From owner-freebsd-smp@FreeBSD.ORG Mon Dec 1 02:12:28 2003 Return-Path: Delivered-To: freebsd-smp@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 963FE16A4CE; Mon, 1 Dec 2003 02:12:28 -0800 (PST) Received: from park.rambler.ru (park.rambler.ru [81.19.64.101]) by mx1.FreeBSD.org (Postfix) with ESMTP id 5711843FB1; Mon, 1 Dec 2003 02:12:26 -0800 (PST) (envelope-from is@rambler-co.ru) Received: from is (is.park.rambler.ru [81.19.64.102]) by park.rambler.ru (8.12.6/8.12.6) with ESMTP id hB1ACNJ6040416; Mon, 1 Dec 2003 13:12:24 +0300 (MSK) (envelope-from is@rambler-co.ru) Date: Mon, 1 Dec 2003 13:12:23 +0300 (MSK) From: Igor Sysoev X-Sender: is@is To: John Baldwin Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII cc: freebsd-smp@freebsd.org Subject: halting HTT X-BeenThere: freebsd-smp@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: FreeBSD SMP implementation group List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 01 Dec 2003 10:12:28 -0000 I had noticed that on Nov 11, 2003 in src/sys/i386/i386/mp_machdep.c HTT CPUs are not halted by default. What is the reason to do it ? Is it not better to allow logical CPUs to handle interrupts ? Besides if I try to halting logical CPUs by machdep.hlt_cpus=10 or machdep.hlt_logical_cpus=1 then computer is partially freezed, i.e top that running in ssh session could contunue to run but I can not login to computer anymore. I'm using 5.1-CURRENT from 2003.11.28.00.00.00 and machdep.logical_cpus_mask=10 Igor Sysoev http://sysoev.ru/en/ From owner-freebsd-smp@FreeBSD.ORG Mon Dec 1 13:01:09 2003 Return-Path: Delivered-To: freebsd-smp@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 3E06E16A4CE for ; Mon, 1 Dec 2003 13:01:09 -0800 (PST) Received: from mail7.speakeasy.net (mail7.speakeasy.net [216.254.0.207]) by mx1.FreeBSD.org (Postfix) with ESMTP id F0AF443FEA for ; Mon, 1 Dec 2003 13:00:00 -0800 (PST) (envelope-from jhb@FreeBSD.org) Received: (qmail 3057 invoked from network); 1 Dec 2003 21:00:00 -0000 Received: from unknown (HELO server.baldwin.cx) ([216.27.160.63]) (envelope-sender )encrypted SMTP for ; 1 Dec 2003 21:00:00 -0000 Received: from laptop.baldwin.cx (gw1.twc.weather.com [216.133.140.1]) by server.baldwin.cx (8.12.9/8.12.9) with ESMTP id hB1KxuFn095798; Mon, 1 Dec 2003 15:59:57 -0500 (EST) (envelope-from jhb@FreeBSD.org) Message-ID: X-Mailer: XFMail 1.5.4 on FreeBSD X-Priority: 3 (Normal) Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 8bit MIME-Version: 1.0 In-Reply-To: Date: Mon, 01 Dec 2003 15:59:59 -0500 (EST) From: John Baldwin To: Igor Sysoev X-Spam-Checker-Version: SpamAssassin 2.55 (1.174.2.19-2003-05-19-exp) cc: freebsd-smp@freebsd.org Subject: RE: halting HTT X-BeenThere: freebsd-smp@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: FreeBSD SMP implementation group List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 01 Dec 2003 21:01:09 -0000 On 01-Dec-2003 Igor Sysoev wrote: > I had noticed that on Nov 11, 2003 in src/sys/i386/i386/mp_machdep.c > HTT CPUs are not halted by default. What is the reason to do it ? > Is it not better to allow logical CPUs to handle interrupts ? HTT CPUs were halted by default because we needed them to always be used to work around other bugs. Since the other bugs are fixed, we no longer need HTT CPUs unless the user wants them. If you disable HT in the BIOS for example, we no longer probe them. Thus, it doesn't make sense to halt the CPUs if the user has explicitly asked for them. > Besides if I try to halting logical CPUs by machdep.hlt_cpus=10 or > machdep.hlt_logical_cpus=1 then computer is partially freezed, i.e > top that running in ssh session could contunue to run but I can not > login to computer anymore. > > I'm using 5.1-CURRENT from 2003.11.28.00.00.00 and > machdep.logical_cpus_mask=10 I do not know why your machine freezes when you halt the logical CPUs. -- John Baldwin <>< http://www.FreeBSD.org/~jhb/ "Power Users Use the Power to Serve!" - http://www.FreeBSD.org/ From owner-freebsd-smp@FreeBSD.ORG Tue Dec 2 01:45:38 2003 Return-Path: Delivered-To: freebsd-smp@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 932EA16A4CE; Tue, 2 Dec 2003 01:45:38 -0800 (PST) Received: from park.rambler.ru (park.rambler.ru [81.19.64.101]) by mx1.FreeBSD.org (Postfix) with ESMTP id BCAC043FF2; Tue, 2 Dec 2003 01:45:36 -0800 (PST) (envelope-from is@rambler-co.ru) Received: from is (is.park.rambler.ru [81.19.64.102]) by park.rambler.ru (8.12.6/8.12.6) with ESMTP id hB29jTJ6073714; Tue, 2 Dec 2003 12:45:29 +0300 (MSK) (envelope-from is@rambler-co.ru) Date: Tue, 2 Dec 2003 12:45:28 +0300 (MSK) From: Igor Sysoev X-Sender: is@is To: John Baldwin In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII cc: Jeff Roberson cc: freebsd-smp@freebsd.org Subject: RE: halting HTT X-BeenThere: freebsd-smp@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: FreeBSD SMP implementation group List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 02 Dec 2003 09:45:38 -0000 On Mon, 1 Dec 2003, John Baldwin wrote: > On 01-Dec-2003 Igor Sysoev wrote: > > I had noticed that on Nov 11, 2003 in src/sys/i386/i386/mp_machdep.c > > HTT CPUs are not halted by default. What is the reason to do it ? > > Is it not better to allow logical CPUs to handle interrupts ? > > HTT CPUs were halted by default because we needed them to always be > used to work around other bugs. Since the other bugs are fixed, we > no longer need HTT CPUs unless the user wants them. If you disable > HT in the BIOS for example, we no longer probe them. Thus, it doesn't > make sense to halt the CPUs if the user has explicitly asked for > them. Thank you for explanation. But what is about a handling interrupts by logical CPUs ? Could you recommend it or not ? I think that a using logical CPUs decreases an overhead while handling interrupts - main CPUs do not need to switch contexts. > > Besides if I try to halting logical CPUs by machdep.hlt_cpus=10 or > > machdep.hlt_logical_cpus=1 then computer is partially freezed, i.e > > top that running in ssh session could contunue to run but I can not > > login to computer anymore. > > > > I'm using 5.1-CURRENT from 2003.11.28.00.00.00 and > > machdep.logical_cpus_mask=10 > > I do not know why your machine freezes when you halt the logical > CPUs. I think it was because of SCHED_ULE. When I rebuild kernel with SCHED_4BSD I can disable logical CPUs without freezing. I added Jeff Roberson to CC to notify him about SCHED_ULE freezing. Igor Sysoev http://sysoev.ru/en/ From owner-freebsd-smp@FreeBSD.ORG Wed Dec 3 14:42:56 2003 Return-Path: Delivered-To: freebsd-smp@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 83D3816A4CE for ; Wed, 3 Dec 2003 14:42:56 -0800 (PST) Received: from mikea.ath.cx (wsip-68-15-203-64.ok.ok.cox.net [68.15.203.64]) by mx1.FreeBSD.org (Postfix) with ESMTP id 5B75843F75 for ; Wed, 3 Dec 2003 14:42:55 -0800 (PST) (envelope-from mikea@mikea.ath.cx) Received: from mikea.ath.cx (bsd [192.168.1.1]) by mikea.ath.cx (8.12.3/8.12.3) with ESMTP id hB3Mgs89030795 for ; Wed, 3 Dec 2003 16:42:54 -0600 (CST) (envelope-from mikea@mikea.ath.cx) Received: (from mikea@localhost) by mikea.ath.cx (8.12.3/8.12.3/Submit) id hB3Mgq0G030794 for freebsd@mikea.ath.cx; Wed, 3 Dec 2003 16:42:52 -0600 (CST) (envelope-from mikea) Date: Wed, 3 Dec 2003 16:42:52 -0600 From: mikea To: freebsd@mikea.ath.cx Message-ID: <20031203164252.B30624@mikea.ath.cx> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5.1i X-Mailman-Approved-At: Thu, 04 Dec 2003 05:34:15 -0800 Subject: Request pointers to current-generation SMP boards X-BeenThere: freebsd-smp@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: FreeBSD SMP implementation group List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 03 Dec 2003 22:42:56 -0000 Hi. I'm speccing out a server for the day job, to run MailScanner and SpamAssassin on the inbound corporate mailstream. I hope to talk the boss into something that is _really_ muscular, in hope that the rising tide of spam, worms, etc., won't swamp the poor thing after 6 months to a year. With any luck, I'll be able to get something that doesn't have lots of integrated on-board controllers, so that I won't end up installing a card to take over the function after that part of the motherboard dies about 3 AM. Or do people find that integrated peripheral support on the motherboard isn't all that bad? The only things I really care about are that o it should support two or more pretty swift Intel P-IV CPUs, preferably the ones with lots of on-board cache, o it should have lots of L2 cache, and o it should support at least 1 GB of fast RAM. 2 GB is better, but I think that more would be overkill. Minimum RAM should be 512 MB, just in case the boss won't (or can't) spring for more right off the bat. I'm also planning to ask for SCSI, in as fast a flavor as I can get approval for, and some _big_ SCSI drives. I'll need card slots for a display adapter, a NIC, and not much more. I'd like those even if there are integrated adapters on the mobo. Of course, if you can point to a ready-to-go server that meets these specs and that we can afford, that's great. We're prepared to spend some money on this, but we're a state government agency, and so have to get bang for the buck. I'm interested in what you folks have to say. -- Mike Andrews mikea@mikea.ath.cx Tired old sysadmin From owner-freebsd-smp@FreeBSD.ORG Thu Dec 4 05:54:09 2003 Return-Path: Delivered-To: freebsd-smp@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id C19C316A4CE for ; Thu, 4 Dec 2003 05:54:09 -0800 (PST) Received: from mgw1.MEIway.com (mgw1.meiway.com [212.73.210.75]) by mx1.FreeBSD.org (Postfix) with ESMTP id 1987243F85 for ; Thu, 4 Dec 2003 05:54:08 -0800 (PST) (envelope-from LConrad@Go2France.com) Received: from VirusGate.MEIway.com (virus-gate.meiway.com [212.73.210.91]) by mgw1.MEIway.com (Postfix Relay Hub) with ESMTP id 4C4B4EF435 for ; Thu, 4 Dec 2003 14:54:06 +0100 (CET) (envelope-from LConrad@Go2France.com) Received: from localhost (localhost.meiway.com [127.0.0.1]) by VirusGate.MEIway.com (Postfix) with SMTP id 7E9FF5D00C for ; Thu, 4 Dec 2003 15:00:05 +0100 (CET) Received: from mail.Go2France.com (ms1.meiway.com [212.73.210.73]) by VirusGate.MEIway.com (Postfix) with ESMTP id 1F4865D00A for ; Thu, 4 Dec 2003 15:00:05 +0100 (CET) Received: from tx0.Go2France.com [24.242.169.51] by mail.Go2France.com with ESMTP (SMTPD32-6.06) id A0221EBE00D2; Thu, 04 Dec 2003 15:09:38 +0100 Message-Id: <6.0.1.1.2.20031204073753.0494c1a0@mail.go2france.com> X-Sender: LConrad@Go2France.com@mail.go2france.com X-Mailer: QUALCOMM Windows Eudora Version 6.0.1.1 Date: Thu, 04 Dec 2003 07:54:00 -0600 To: freebsd-smp@freebsd.org From: Len Conrad In-Reply-To: <20031203164252.B30624@mikea.ath.cx> References: <20031203164252.B30624@mikea.ath.cx> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Subject: Re: Request pointers to current-generation SMP boards X-BeenThere: freebsd-smp@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: FreeBSD SMP implementation group List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 04 Dec 2003 13:54:09 -0000 >I hope to talk the boss into something that is _really_ muscular, in >hope that the rising tide of spam, worms, etc., won't swamp the poor >thing after 6 months to a year. Have the MX machine reject based on the envelope info, which will stop 90+% of the spam. You don't need a very powerful box as MX to do envelope rejection Then pass the remainder to a box much less powerful than you imagine for content-scanning. What volume of msgs do you have today? 100K? 500K msgs? here's the stats for yesterday: Grand Totals ------------ messages 186580 received 26185 delivered 0 forwarded 109 deferred (199 deferrals) 252 bounced 199942 rejected (88%) .... for an MX machine that is 330 MHz, 256 MB RAM, one ATA33 disk. %uptime 7:59AM up 35 days, 9:56, 1 user, load averages: 0.84, 0.36, 0.18 The .84 is due to a lot of hourly maillog scanning done by reporting programs in the last 10 minutes of the hour (ie, not mail relaying activity). The .18 is a more accurate indication of the load. So you can see that scanning 26k mgs received vs the 200K rejected is huge difference in volume, with a huge reduction in the requirements for expensive content-scanning. Note the 26K is really about 19K inbound and 7k outbound, meaning the content-scanning would only look at the 19K. Len _____________________________________________________________________ http://MenAndMice.com/DNS-training: Orlando; San Jose IMGate.MEIway.com: anti-spam gateway, effective on 1000's of sites, free From owner-freebsd-smp@FreeBSD.ORG Sat Dec 6 19:08:12 2003 Return-Path: Delivered-To: freebsd-smp@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 32DDF16A4CE for ; Sat, 6 Dec 2003 19:08:12 -0800 (PST) Received: from mail.drunkencomputing.de (ratz.drunkencomputing.de [195.244.235.248]) by mx1.FreeBSD.org (Postfix) with ESMTP id A5ED743F75 for ; Sat, 6 Dec 2003 19:08:10 -0800 (PST) (envelope-from hscholz@pandemonium.lan.raisdorf.net) Received: from localhost (localhost [127.0.0.1]) by mail.drunkencomputing.de (Postfix) with ESMTP id 1583C8AF94 for ; Sun, 7 Dec 2003 04:08:03 +0100 (CET) Received: from mail.drunkencomputing.de ([127.0.0.1])port 10024) with ESMTP id 31046-06 for ; Sun, 7 Dec 2003 04:08:02 +0100 (CET) Received: by mail.drunkencomputing.de (Postfix, from userid 66) id EB2348AF9F; Sun, 7 Dec 2003 04:08:01 +0100 (CET) Received: by pandemonium.lan.raisdorf.net (Postfix, from userid 1001) id 8E601BBD45; Sun, 7 Dec 2003 03:48:51 +0100 (CET) Date: Sun, 7 Dec 2003 03:48:51 +0100 From: Hendrik Scholz To: freebsd-smp@freebsd.org Message-ID: <20031207024851.GA10570@pandemonium.lan.raisdorf.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.5.4i X-Virus-Scanned: by amavisd-new at drunkencomputing.de Subject: hyperthreading aware scheduler? X-BeenThere: freebsd-smp@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: FreeBSD SMP implementation group List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 07 Dec 2003 03:08:12 -0000 Hi! Does the scheduler actually know the differences between physical and logical CPUs? Is there are workaround the situation where two processes on a Dual CPU box (total of 4 virtual CPUS) are runable and get processed on one physical & one logical (and slower) CPU instead of utilizing both physical CPUs? Thanks, Hendrik -- Hendrik Scholz - - http://raisdorf.net/ cell phone: 404-606-5324 (US) 0160-1570-272 (DE)