 process 1 ? 5375 plond 251 25 34756K 15900K run 2173:38 46.66 46.58 l502.exe So the idea was in a first step to display the load of each CPU in a seperate graph of xperfmon++ . Perhaps it's a better idea to display the other parameters like IO rate, interrupts, ... but I don't see a way to get them CPU dependent. Is there any CPU-private parameter in the kernel? Thanks and regards Lars -- E-Mail: | Lars Köller Lars.Koeller@Uni-Bielefeld.DE | UNIX Sysadmin lkoeller@cc.FH-Lippe.DE | Computing Center PGP-key: | University of Bielefeld http://www.nic.surfnet.nl/pgp/pks-toplev.html | Germany ----------- FreeBSD, what else? ---- http://www.freebsd.org ------------- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Tue Aug 18 00:51:50 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id AAA14547 for freebsd-smp-outgoing; Tue, 18 Aug 1998 00:51:50 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from dingo.cdrom.com ([203.8.14.118]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id AAA14541 for ; Tue, 18 Aug 1998 00:51:45 -0700 (PDT) (envelope-from mike@dingo.cdrom.com) Received: from dingo.cdrom.com (localhost.hilink.com.au [127.0.0.1] (may be forged)) by dingo.cdrom.com (8.8.8/8.8.8) with ESMTP id AAA14742; Tue, 18 Aug 1998 00:49:30 GMT (envelope-from mike@dingo.cdrom.com) Message-Id: <199808180049.AAA14742@dingo.cdrom.com> X-Mailer: exmh version 2.0.2 2/24/98 To: Luoqi Chen cc: mal@algonet.se, freebsd-smp@FreeBSD.ORG Subject: Re: SMP + Wine => panic In-reply-to: Your message of "Wed, 12 Aug 1998 14:04:04 -0400." <199808121804.OAA09655@lor.watermarkgroup.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Tue, 18 Aug 1998 00:49:30 +0000 From: Mike Smith Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > The last patches did cure the problem. It looks like the program > > I wanted to run is a no-go, but that is a problem with wine, not freebsd. > > > > Thanks for the help! Tell me if you need help with more testing to get > > the code committed to the main tree. > > > > _ > > Mats Lofkvist > > mal@algonet.se > > > Thanks for the testing result, it worked on my PC too. I don't know what > exactly needs to be done to get the fix committed, I'll just file a PR > and hopefully someone would take a look and commit it. > > -lq > > PS: I filed the PR and the number is i386/7591. Thanks for this; your changes looked good, so I've committed them and closed both the PRs you reference. Until we can convince you to become a committer yourself, filing PRs and making sure we know about them is the best way to get stuff committed. -- \\ Sometimes you're ahead, \\ Mike Smith \\ sometimes you're behind. \\ mike@smith.net.au \\ The race is long, and in the \\ msmith@freebsd.org \\ end it's only with yourself. \\ msmith@cdrom.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Tue Aug 18 02:02:56 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id CAA21590 for freebsd-smp-outgoing; Tue, 18 Aug 1998 02:02:56 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from mail.ftf.dk (mail.ftf.dk [129.142.64.2]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id CAA21579 for ; Tue, 18 Aug 1998 02:02:45 -0700 (PDT) (envelope-from regnauld@deepo.prosa.dk) Received: from mail.prosa.dk ([192.168.100.254]) by mail.ftf.dk (8.8.8/8.8.8/gw-ftf-1.0) with ESMTP id LAA23788; Tue, 18 Aug 1998 11:07:15 +0200 (CEST) (envelope-from regnauld@deepo.prosa.dk) Received: from deepo.prosa.dk (deepo.prosa.dk [192.168.100.10]) by mail.prosa.dk (8.8.8/8.8.5/prosa-1.1) with ESMTP id LAA28091; Tue, 18 Aug 1998 11:10:25 +0200 (CEST) Received: (from regnauld@localhost) by deepo.prosa.dk (8.8.8/8.8.5/prosa-1.1) id LAA07030; Tue, 18 Aug 1998 11:01:52 +0200 (CEST) Message-ID: <19980818110152.12065@deepo.prosa.dk> Date: Tue, 18 Aug 1998 11:01:52 +0200 From: Philippe Regnauld To: Garance A Drosihn Cc: freebsd-smp@FreeBSD.ORG Subject: Re: PR440fx (Toshiba) + SMP ? References: <19980817220046.44173@deepo.prosa.dk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.88e In-Reply-To: ; from Garance A Drosihn on Mon, Aug 17, 1998 at 09:33:00PM -0400 X-Operating-System: FreeBSD 2.2.6-RELEASE i386 Phone: +45 3336 4148 Address: Ahlefeldtsgade 16, 1359 Copenhagen K, Denmark Organization: PROSA Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Garance A Drosihn writes: > > Turns out I had put the CPU in wrong. I knew it was a "zero > insertion force" socket, so I dropped the chip on the socket > and sure enough it seemed to fit there with zero force. Heh. I feel less alone :-) > If you want to eliminate FreeBSD from the equation, you could > check the BIOS to see if *it* sees the second CPU. It does now. > Also, at least on my machine (a Toshiba 6200M), you will *not* > see the name "Toshiba" displayed at boot if you have updated > to the latest BIOS from Intel for the PR440FX. Very true -- the latest (01.00.08, I think) displays a plain "Intel Corporation" message. > Upgrading the BIOS (at least > on my machine) is somewhat tricky, as the only way to get > the BIOS upgrade to work is to change the jumper on the > motherboard to use the "emergency" BIOS upgrade procedure. This was mentioned in the list archives, yes: it always fails in standard mode -- you have to move the jumper to the up position, boot blind, wait for two beeps, turn off the system, and put the jumper back down. -- -[ Philippe Regnauld / sysadmin / regnauld@deepo.prosa.dk / +55.4N +11.3E ]- The Internet is busy. Please try again later. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Tue Aug 18 07:14:49 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id HAA24027 for freebsd-smp-outgoing; Tue, 18 Aug 1998 07:14:49 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from mail.ftf.dk (mail.ftf.dk [129.142.64.2]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id HAA24008; Tue, 18 Aug 1998 07:14:45 -0700 (PDT) (envelope-from regnauld@deepo.prosa.dk) Received: from mail.prosa.dk ([192.168.100.254]) by mail.ftf.dk (8.8.8/8.8.8/gw-ftf-1.0) with ESMTP id QAA07178; Tue, 18 Aug 1998 16:19:42 +0200 (CEST) (envelope-from regnauld@deepo.prosa.dk) Received: from deepo.prosa.dk (deepo.prosa.dk [192.168.100.10]) by mail.prosa.dk (8.8.8/8.8.5/prosa-1.1) with ESMTP id QAA28650; Tue, 18 Aug 1998 16:22:08 +0200 (CEST) Received: (from regnauld@localhost) by deepo.prosa.dk (8.8.8/8.8.5/prosa-1.1) id QAA23756; Tue, 18 Aug 1998 16:13:35 +0200 (CEST) Message-ID: <19980818161335.30289@deepo.prosa.dk> Date: Tue, 18 Aug 1998 16:13:35 +0200 From: Philippe Regnauld To: Werner Griessl Cc: current@FreeBSD.ORG, smp@FreeBSD.ORG Subject: Re: softupdates and smp crash References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.88e In-Reply-To: ; from Werner Griessl on Tue, Aug 18, 1998 at 01:08:12PM +0200 X-Operating-System: FreeBSD 2.2.6-RELEASE i386 Phone: +45 3336 4148 Address: Ahlefeldtsgade 16, 1359 Copenhagen K, Denmark Organization: PROSA Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Werner Griessl writes: > > My system with current-smp from yesterday (Aug. 17) crashes with > "make buildworld -jx" with x>1 and softupdates on for /usr/obj. > Have no problems with j1,softupdates on and j4 or j8 with softupdates off. Same here last night. 2xPPro-200, 64MB, UW IBM DCAS. > Appended is a crash-log "gdb6.log" with j4 #0 boot (howto=260) at ../../kern/kern_shutdown.c:286 #1 0xf01151c1 in panic (fmt=0xf01ab861 "softdep_lock: locking against myself") at ../../kern/kern_shutdown.c:429 #2 0xf01ab8c9 in acquire_lock (lk=0xf02265ac) at ../../ufs/ffs/ffs_softdep.c:263 +-> #3 0xf01aee54 in initiate_write_filepage (pagedep=0xf0c49e00, bp=0xf35300b0) | at ../../ufs/ffs/ffs_softdep.c:2745 | #4 0xf01aecbf in softdep_disk_io_initiation (bp=0xf35300b0) | I crashed here. -- -[ Philippe Regnauld / sysadmin / regnauld@deepo.prosa.dk / +55.4N +11.3E ]- The Internet is busy. Please try again later. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Tue Aug 18 11:38:59 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id LAA04789 for freebsd-smp-outgoing; Tue, 18 Aug 1998 11:38:59 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from smtp04.primenet.com (smtp04.primenet.com [206.165.6.134]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id LAA04784 for ; Tue, 18 Aug 1998 11:38:57 -0700 (PDT) (envelope-from tlambert@usr06.primenet.com) Received: (from daemon@localhost) by smtp04.primenet.com (8.8.8/8.8.8) id LAA28446; Tue, 18 Aug 1998 11:38:19 -0700 (MST) Received: from usr06.primenet.com(206.165.6.206) via SMTP by smtp04.primenet.com, id smtpd028399; Tue Aug 18 11:38:16 1998 Received: (from tlambert@localhost) by usr06.primenet.com (8.8.5/8.8.5) id LAA20956; Tue, 18 Aug 1998 11:38:13 -0700 (MST) From: Terry Lambert Message-Id: <199808181838.LAA20956@usr06.primenet.com> Subject: Re: Per processor load? To: Lars.Koeller@post.uni-bielefeld.de (Lars =?iso-8859-1?Q?K=F6ller?=) Date: Tue, 18 Aug 1998 18:38:13 +0000 (GMT) Cc: tlambert@primenet.com, chuckr@glue.umd.edu, freebsd-smp@FreeBSD.ORG In-Reply-To: <199808180539.FAA26168@mitch.hrz.uni-bielefeld.de> from "Lars =?iso-8859-1?Q?K=F6ller?=" at Aug 18, 98 07:39:30 am X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > For a symmetric system, if the load is 1.0 or above, both CPU's > > should be actively working. > > > > I suppose that what you are asking for is a "processor not idle > > in the case of 1.0 >= load >= 0.0". > > > > To get this, you would have to insert counters into the per CPU > > idle loops, probably using the Appendix H cycle counter before and > > after the per CPU HLT instruction, subtracting the count at exit > > of the last HLT from both, and then subtracting the entry from the > > exit, and dividing to get an "idle ratio". > > > > Gathering this type of statistic could be actively harmful to CPU > > latency coming out of the HLT condition, and could be as high as 10% > > to 20% of the systems ability to do work. > > The basic idea was to treat the CPU's as seperate systems each with > it's own load. This is well known from HPUX, Linux, Solaris, ... > They display the following in, e.g. top: > > System: share Tue Aug 18 07:30:58 1998 > Load averages: 2.42, 2.29, 2.28 > 280 processes: 273 sleeping, 5 running, 2 zombies > Cpu states: > CPU LOAD USER NICE SYS IDLE BLOCK SWAIT INTR SSYS > 0 2.62 0.4% 97.6% 2.0% 0.0% 0.0% 0.0% 0.0% 0.0% > 1 2.22 0.8% 97.0% 2.2% 0.0% 0.0% 0.0% 0.0% 0.0% > --- ---- ----- ----- ----- ----- ----- ----- ----- ----- > avg 2.42 0.6% 97.2% 2.2% 0.0% 0.0% 0.0% 0.0% 0.0% This basically implies a scheduler artifact; each CPU must have its own ready-to-run queue for you to get this statistic; I'm sure that on Solaris, at least, you have to know how to grovel /dev/kmem for the information. FreeBSD is symmetric. That is, there is only one ready-to-run queue for all processors. Anything else would either result in potential job starvation (inequity because on one processor, the jobs you are competing with use 75% of their quantum, being compute intensive, and on the other, they use only 10% of their quantum, being I/O intensive). To combat this, there must be highly complex scheduler changes to ensure CPU affinity and assure process migration based on average behaviour. This is an imperfect homeostasis, at best. To my knowledge, only Solaris attempts this, and then only started doing it halfway well in 2.5. CPU affinity s a big win for non-cache-busting programs (it ensures some cache will be valid the next time the process is run), but for most servers, it's a NOP. Depending on your application mix, expect to perhaps lose (or gain) total compute power over it. As far as INTR time goes, I notice it's not reported. This is not surprising. In Symmetric (APIC) I/O, or "virtual wire mode", the interrupt is directed to any available processor, lowest APIC ID first (see the Intel MP Spec version 1.4). It's really not possible, unless you modify the ISR to record APIC ID, and reverse look it up (an expensive operation) on each interrupt, to determine which CPU is actually getting the interrupt. I notice the other fields are not reported as well, probably for similar reasons. > Memory: 180344K (29336K) real, 256220K (66940K) virtual, 5160K free Page# 1/26 > > CPU TTY PID USERNAME PRI NI SIZE RES STATE TIME %WCPU %CPU COMMAND > 0 ? 19703 mcfutz 251 25 632K 116K run 6:05 80.27 80.13 schlu > 1 ? 19721 physik 251 25 632K 112K run 4:52 49.42 49.34 process > 1 ? 5375 plond 251 25 34756K 15900K run 2173:38 46.66 46.58 l502.exe Pretty obviously, there aren't two running process on that one CPU. A CPU can be in user space in only one process at a time. 8-). I think what they are doing, since they can tell you the CPU, is either recording what CPU they last ran on, *or*, they are reporting which of the multiple run queues that the program is on. The way to tell this would be to dump this information, and then count the number of processes on one CPU or the other. If there isn't an imbalance, then they are talking ready-to-run-queue. If there *is* an imbalance, then they *may* be talking ready-to-run queue, if they are cache-busting round-robin. This would be a design error, but may be what is happening in an attempt to achieve load balance between the CPU's. To do this, if you last ran on CPU M of N, then you next run on CPU M + 1; when M = N, you next run on CPU 0. For programs that don't benefit from L1 cache, this is a normative win; but again, they would be special casing the code for something that wasn't a very general purpose use. In either case, the statistics that *I* would find interesting is "process migration rate" and "cache miss rate"; the second would be as hard to do as "idle time". 8-(. > So the idea was in a first step to display the load of each CPU in a > seperate graph of xperfmon++ . Perhaps it's a better idea to display > the other parameters like IO rate, interrupts, ... but I don't see a > way to get them CPU dependent. > > Is there any CPU-private parameter in the kernel? There are CPU private areas, certainly; these are memory regions mapped to a single CPU. The processor data area and processor stack are examples; you can get all of them by looking in the locore code for the SMP case, if you are interested. I think the thing to do is to better understand the scheduler and the model to determine which metrics are truly useful, and which statistics are "too expensive". Both Steve Passe and John Dyson would be good resources on this. Note that the FreeBSD SMP scheduling algorithm is not really set in stone yet; for example, there is experimental kernel threading code and CPU affinity code (I'm not sure how complex this is; certainly it's no Solaris) that would make some of what I said weigh differently based on type of load expected. Unfortunately, displaying this information is complicated, in that you have to know what you are displaying... Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Tue Aug 18 22:41:02 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id WAA24154 for freebsd-smp-outgoing; Tue, 18 Aug 1998 22:41:02 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from hermes.hrz.uni-bielefeld.de (hermes.hrz.uni-bielefeld.de [129.70.4.55]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id WAA24146 for ; Tue, 18 Aug 1998 22:41:00 -0700 (PDT) (envelope-from lkoeller@post.uni-bielefeld.de) Received: from mitch.hrz.uni-bielefeld.de (9134@mitch.hrz.uni-bielefeld.de [129.70.4.17]) by hermes.hrz.uni-bielefeld.de (8.8.6/8.8.6) with ESMTP id HAA23729; Wed, 19 Aug 1998 07:40:44 +0200 (METDST) Received: from localhost by mitch.hrz.uni-bielefeld.de with SMTP (8.8.6/16.2) id FAA24476; Wed, 19 Aug 1998 05:40:16 GMT Message-Id: <199808190540.FAA24476@mitch.hrz.uni-bielefeld.de> X-Mailer: exmh version 2.0.2 2/24/98 From: Lars =?iso-8859-1?Q?K=F6ller?= To: Terry Lambert cc: Lars.Koeller@post.uni-bielefeld.de (Lars =?iso-8859-1?Q?K=F6ller?=), chuckr@glue.umd.edu, freebsd-smp@FreeBSD.ORG Subject: Re: Per processor load? In-reply-to: tlambert's message of Tue, 18 Aug 1998 18:38:13 -0000. <199808181838.LAA20956@usr06.primenet.com> X-Face: eCcoCV}FjV*O{6>[1$XP/e%]TJhEw2MF33dFh)^HM7Gfd=[/(4+0a$~ > > Gathering this type of statistic could be actively harmful to CPU > > > latency coming out of the HLT condition, and could be as high as 10% > > > to 20% of the systems ability to do work. > > > > The basic idea was to treat the CPU's as seperate systems each with > > it's own load. This is well known from HPUX, Linux, Solaris, ... > > They display the following in, e.g. top: > > > > System: share Tue Aug 18 07:30:58 1 998 > > Load averages: 2.42, 2.29, 2.28 > > 280 processes: 273 sleeping, 5 running, 2 zombies > > Cpu states: > > CPU LOAD USER NICE SYS IDLE BLOCK SWAIT INTR SSYS > > 0 2.62 0.4% 97.6% 2.0% 0.0% 0.0% 0.0% 0.0% 0.0% > > 1 2.22 0.8% 97.0% 2.2% 0.0% 0.0% 0.0% 0.0% 0.0% > > --- ---- ----- ----- ----- ----- ----- ----- ----- ----- > > avg 2.42 0.6% 97.2% 2.2% 0.0% 0.0% 0.0% 0.0% 0.0% Sorry, I forgot to mention this top was running on HPUX 10.20. > This basically implies a scheduler artifact; each CPU must have its own > ready-to-run queue for you to get this statistic; I'm sure that on > Solaris, at least, you have to know how to grovel /dev/kmem for the > information. > > FreeBSD is symmetric. That is, there is only one ready-to-run queue > for all processors. Anything else would either result in potential > job starvation (inequity because on one processor, the jobs you are > competing with use 75% of their quantum, being compute intensive, > and on the other, they use only 10% of their quantum, being I/O > intensive). > > ..... snip ... > > As far as INTR time goes, I notice it's not reported. This is not > surprising. In Symmetric (APIC) I/O, or "virtual wire mode", the > interrupt is directed to any available processor, lowest APIC ID > first (see the Intel MP Spec version 1.4). It's really not possible, > unless you modify the ISR to record APIC ID, and reverse look it up > (an expensive operation) on each interrupt, to determine which CPU > is actually getting the interrupt. You are right, only LOAD, USER, NICE, SYS, IDLE are displayed, the other are always zero! > > I notice the other fields are not reported as well, probably for > similar reasons. > > > > Memory: 180344K (29336K) real, 256220K (66940K) virtual, 5160K free Page# 1/26 > > > > CPU TTY PID USERNAME PRI NI SIZE RES STATE TIME %WCPU %CPU COMM AND > > 0 ? 19703 mcfutz 251 25 632K 116K run 6:05 80.27 80.13 schl u > > 1 ? 19721 physik 251 25 632K 112K run 4:52 49.42 49.34 proc ess > > 1 ? 5375 plond 251 25 34756K 15900K run 2173:38 46.66 46.58 l502 .exe > > Pretty obviously, there aren't two running process on that one CPU. A > CPU can be in user space in only one process at a time. 8-). Grinnnn! > I think what they are doing, since they can tell you the CPU, is either > recording what CPU they last ran on, *or*, they are reporting which of > the multiple run queues that the program is on. > > ... snip ... > > Unfortunately, displaying this information is complicated, in that > you have to know what you are displaying... I see ... better I concentrate on porting xperfmon++ to libdevstat, to bring it up with the new CAM code! Thanks again and best wishes Lars -- E-Mail: | Lars Köller Lars.Koeller@Uni-Bielefeld.DE | UNIX Sysadmin lkoeller@cc.FH-Lippe.DE | Computing Center PGP-key: | University of Bielefeld http://www.nic.surfnet.nl/pgp/pks-toplev.html | Germany ----------- FreeBSD, what else? ---- http://www.freebsd.org ------------- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Aug 19 06:13:55 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id GAA07714 for freebsd-smp-outgoing; Wed, 19 Aug 1998 06:13:55 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from sos.freebsd.dk (sos.freebsd.dk [212.242.40.180]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id GAA07708; Wed, 19 Aug 1998 06:13:49 -0700 (PDT) (envelope-from sos@sos.freebsd.dk) Received: (from sos@localhost) by sos.freebsd.dk (8.9.1/8.8.8) id PAA10187; Wed, 19 Aug 1998 15:03:40 +0200 (CEST) (envelope-from sos) Message-Id: <199808191303.PAA10187@sos.freebsd.dk> Subject: Re: softupdates and smp crash In-Reply-To: <19980818161335.30289@deepo.prosa.dk> from Philippe Regnauld at "Aug 18, 98 04:13:35 pm" To: regnauld@deepo.prosa.dk (Philippe Regnauld) Date: Wed, 19 Aug 1998 15:03:40 +0200 (CEST) Cc: croot@btp1da.phy.uni-bayreuth.de, current@FreeBSD.ORG, smp@FreeBSD.ORG From: Søren Schmidt Reply-to: sos@FreeBSD.ORG X-Mailer: ELM [version 2.4ME+ PL38 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org In reply to Philippe Regnauld who wrote: > #0 boot (howto=260) at ../../kern/kern_shutdown.c:286 > #1 0xf01151c1 in panic (fmt=0xf01ab861 "softdep_lock: locking against myself") > at ../../kern/kern_shutdown.c:429 > #2 0xf01ab8c9 in acquire_lock (lk=0xf02265ac) > at ../../ufs/ffs/ffs_softdep.c:263 > +-> #3 0xf01aee54 in initiate_write_filepage (pagedep=0xf0c49e00, bp=0xf35300b0) > | at ../../ufs/ffs/ffs_softdep.c:2745 > | #4 0xf01aecbf in softdep_disk_io_initiation (bp=0xf35300b0) > | > I crashed here. I tried it on my play-smp machine it panics also... We can still say for sure that softupdates are NOT SMP safe... -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Søren Schmidt (sos@FreeBSD.org) FreeBSD Core Team Even more code to hack -- will it ever end? .. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Aug 19 06:40:33 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id GAA11014 for freebsd-smp-outgoing; Wed, 19 Aug 1998 06:40:33 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from btp1da.phy.uni-bayreuth.de (btp1da.phy.uni-bayreuth.de [132.180.20.32]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id GAA10986; Wed, 19 Aug 1998 06:40:25 -0700 (PDT) (envelope-from croot@btp1da.phy.uni-bayreuth.de) Received: (from root@localhost) by btp1da.phy.uni-bayreuth.de (8.8.8/8.7.3) id PAA11932; Wed, 19 Aug 1998 15:39:47 +0200 (MEST) Message-ID: X-Mailer: XFMail 1.3 [p0] on FreeBSD X-Priority: 3 (Normal) Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 8bit MIME-Version: 1.0 In-Reply-To: <199808191303.PAA10187@sos.freebsd.dk> Date: Wed, 19 Aug 1998 15:39:47 +0200 (MEST) Organization: university of bayreuth From: Werner Griessl To: "S\xren Schmidt" Subject: Re: softupdates and smp crash Cc: smp@FreeBSD.ORG, current@FreeBSD.ORG, (Philippe Regnauld) Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On 19-Aug-98 S\xren Schmidt wrote: > In reply to Philippe Regnauld who wrote: >> #0 boot (howto=260) at ../../kern/kern_shutdown.c:286 >> #1 0xf01151c1 in panic (fmt=0xf01ab861 "softdep_lock: locking against >> #myself") >> at ../../kern/kern_shutdown.c:429 >> #2 0xf01ab8c9 in acquire_lock (lk=0xf02265ac) >> at ../../ufs/ffs/ffs_softdep.c:263 >> +-> #3 0xf01aee54 in initiate_write_filepage (pagedep=0xf0c49e00, >> bp=0xf35300b0) >> | at ../../ufs/ffs/ffs_softdep.c:2745 >> | #4 0xf01aecbf in softdep_disk_io_initiation (bp=0xf35300b0) >> | >> I crashed here. > > > I tried it on my play-smp machine it panics also... > We can still say for sure that softupdates are NOT SMP safe... > > -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- > Søren Schmidt (sos@FreeBSD.org) FreeBSD Cor ---------------------------------- E-Mail: Werner Griessl Date: 19-Aug-98 Time: 15:37:25 This message was sent by XFMail ---------------------------------- But with current from today (Aug. 19) i cannot reproduce the crash. Tried 3 times buildworld -j4 with softupdats enabled without problems. Werner To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Aug 19 06:43:48 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id GAA11986 for freebsd-smp-outgoing; Wed, 19 Aug 1998 06:43:48 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from mail.ftf.dk (mail.ftf.dk [129.142.64.2]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id GAA11980; Wed, 19 Aug 1998 06:43:46 -0700 (PDT) (envelope-from regnauld@deepo.prosa.dk) Received: from mail.prosa.dk ([192.168.100.254]) by mail.ftf.dk (8.8.8/8.8.8/gw-ftf-1.0) with ESMTP id PAA06606; Wed, 19 Aug 1998 15:48:46 +0200 (CEST) (envelope-from regnauld@deepo.prosa.dk) Received: from deepo.prosa.dk (deepo.prosa.dk [192.168.100.10]) by mail.prosa.dk (8.8.8/8.8.5/prosa-1.1) with ESMTP id PAA00776; Wed, 19 Aug 1998 15:51:25 +0200 (CEST) Received: (from regnauld@localhost) by deepo.prosa.dk (8.8.8/8.8.5/prosa-1.1) id PAA00981; Wed, 19 Aug 1998 15:42:56 +0200 (CEST) Message-ID: <19980819154255.54405@deepo.prosa.dk> Date: Wed, 19 Aug 1998 15:42:55 +0200 From: Philippe Regnauld To: Werner Griessl Cc: Sxren Schmidt , smp@FreeBSD.ORG, current@FreeBSD.ORG Subject: Re: softupdates and smp crash References: <199808191303.PAA10187@sos.freebsd.dk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.88e In-Reply-To: ; from Werner Griessl on Wed, Aug 19, 1998 at 03:39:47PM +0200 X-Operating-System: FreeBSD 2.2.6-RELEASE i386 Phone: +45 3336 4148 Address: Ahlefeldtsgade 16, 1359 Copenhagen K, Denmark Organization: PROSA Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Werner Griessl writes: > > >> +-> #3 0xf01aee54 in initiate_write_filepage (pagedep=0xf0c49e00, > >> bp=0xf35300b0) > >> | at ../../ufs/ffs/ffs_softdep.c:2745 > >> | #4 0xf01aecbf in softdep_disk_io_initiation (bp=0xf35300b0) > >> | > >> I crashed here. > > > > But with current from today (Aug. 19) i cannot reproduce the crash. > Tried 3 times buildworld -j4 with softupdats enabled without problems. > Werner I also have AHC_TAGENABLE turned on. (no CAM). I will try to turn it off tonight and see. -- -[ Philippe Regnauld / sysadmin / regnauld@deepo.prosa.dk / +55.4N +11.3E ]- The Internet is busy. Please try again later. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Aug 19 06:46:17 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id GAA12793 for freebsd-smp-outgoing; Wed, 19 Aug 1998 06:46:17 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from sos.freebsd.dk (sos.freebsd.dk [212.242.40.180]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id GAA12769; Wed, 19 Aug 1998 06:46:04 -0700 (PDT) (envelope-from sos@sos.freebsd.dk) Received: (from sos@localhost) by sos.freebsd.dk (8.9.1/8.8.8) id PAA10312; Wed, 19 Aug 1998 15:44:19 +0200 (CEST) (envelope-from sos) Message-Id: <199808191344.PAA10312@sos.freebsd.dk> Subject: Re: softupdates and smp crash In-Reply-To: from Werner Griessl at "Aug 19, 98 03:39:47 pm" To: croot@btp1da.phy.uni-bayreuth.de (Werner Griessl) Date: Wed, 19 Aug 1998 15:44:19 +0200 (CEST) Cc: sos@FreeBSD.ORG, smp@FreeBSD.ORG, current@FreeBSD.ORG, regnauld@deepo.prosa.dk From: Søren Schmidt Reply-to: sos@FreeBSD.ORG X-Mailer: ELM [version 2.4ME+ PL38 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org In reply to Werner Griessl who wrote: [lots about smp + softupdates crash] > But with current from today (Aug. 19) i cannot reproduce the crash. > Tried 3 times buildworld -j4 with softupdats enabled without problems. > Werner Hmm, I dont see where this should have been fixed, you should try at least -j8, there it crashed reliably for me, -j4 only crashed now and then. I'll try again tonight with a fresh src tree... -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Søren Schmidt (sos@FreeBSD.org) FreeBSD Core Team Even more code to hack -- will it ever end? .. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Aug 19 06:53:29 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id GAA14068 for freebsd-smp-outgoing; Wed, 19 Aug 1998 06:53:29 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from sos.freebsd.dk (sos.freebsd.dk [212.242.40.180]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id GAA14039; Wed, 19 Aug 1998 06:53:15 -0700 (PDT) (envelope-from sos@sos.freebsd.dk) Received: (from sos@localhost) by sos.freebsd.dk (8.9.1/8.8.8) id PAA10345; Wed, 19 Aug 1998 15:51:35 +0200 (CEST) (envelope-from sos) Message-Id: <199808191351.PAA10345@sos.freebsd.dk> Subject: Re: softupdates and smp crash In-Reply-To: <19980819154255.54405@deepo.prosa.dk> from Philippe Regnauld at "Aug 19, 98 03:42:55 pm" To: regnauld@deepo.prosa.dk (Philippe Regnauld) Date: Wed, 19 Aug 1998 15:51:35 +0200 (CEST) Cc: croot@btp1da.phy.uni-bayreuth.de, sos@FreeBSD.ORG, smp@FreeBSD.ORG, current@FreeBSD.ORG From: Søren Schmidt Reply-to: sos@FreeBSD.ORG X-Mailer: ELM [version 2.4ME+ PL38 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org In reply to Philippe Regnauld who wrote: > > But with current from today (Aug. 19) i cannot reproduce the crash. > > Tried 3 times buildworld -j4 with softupdats enabled without problems. > > Werner > > I also have AHC_TAGENABLE turned on. (no CAM). I will try to > turn it off tonight and see. Wild shot but not the (total) reason as I run it on an EIDE system and it still crashes.... -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Søren Schmidt (sos@FreeBSD.org) FreeBSD Core Team Even more code to hack -- will it ever end? .. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Aug 19 06:57:31 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id GAA14829 for freebsd-smp-outgoing; Wed, 19 Aug 1998 06:57:31 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from btp1da.phy.uni-bayreuth.de (btp1da.phy.uni-bayreuth.de [132.180.20.32]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id GAA14776; Wed, 19 Aug 1998 06:57:18 -0700 (PDT) (envelope-from croot@btp1da.phy.uni-bayreuth.de) Received: (from root@localhost) by btp1da.phy.uni-bayreuth.de (8.8.8/8.7.3) id PAA12221; Wed, 19 Aug 1998 15:56:36 +0200 (MEST) Message-ID: X-Mailer: XFMail 1.3 [p0] on FreeBSD X-Priority: 3 (Normal) Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 8bit MIME-Version: 1.0 In-Reply-To: <199808191344.PAA10312@sos.freebsd.dk> Date: Wed, 19 Aug 1998 15:56:36 +0200 (MEST) Organization: university of bayreuth From: Werner Griessl To: "S\xren Schmidt" Subject: Re: softupdates and smp crash Cc: regnauld@deepo.prosa.dk, current@FreeBSD.ORG, smp@FreeBSD.ORG Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On 19-Aug-98 S\xren Schmidt wrote: > In reply to Werner Griessl who wrote: > > [lots about smp + softupdates crash] > >> But with current from today (Aug. 19) i cannot reproduce the crash. >> Tried 3 times buildworld -j4 with softupdats enabled without problems. >> Werner > > Hmm, I dont see where this should have been fixed, you should try > at least -j8, there it crashed reliably for me, -j4 only crashed > now and then. I'll try again tonight with a fresh src tree... > > -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- > Søren Schmidt (sos@FreeBSD.org) FreeBSD Core Team > Even more code to hack -- will it ever end? > .. I'll try -j8 . Werner To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Aug 19 07:04:43 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id HAA16003 for freebsd-smp-outgoing; Wed, 19 Aug 1998 07:04:43 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from btp1da.phy.uni-bayreuth.de (btp1da.phy.uni-bayreuth.de [132.180.20.32]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id HAA15988; Wed, 19 Aug 1998 07:04:31 -0700 (PDT) (envelope-from croot@btp1da.phy.uni-bayreuth.de) Received: (from root@localhost) by btp1da.phy.uni-bayreuth.de (8.8.8/8.7.3) id QAA12332; Wed, 19 Aug 1998 16:03:51 +0200 (MEST) Message-ID: X-Mailer: XFMail 1.3 [p0] on FreeBSD X-Priority: 3 (Normal) Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 8bit MIME-Version: 1.0 In-Reply-To: <199808191351.PAA10345@sos.freebsd.dk> Date: Wed, 19 Aug 1998 16:03:51 +0200 (MEST) Organization: university of bayreuth From: Werner Griessl To: "S\xren Schmidt" Subject: Re: softupdates and smp crash Cc: current@FreeBSD.ORG, smp@FreeBSD.ORG, (Philippe Regnauld) Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On 19-Aug-98 S\xren Schmidt wrote: > In reply to Philippe Regnauld who wrote: >> > But with current from today (Aug. 19) i cannot reproduce the crash. >> > Tried 3 times buildworld -j4 with softupdats enabled without problems. >> > Werner >> >> I also have AHC_TAGENABLE turned on. (no CAM). I will try to >> turn it off tonight and see. > > > Wild shot but not the (total) reason as I run it on an EIDE system > and it still crashes.... > > -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- > Søren Schmidt (sos@FreeBSD.org) FreeBSD Cor ---------------------------------- I read from EIDE (/usr/src) and write to SCSI (/usr/obj), ncr810 without AHC_TAGENABLE. Werner To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Aug 19 07:16:47 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id HAA17831 for freebsd-smp-outgoing; Wed, 19 Aug 1998 07:16:47 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from btp1da.phy.uni-bayreuth.de (btp1da.phy.uni-bayreuth.de [132.180.20.32]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id HAA17810; Wed, 19 Aug 1998 07:16:31 -0700 (PDT) (envelope-from croot@btp1da.phy.uni-bayreuth.de) Received: (from root@localhost) by btp1da.phy.uni-bayreuth.de (8.8.8/8.7.3) id QAA12621; Wed, 19 Aug 1998 16:15:43 +0200 (MEST) Message-ID: X-Mailer: XFMail 1.3 [p0] on FreeBSD X-Priority: 3 (Normal) Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 8bit MIME-Version: 1.0 In-Reply-To: <199808191344.PAA10312@sos.freebsd.dk> Date: Wed, 19 Aug 1998 16:15:43 +0200 (MEST) Organization: university of bayreuth From: Werner Griessl To: "S\xren Schmidt" Subject: Re: softupdates and smp crash Cc: regnauld@deepo.prosa.dk, current@FreeBSD.ORG, smp@FreeBSD.ORG Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On 19-Aug-98 S\xren Schmidt wrote: > In reply to Werner Griessl who wrote: > > [lots about smp + softupdates crash] > >> But with current from today (Aug. 19) i cannot reproduce the crash. >> Tried 3 times buildworld -j4 with softupdats enabled without problems. >> Werner > > Hmm, I dont see where this should have been fixed, you should try > at least -j8, there it crashed reliably for me, -j4 only crashed > now and then. I'll try again tonight with a fresh src tree... > > -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- > Søren Schmidt (sos@FreeBSD.org) FreeBSD Cor ---------------------------------- E-Mail: Werner Griessl Date: 19-Aug-98 Time: 16:14:55 This message was sent by XFMail ---------------------------------- You are right, crash after 10 mins with -j8 ! Werner To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Aug 19 07:54:39 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id HAA23497 for freebsd-smp-outgoing; Wed, 19 Aug 1998 07:54:39 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from sos.freebsd.dk (sos.freebsd.dk [212.242.40.180]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id HAA23473; Wed, 19 Aug 1998 07:54:30 -0700 (PDT) (envelope-from sos@sos.freebsd.dk) Received: (from sos@localhost) by sos.freebsd.dk (8.9.1/8.8.8) id QAA10472; Wed, 19 Aug 1998 16:52:42 +0200 (CEST) (envelope-from sos) Message-Id: <199808191452.QAA10472@sos.freebsd.dk> Subject: Re: softupdates and smp crash In-Reply-To: from Werner Griessl at "Aug 19, 98 04:15:43 pm" To: croot@btp1da.phy.uni-bayreuth.de (Werner Griessl) Date: Wed, 19 Aug 1998 16:52:42 +0200 (CEST) Cc: sos@FreeBSD.ORG, regnauld@deepo.prosa.dk, current@FreeBSD.ORG, smp@FreeBSD.ORG From: Søren Schmidt Reply-to: sos@FreeBSD.ORG X-Mailer: ELM [version 2.4ME+ PL38 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org In reply to Werner Griessl who wrote: > > > > Hmm, I dont see where this should have been fixed, you should try > > at least -j8, there it crashed reliably for me, -j4 only crashed > > now and then. I'll try again tonight with a fresh src tree... > > > > You are right, crash after 10 mins with -j8 ! > Werner So we can still say that softupdates are broken, at least for SMP. -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Søren Schmidt (sos@FreeBSD.org) FreeBSD Core Team Even more code to hack -- will it ever end? .. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Aug 19 08:05:58 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id IAA25387 for freebsd-smp-outgoing; Wed, 19 Aug 1998 08:05:58 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from lor.watermarkgroup.com (lor.watermarkgroup.com [207.202.73.33]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id IAA25371; Wed, 19 Aug 1998 08:05:56 -0700 (PDT) (envelope-from luoqi@watermarkgroup.com) Received: (from luoqi@localhost) by lor.watermarkgroup.com (8.8.8/8.8.8) id LAA05903; Wed, 19 Aug 1998 11:04:33 -0400 (EDT) (envelope-from luoqi) Date: Wed, 19 Aug 1998 11:04:33 -0400 (EDT) From: Luoqi Chen Message-Id: <199808191504.LAA05903@lor.watermarkgroup.com> To: regnauld@deepo.prosa.dk, sos@FreeBSD.ORG Subject: Re: softupdates and smp crash Cc: croot@btp1da.phy.uni-bayreuth.de, current@FreeBSD.ORG, smp@FreeBSD.ORG Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > In reply to Philippe Regnauld who wrote: > > #0 boot (howto=260) at ../../kern/kern_shutdown.c:286 > > #1 0xf01151c1 in panic (fmt=0xf01ab861 "softdep_lock: locking against myself") > > at ../../kern/kern_shutdown.c:429 > > #2 0xf01ab8c9 in acquire_lock (lk=0xf02265ac) > > at ../../ufs/ffs/ffs_softdep.c:263 > > +-> #3 0xf01aee54 in initiate_write_filepage (pagedep=0xf0c49e00, bp=0xf35300b0) > > | at ../../ufs/ffs/ffs_softdep.c:2745 > > | #4 0xf01aecbf in softdep_disk_io_initiation (bp=0xf35300b0) > > | > > I crashed here. > > > I tried it on my play-smp machine it panics also... > We can still say for sure that softupdates are NOT SMP safe... > > -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- > Søren Schmidt (sos@FreeBSD.org) FreeBSD Core Team > Even more code to hack -- will it ever end? > .. > There is a race window between a directory entry is removed and dependency for that directory entry is deallocated. The reason that the race window has manifested itself on SMP is that the syncer process is a kernel thread, so we can have two processes both running in kernel mode. The simplest solution for now, IMO, is to make syncer process honor the giant lock. The long term solution is, of course, close those race windows one by one. -lq To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Aug 19 10:58:48 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id KAA25062 for freebsd-smp-outgoing; Wed, 19 Aug 1998 10:58:48 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from sos.freebsd.dk (sos.freebsd.dk [212.242.40.180]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id KAA25056; Wed, 19 Aug 1998 10:58:42 -0700 (PDT) (envelope-from sos@sos.freebsd.dk) Received: (from sos@localhost) by sos.freebsd.dk (8.9.1/8.8.8) id TAA00646; Wed, 19 Aug 1998 19:57:01 +0200 (CEST) (envelope-from sos) Message-Id: <199808191757.TAA00646@sos.freebsd.dk> Subject: Re: softupdates and smp crash In-Reply-To: <199808191504.LAA05903@lor.watermarkgroup.com> from Luoqi Chen at "Aug 19, 98 11:04:33 am" To: luoqi@watermarkgroup.com (Luoqi Chen) Date: Wed, 19 Aug 1998 19:57:01 +0200 (CEST) Cc: regnauld@deepo.prosa.dk, sos@FreeBSD.ORG, croot@btp1da.phy.uni-bayreuth.de, current@FreeBSD.ORG, smp@FreeBSD.ORG From: Søren Schmidt Reply-to: sos@FreeBSD.ORG X-Mailer: ELM [version 2.4ME+ PL38 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org In reply to Luoqi Chen who wrote: > There is a race window between a directory entry is removed and dependency > for that directory entry is deallocated. The reason that the race window > has manifested itself on SMP is that the syncer process is a kernel thread, > so we can have two processes both running in kernel mode. The simplest > solution for now, IMO, is to make syncer process honor the giant lock. > The long term solution is, of course, close those race windows one by one. Care to fix it, the giant lock method should be fine for now ?? -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Søren Schmidt (sos@FreeBSD.org) FreeBSD Core Team Even more code to hack -- will it ever end? .. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Aug 19 15:15:49 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id PAA14605 for freebsd-smp-outgoing; Wed, 19 Aug 1998 15:15:49 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from smtp.algonet.se (angel.algonet.se [194.213.74.112]) by hub.freebsd.org (8.8.8/8.8.8) with SMTP id PAA14583 for ; Wed, 19 Aug 1998 15:15:38 -0700 (PDT) (envelope-from mal@algonet.se) Received: (qmail 14594 invoked from network); 19 Aug 1998 22:14:57 -0000 Received: from kairos.algonet.se (HELO kairos) (194.213.74.18) by angel.algonet.se with SMTP; 19 Aug 1998 22:14:57 -0000 Received: (mal@localhost) by kairos (SMI-8.6/8.6.12) id AAA29928; Thu, 20 Aug 1998 00:14:56 +0200 Date: Thu, 20 Aug 1998 00:14:56 +0200 Message-Id: <199808192214.AAA29928@kairos> From: Mats Lofkvist To: sos@FreeBSD.ORG CC: croot@btp1da.phy.uni-bayreuth.de, sos@FreeBSD.ORG, regnauld@deepo.prosa.dk, current@FreeBSD.ORG, smp@FreeBSD.ORG In-reply-to: <199808191452.QAA10472@sos.freebsd.dk> Subject: Re: softupdates and smp crash References: <199808191452.QAA10472@sos.freebsd.dk> Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > > > Hmm, I dont see where this should have been fixed, you should try > > at least -j8, there it crashed reliably for me, -j4 only crashed > > now and then. I'll try again tonight with a fresh src tree... > > > > You are right, crash after 10 mins with -j8 ! > Werner So we can still say that softupdates are broken, at least for SMP. Is this a new problem? I'm running current from July 21 on a dual ppro with softupdates and ahc_tagenable and have had no crashes. (I have run a few -j4 buildworlds and just completed a -j8 ditto.) _ Mats Lofkvist mal@algonet.se To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Aug 19 18:01:15 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id SAA13708 for freebsd-smp-outgoing; Wed, 19 Aug 1998 18:01:15 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from smtp01.primenet.com (smtp01.primenet.com [206.165.6.131]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id SAA13683; Wed, 19 Aug 1998 18:01:04 -0700 (PDT) (envelope-from tlambert@usr05.primenet.com) Received: (from daemon@localhost) by smtp01.primenet.com (8.8.8/8.8.8) id SAA15261; Wed, 19 Aug 1998 18:00:26 -0700 (MST) Received: from usr05.primenet.com(206.165.6.205) via SMTP by smtp01.primenet.com, id smtpd015210; Wed Aug 19 18:00:21 1998 Received: (from tlambert@localhost) by usr05.primenet.com (8.8.5/8.8.5) id SAA01161; Wed, 19 Aug 1998 18:00:17 -0700 (MST) From: Terry Lambert Message-Id: <199808200100.SAA01161@usr05.primenet.com> Subject: Re: softupdates and smp crash To: sos@FreeBSD.ORG Date: Thu, 20 Aug 1998 01:00:17 +0000 (GMT) Cc: croot@btp1da.phy.uni-bayreuth.de, regnauld@deepo.prosa.dk, current@FreeBSD.ORG, smp@FreeBSD.ORG In-Reply-To: <199808191452.QAA10472@sos.freebsd.dk> from "S?ren Schmidt" at Aug 19, 98 04:52:42 pm X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > You are right, crash after 10 mins with -j8 ! > > Werner > > So we can still say that softupdates are broken, at least for SMP. It would be more useful (than saying this) for people to characterize the problem they are getting so that we can find a discriminitory value. I suggest: Result: crash/stable Disk1: IDE/EIDE/SCSI Disk2: IDE/EIDE/SCSI ... DiskN: IDE/EIDE/SCSI CAM: Yes/No AHC_TAGENABLE: Yes/No /tmp is MFS Yes/No Processor: Pentium/P2/etc. SMP BIOS: 1.1/1.4 APIC INT: Yes/No Motherboard: ASUS/Compaq/HP/etc. I also suggest using the SMP list for this instead of -current. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Aug 19 23:31:26 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id XAA01174 for freebsd-smp-outgoing; Wed, 19 Aug 1998 23:31:26 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from sos.freebsd.dk (sos.freebsd.dk [212.242.40.180]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id XAA01166; Wed, 19 Aug 1998 23:31:22 -0700 (PDT) (envelope-from sos@sos.freebsd.dk) Received: (from sos@localhost) by sos.freebsd.dk (8.9.1/8.8.8) id IAA02030; Thu, 20 Aug 1998 08:29:44 +0200 (CEST) (envelope-from sos) Message-Id: <199808200629.IAA02030@sos.freebsd.dk> Subject: Re: softupdates and smp crash In-Reply-To: <199808200100.SAA01161@usr05.primenet.com> from Terry Lambert at "Aug 20, 98 01:00:17 am" To: tlambert@primenet.com (Terry Lambert) Date: Thu, 20 Aug 1998 08:29:44 +0200 (CEST) Cc: sos@FreeBSD.ORG, croot@btp1da.phy.uni-bayreuth.de, regnauld@deepo.prosa.dk, current@FreeBSD.ORG, smp@FreeBSD.ORG From: Søren Schmidt Reply-to: sos@FreeBSD.ORG X-Mailer: ELM [version 2.4ME+ PL38 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org In reply to Terry Lambert who wrote: > > > You are right, crash after 10 mins with -j8 ! > > > Werner > > > > So we can still say that softupdates are broken, at least for SMP. > > It would be more useful (than saying this) for people to characterize > the problem they are getting so that we can find a discriminitory > value. I suggest: I suggest reading Luoqi Chen's exelent explanation as why it breaks under SMP (hint: it has nothing todo with HW setups).. -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Søren Schmidt (sos@FreeBSD.org) FreeBSD Core Team Even more code to hack -- will it ever end? .. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Aug 19 23:42:56 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id XAA02648 for freebsd-smp-outgoing; Wed, 19 Aug 1998 23:42:56 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from smtp04.primenet.com (smtp04.primenet.com [206.165.6.134]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id XAA02627; Wed, 19 Aug 1998 23:42:47 -0700 (PDT) (envelope-from tlambert@usr02.primenet.com) Received: (from daemon@localhost) by smtp04.primenet.com (8.8.8/8.8.8) id XAA08274; Wed, 19 Aug 1998 23:42:09 -0700 (MST) Received: from usr02.primenet.com(206.165.6.202) via SMTP by smtp04.primenet.com, id smtpd008231; Wed Aug 19 23:42:02 1998 Received: (from tlambert@localhost) by usr02.primenet.com (8.8.5/8.8.5) id XAA18028; Wed, 19 Aug 1998 23:41:54 -0700 (MST) From: Terry Lambert Message-Id: <199808200641.XAA18028@usr02.primenet.com> Subject: Re: softupdates and smp crash To: sos@FreeBSD.ORG Date: Thu, 20 Aug 1998 06:41:53 +0000 (GMT) Cc: tlambert@primenet.com, croot@btp1da.phy.uni-bayreuth.de, regnauld@deepo.prosa.dk, current@FreeBSD.ORG, smp@FreeBSD.ORG In-Reply-To: <199808200629.IAA02030@sos.freebsd.dk> from "S?ren Schmidt" at Aug 20, 98 08:29:44 am X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > > So we can still say that softupdates are broken, at least for SMP. > > > > It would be more useful (than saying this) for people to characterize > > the problem they are getting so that we can find a discriminitory > > value. I suggest: > > I suggest reading Luoqi Chen's exelent explanation as why it breaks > under SMP (hint: it has nothing todo with HW setups).. Perhaps clock speed? 8-). His post arrived in my mailbox after I responded to your post... I'm also not convinced that this is the only possible cause of the problem; the VM code is hardly "assert" protected everywhere, so diagnosing this thing is not trivial. Look at the VM fixes I recently did, which killed the bugs Karl Denniger was seeing in 75% of the cases, leaving 25% of the cases "clustered" (in his words), indicating a seperate problem, in addition to the ones I fixed, in a periodically executing code path. I had suspected that this would be the case when I made the fix, since it doesn't account for the buggy behaviour I'm personally seeing. 8-(. I myself haven't seen the problems described on my dual P90 system. I suspect the people seeing the bug Luoqi pointed to have very fast processors in order to have one "win" the race. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Aug 20 07:52:11 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id HAA23267 for freebsd-smp-outgoing; Thu, 20 Aug 1998 07:52:11 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from thor.tvol.net (mail.tvol.com [38.219.83.4]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id HAA23238 for ; Thu, 20 Aug 1998 07:52:04 -0700 (PDT) (envelope-from agrosky@wgate.com) Received: from dadu.eng.tvol.net (dadu.eng.tvol.net [10.32.2.13]) by thor.tvol.net (8.8.8/8.8.3) with SMTP id KAA13447; Thu, 20 Aug 1998 10:50:29 -0400 (EDT) Date: Thu, 20 Aug 1998 10:49:18 -0400 (EDT) From: Aaron Grosky X-Sender: asg@dadu.eng.tvol.net Reply-To: Aaron Grosky To: Lars =?iso-8859-1?Q?K=F6ller?= cc: Terry Lambert , chuckr@glue.umd.edu, freebsd-smp@FreeBSD.ORG Subject: Re: Per processor load? In-Reply-To: <199808190540.FAA24476@mitch.hrz.uni-bielefeld.de> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from QUOTED-PRINTABLE to 8bit by hub.freebsd.org id HAA23263 Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Hello Lars, Terry, ... I spent many years working with Cray's SMP UNICOS. On Wed, 19 Aug 1998, Lars [iso-8859-1] Köller wrote: > Hello Terry! > > First of all thanks for the large answer! > > In reply to Terry Lambert who wrote: > >>> ... >>> System: share Tue Aug 18 07:30:58 1998 >>> Load averages: 2.42, 2.29, 2.28 >>> 280 processes: 273 sleeping, 5 running, 2 zombies >>> Cpu states: >>> CPU LOAD USER NICE SYS IDLE BLOCK SWAIT INTR SSYS >>> 0 2.62 0.4% 97.6% 2.0% 0.0% 0.0% 0.0% 0.0% 0.0% >>> 1 2.22 0.8% 97.0% 2.2% 0.0% 0.0% 0.0% 0.0% 0.0% >>> --- ---- ----- ----- ----- ----- ----- ----- ----- ----- >>> avg 2.42 0.6% 97.2% 2.2% 0.0% 0.0% 0.0% 0.0% 0.0% > > Sorry, I forgot to mention this top was running on HPUX 10.20. > >> This basically implies a scheduler artifact; each CPU must have its >> own ready-to-run queue for you to get this statistic; ... A process which is not multi-threaded can only be runnable on one processor at a time. Therefore, given a mix of simple processes (not threaded) and threaded processes running on only one processor: the above example shows, as one possibility, processor 0 running a simple process for the previous load interval while processor 1 ran a simple process 60% of that time and a multi-threaded process 40% of the time. > ... > > Thanks again and best wishes > > Lars Aaron Aaron Grosky agrosky@wgate.com WorldGate Communications "Internet TV Over Cable" 3220 Tillman Drive 215-633-5125 Bensalem, PA 19020 215-633-9590 (FAX) To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Aug 21 22:24:32 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id WAA27577 for freebsd-smp-outgoing; Fri, 21 Aug 1998 22:24:32 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from dyson.iquest.net (dyson.iquest.net [198.70.144.127]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id WAA27566; Fri, 21 Aug 1998 22:24:27 -0700 (PDT) (envelope-from toor@dyson.iquest.net) Received: (from root@localhost) by dyson.iquest.net (8.8.8/8.8.8) id AAA19739; Sat, 22 Aug 1998 00:23:34 -0500 (EST) (envelope-from toor) Message-Id: <199808220523.AAA19739@dyson.iquest.net> Subject: Re: softupdates and smp crash In-Reply-To: <199808200641.XAA18028@usr02.primenet.com> from Terry Lambert at "Aug 20, 98 06:41:53 am" To: tlambert@primenet.com (Terry Lambert) Date: Sat, 22 Aug 1998 00:23:34 -0500 (EST) Cc: sos@FreeBSD.ORG, tlambert@primenet.com, croot@btp1da.phy.uni-bayreuth.de, regnauld@deepo.prosa.dk, current@FreeBSD.ORG, smp@FreeBSD.ORG From: "John S. Dyson" Reply-To: dyson@iquest.net X-Mailer: ELM [version 2.4ME+ PL38 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Terry Lambert said: > > > > So we can still say that softupdates are broken, at least for SMP. > > > > > > It would be more useful (than saying this) for people to characterize > > > the problem they are getting so that we can find a discriminitory > > > value. I suggest: > > > > I suggest reading Luoqi Chen's exelent explanation as why it breaks > > under SMP (hint: it has nothing todo with HW setups).. > > Perhaps clock speed? 8-). > > His post arrived in my mailbox after I responded to your post... > > I'm also not convinced that this is the only possible cause of > the problem; the VM code is hardly "assert" protected everywhere, > so diagnosing this thing is not trivial. Look at the VM fixes > I recently did, which killed the bugs Karl Denniger was seeing > in 75% of the cases, leaving 25% of the cases "clustered" (in his > words), indicating a seperate problem, in addition to the ones I > fixed, in a periodically executing code path. I had suspected > that this would be the case when I made the fix, since it doesn't > account for the buggy behaviour I'm personally seeing. 8-(. > I have to chime in here -- some of the "fixes" are work-arounds, and there are still underlying VM problems. It might be "good enough" for 3.0, but I would suggest preparing for some rework to find the root cause for the problem. -- John | Never try to teach a pig to sing, dyson@iquest.net | it makes one look stupid jdyson@nc.com | and it irritates the pig. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Sat Aug 22 08:11:55 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id IAA10260 for freebsd-smp-outgoing; Sat, 22 Aug 1998 08:11:55 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from alpo.whistle.com (alpo.whistle.com [207.76.204.38]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id IAA10254; Sat, 22 Aug 1998 08:11:54 -0700 (PDT) (envelope-from julian@whistle.com) Received: (from daemon@localhost) by alpo.whistle.com (8.8.5/8.8.5) id IAA01902; Sat, 22 Aug 1998 08:07:23 -0700 (PDT) Received: from current1.whistle.com(207.76.205.22) via SMTP by alpo.whistle.com, id smtpdVj1897; Sat Aug 22 15:07:14 1998 Date: Sat, 22 Aug 1998 08:07:12 -0700 (PDT) From: Julian Elischer To: "John S. Dyson" cc: Terry Lambert , sos@FreeBSD.ORG, croot@btp1da.phy.uni-bayreuth.de, regnauld@deepo.prosa.dk, current@FreeBSD.ORG, smp@FreeBSD.ORG Subject: Re: softupdates and smp crash In-Reply-To: <199808220523.AAA19739@dyson.iquest.net> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Sat, 22 Aug 1998, John S. Dyson wrote: > I have to chime in here -- some of the "fixes" are work-arounds, and > there are still underlying VM problems. It might be "good enough" > for 3.0, but I would suggest preparing for some rework to find the > root cause for the problem. > We all agree john, but we have not got a suitable set of people who really understand that code yet... several people are doing "voyages of discovery" but we do not yet have a "slot-in replacement" for you :-) julian To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Sat Aug 22 10:21:27 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id KAA20822 for freebsd-smp-outgoing; Sat, 22 Aug 1998 10:21:27 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from dyson.iquest.net (dyson.iquest.net [198.70.144.127]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id KAA20816; Sat, 22 Aug 1998 10:21:24 -0700 (PDT) (envelope-from toor@dyson.iquest.net) Received: (from root@localhost) by dyson.iquest.net (8.8.8/8.8.8) id MAA02475; Sat, 22 Aug 1998 12:20:27 -0500 (EST) (envelope-from toor) Message-Id: <199808221720.MAA02475@dyson.iquest.net> Subject: Re: softupdates and smp crash In-Reply-To: from Julian Elischer at "Aug 22, 98 08:07:12 am" To: julian@whistle.com (Julian Elischer) Date: Sat, 22 Aug 1998 12:20:27 -0500 (EST) Cc: dyson@iquest.net, tlambert@primenet.com, sos@FreeBSD.ORG, croot@btp1da.phy.uni-bayreuth.de, regnauld@deepo.prosa.dk, current@FreeBSD.ORG, smp@FreeBSD.ORG From: "John S. Dyson" Reply-To: dyson@iquest.net X-Mailer: ELM [version 2.4ME+ PL38 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Julian Elischer said: > > > On Sat, 22 Aug 1998, John S. Dyson wrote: > > > I have to chime in here -- some of the "fixes" are work-arounds, and > > there are still underlying VM problems. It might be "good enough" > > for 3.0, but I would suggest preparing for some rework to find the > > root cause for the problem. > > > > We all agree john, but we have not got a suitable set of people who really > understand that code yet... several people are doing "voyages of > discovery" but we do not yet have a "slot-in replacement" for you :-) > That wasn't clear. The "fix" to the page_remove code has really concerned me. The effort was good, but I am still thinking about where the problem really is. I wasn't meaning to put the effort down, and I think that Terry is a person who has a good chance of figuring the code out to the point of being creative with it. (Even though some-people think that he is sometimes "off the wall", I also don't think that some-people listen to him *enough*.) He has been in contact with me, and I am certainly willing to help him. Especially after meetings personally with him, I trust his ability to comprehend and create. (For VM code, I also believe that some of the Alpha people, and others who were helping even when I was actively working on the code, are able to take off where I left it.) None of the VM stuff is brain surgery, but there are some end cases that I believe have BOTH existed since day one, and also some problems that I might have caused. Those who are actually working on/with the code should feel very free to contact me. This strategy has apparently been fairly successful, with not only Terry, but others who have just chatted with me about other technical things. The locking situations and critical region problems are existant throughout the VFS and VM code (and probably in other places too.) I am working to help solve those sorts of problems coherently and completely in another forum. However, when I can help (without totally distracting me in my new work), I will (and do.) FreeBSD's success is still important to me, so using me as a resource when I can help effectively is okay. My current schedule is quite full, but use me when you need to. -- John | Never try to teach a pig to sing, dyson@iquest.net | it makes one look stupid jdyson@nc.com | and it irritates the pig. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Sat Aug 22 20:20:20 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id UAA13212 for freebsd-smp-outgoing; Sat, 22 Aug 1998 20:20:20 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from smtp02.primenet.com (smtp02.primenet.com [206.165.6.132]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id UAA13172; Sat, 22 Aug 1998 20:20:14 -0700 (PDT) (envelope-from tlambert@usr04.primenet.com) Received: (from daemon@localhost) by smtp02.primenet.com (8.8.8/8.8.8) id UAA04959; Sat, 22 Aug 1998 20:19:31 -0700 (MST) Received: from usr04.primenet.com(206.165.6.204) via SMTP by smtp02.primenet.com, id smtpd004906; Sat Aug 22 20:19:23 1998 Received: (from tlambert@localhost) by usr04.primenet.com (8.8.5/8.8.5) id UAA21616; Sat, 22 Aug 1998 20:19:17 -0700 (MST) From: Terry Lambert Message-Id: <199808230319.UAA21616@usr04.primenet.com> Subject: Re: softupdates and smp crash To: dyson@iquest.net Date: Sun, 23 Aug 1998 03:19:16 +0000 (GMT) Cc: tlambert@primenet.com, sos@FreeBSD.ORG, croot@btp1da.phy.uni-bayreuth.de, regnauld@deepo.prosa.dk, current@FreeBSD.ORG, smp@FreeBSD.ORG In-Reply-To: <199808220523.AAA19739@dyson.iquest.net> from "John S. Dyson" at Aug 22, 98 00:23:34 am X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > I'm also not convinced that this is the only possible cause of > > the problem; the VM code is hardly "assert" protected everywhere, > > so diagnosing this thing is not trivial. Look at the VM fixes > > I recently did, which killed the bugs Karl Denniger was seeing > > in 75% of the cases, leaving 25% of the cases "clustered" (in his > > words), indicating a seperate problem, in addition to the ones I > > fixed, in a periodically executing code path. I had suspected > > that this would be the case when I made the fix, since it doesn't > > account for the buggy behaviour I'm personally seeing. 8-(. > > > I have to chime in here -- some of the "fixes" are work-arounds, and > there are still underlying VM problems. It might be "good enough" > for 3.0, but I would suggest preparing for some rework to find the > root cause for the problem. Can you identify which of the "better fixes" are workarounds? The two fixes I have done, and now have enough confidence in to want them committed, are: o The "valid = 0 at wrong time" that you told me about. o The "setting the recorded size of a backing object to a page boundary instead of to the actual size". You could argue that this second, which promiscuously sets the vnode object size after instancing the object, is a workaround which should be repaired by adding a "real_size" parameter to the allocator, but the fact is that the setsize code path is not a problem at the only time when it is called (ie: it can't be called at interrupt level as a result of a disk I/O completion interrupt); so the window I noted has been analyzed, and is not there. The code is ugly, but it does the intended job, without side effects. The other "fix", the "back up one" is, indeed, a kludge that happens to work for some cases, but I would not want that one committed (I explicitly posted that it should be tried as a dianostic). The only other changes packaged with the two real changes, above, are panics in the diagnostic case, which is basically an "assert" that map contents aren't being stomped on page insertion, and a lock acquisition logging that was arguably erroneously missing. I haven't been able to get anyone to run with the "DIAGNOSTIC" flag to test the first nor the "MAP_LOCK_DIAGNOSTIC" for the second (but they run without error here, where I can't trigger the failures at will). Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Sat Aug 22 20:23:59 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id UAA13907 for freebsd-smp-outgoing; Sat, 22 Aug 1998 20:23:59 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from dyson.iquest.net (dyson.iquest.net [198.70.144.127]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id UAA13901; Sat, 22 Aug 1998 20:23:56 -0700 (PDT) (envelope-from toor@dyson.iquest.net) Received: (from root@localhost) by dyson.iquest.net (8.8.8/8.8.8) id WAA01043; Sat, 22 Aug 1998 22:22:58 -0500 (EST) (envelope-from toor) Message-Id: <199808230322.WAA01043@dyson.iquest.net> Subject: Re: softupdates and smp crash In-Reply-To: <199808221720.MAA02475@dyson.iquest.net> from "John S. Dyson" at "Aug 22, 98 12:20:27 pm" To: dyson@iquest.net Date: Sat, 22 Aug 1998 22:22:58 -0500 (EST) Cc: julian@whistle.com, dyson@iquest.net, tlambert@primenet.com, sos@FreeBSD.ORG, croot@btp1da.phy.uni-bayreuth.de, regnauld@deepo.prosa.dk, current@FreeBSD.ORG, smp@FreeBSD.ORG From: "John S. Dyson" Reply-To: dyson@iquest.net X-Mailer: ELM [version 2.4ME+ PL38 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org John S. Dyson said: > completely in another forum. However, when I can help (without totally > distracting me in my new work), I will (and do.) > > FreeBSD's success is still important to me, so using me as a resource > when I can help effectively is okay. My current schedule is quite full, > but use me when you need to. > Following up on my posting: I am porting my (SMP) patches to -current for my own use. I have found a few bugs and incompatibilites, bug will make the patches available. (I want to upgrade to a more current -current for stability and feature reasons, but don't want to loose any of the performance features that I have been using.) I will be leaving out the improved vnode dirty buffer queuing (which improves sync times, with non-softupdate filesystems.) -- John | Never try to teach a pig to sing, dyson@iquest.net | it makes one look stupid jdyson@nc.com | and it irritates the pig. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Sat Aug 22 20:27:03 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id UAA14519 for freebsd-smp-outgoing; Sat, 22 Aug 1998 20:27:03 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from dyson.iquest.net (dyson.iquest.net [198.70.144.127]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id UAA14489; Sat, 22 Aug 1998 20:26:52 -0700 (PDT) (envelope-from toor@dyson.iquest.net) Received: (from root@localhost) by dyson.iquest.net (8.8.8/8.8.8) id WAA01058; Sat, 22 Aug 1998 22:26:04 -0500 (EST) (envelope-from toor) Message-Id: <199808230326.WAA01058@dyson.iquest.net> Subject: Re: softupdates and smp crash In-Reply-To: <199808230319.UAA21616@usr04.primenet.com> from Terry Lambert at "Aug 23, 98 03:19:16 am" To: tlambert@primenet.com (Terry Lambert) Date: Sat, 22 Aug 1998 22:26:04 -0500 (EST) Cc: dyson@iquest.net, tlambert@primenet.com, sos@FreeBSD.ORG, croot@btp1da.phy.uni-bayreuth.de, regnauld@deepo.prosa.dk, current@FreeBSD.ORG, smp@FreeBSD.ORG From: "John S. Dyson" Reply-To: dyson@iquest.net X-Mailer: ELM [version 2.4ME+ PL38 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Terry Lambert said: > > The other "fix", the "back up one" is, indeed, a kludge that happens > to work for some cases, but I would not want that one committed (I > explicitly posted that it should be tried as a dianostic). > That is what worried me. Since I don't track the cvs tree very closely any more, I was mistakenly under the false impression that change was adopted. Other than that, your stuff appears to be pretty good!!! As I posted (perhaps in another forum -- or not?!?!), I'll be making available an updated version of my SMP and pmap improvements (Since I don't want to loose anything.) It might be a good time for someone to look at the code seriously for bugs/etc, and perhaps adoption. -- John | Never try to teach a pig to sing, dyson@iquest.net | it makes one look stupid jdyson@nc.com | and it irritates the pig. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Sat Aug 22 20:33:36 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id UAA15795 for freebsd-smp-outgoing; Sat, 22 Aug 1998 20:33:36 -0700 (PDT) (envelope-from owner-freebsd-smp@FreeBSD.ORG) Received: from smtp02.primenet.com (smtp02.primenet.com [206.165.6.132]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id UAA15775; Sat, 22 Aug 1998 20:33:33 -0700 (PDT) (envelope-from tlambert@usr04.primenet.com) Received: (from daemon@localhost) by smtp02.primenet.com (8.8.8/8.8.8) id UAA07547; Sat, 22 Aug 1998 20:32:50 -0700 (MST) Received: from usr04.primenet.com(206.165.6.204) via SMTP by smtp02.primenet.com, id smtpd007505; Sat Aug 22 20:32:43 1998 Received: (from tlambert@localhost) by usr04.primenet.com (8.8.5/8.8.5) id UAA22340; Sat, 22 Aug 1998 20:32:35 -0700 (MST) From: Terry Lambert Message-Id: <199808230332.UAA22340@usr04.primenet.com> Subject: Re: softupdates and smp crash To: dyson@iquest.net Date: Sun, 23 Aug 1998 03:32:35 +0000 (GMT) Cc: julian@whistle.com, tlambert@primenet.com, sos@FreeBSD.ORG, croot@btp1da.phy.uni-bayreuth.de, regnauld@deepo.prosa.dk, current@FreeBSD.ORG, smp@FreeBSD.ORG In-Reply-To: <199808221720.MAA02475@dyson.iquest.net> from "John S. Dyson" at Aug 22, 98 12:20:27 pm X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > That wasn't clear. The "fix" to the page_remove code has really concerned > me. The effort was good, but I am still thinking about where the problem > really is. Are you talking about the initial diagnostic fix, which was never intended for integration (I would have squawked over a commit), the "backup one", or the "valid = 0" fix, based on your suggestion? I'm pretty sure that I understand the ramifications of the "valid = 0" fix enough that I think that it is "right". I'd like to put another panic in vm_page.c, where the kludge for "9-Jan-93" XXX; I think it's a kludge, but I doubt a panic there would make me very popular. 8-). > The locking situations and critical region problems are existant > throughout the VFS and VM code (and probably in other places too.) I > am working to help solve those sorts of problems coherently and > completely in another forum. However, when I can help (without totally > distracting me in my new work), I will (and do.) The locking stuff is noted; I have the copy of your diffs, and I'm slowly going through them with your list of what should and should not be taken, in hand. For the most part, these appear to be SMP issues, and unrelated to the problems I'm seeing. For what it's worth, I believe the two fixes I want in will do nothing for the problems I'm seeing; one was from you, and the other was from my own code review of what I felt was the place to look for the contact lens "because that's where the light is". David Greenamn made the suggestion that I look instead at the vfs_bio code for my own problems; that, of course, is "where the light isn't", but it's a good suggestion anyway, and I'm looking... 8-(. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message