From owner-freebsd-performance@FreeBSD.ORG Tue Apr 15 03:40:38 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 8406337B401 for ; Tue, 15 Apr 2003 03:40:38 -0700 (PDT) Received: from bart.esiee.fr (bart.esiee.fr [147.215.1.20]) by mx1.FreeBSD.org (Postfix) with ESMTP id 77BE643FDD for ; Tue, 15 Apr 2003 03:40:37 -0700 (PDT) (envelope-from bonnetf@bart.esiee.fr) Received: from bart.esiee.fr (localhost.esiee.fr [127.0.0.1]) by bart.esiee.fr (8.12.9/8.12.7) with ESMTP id h3FAeZiI025262 for ; Tue, 15 Apr 2003 12:40:36 +0200 (METDST) Received: (from bonnetf@localhost) by bart.esiee.fr (8.12.9/8.12.7/Submit) id h3FAeZjC025261 for freebsd-performance@freebsd.org; Tue, 15 Apr 2003 12:40:35 +0200 (METDST) Date: Tue, 15 Apr 2003 12:40:35 +0200 From: Frank Bonnet To: freebsd-performance@freebsd.org Message-ID: <20030415124035.B25027@bart.esiee.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i Subject: tweaking FreeBSD for Squid using X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 15 Apr 2003 10:40:38 -0000 Hi I need infos in order to tweak a FreeBSD box kernel to use the Squid cache/proxy software on a dedicated machine. The hardware is OK P4 2.2 Ghz with 2.5 Gb RAM and a 36 Gb / 15k rpm disk for caching. thank for any infos -- Frank Bonnet From owner-freebsd-performance@FreeBSD.ORG Tue Apr 15 05:38:57 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id F327137B401 for ; Tue, 15 Apr 2003 05:38:56 -0700 (PDT) Received: from relay2.mecon.ar (relay2.mecon.ar [168.101.16.11]) by mx1.FreeBSD.org (Postfix) with ESMTP id 859AB43FCB for ; Tue, 15 Apr 2003 05:38:55 -0700 (PDT) (envelope-from fernando@mecon.gov.ar) Received: from racing.mecon.ar (racing.mecon.ar [168.101.133.15]) by relay2.mecon.ar (8.12.6/8.12.6) with ESMTP id h3FCcrMp059773 for ; Tue, 15 Apr 2003 09:38:53 -0300 (ART) (envelope-from fernando@mecon.gov.ar) Received: from racing.mecon.ar (meyosp.mecon.gov.ar [10.11.0.149]) by racing.mecon.ar (8.12.6/8.12.6) with ESMTP id h3FCcmba076746; Tue, 15 Apr 2003 09:38:48 -0300 (ART) (envelope-from fernando@mecon.gov.ar) Received: from bal740r0.mecon.gov.ar (bal740r0.mecon.ar [10.11.1.11]) by racing.mecon.ar (8.12.6/8.12.6) with ESMTP id h3FCcmk1076744; Tue, 15 Apr 2003 09:38:48 -0300 (ART) (envelope-from fernando@mecon.gov.ar) Received: from bal740r0.mecon.gov.ar (localhost [127.0.0.1]) by bal740r0.mecon.gov.ar (8.12.6/8.12.6) with ESMTP id h3FCcmVk000501; Tue, 15 Apr 2003 09:38:48 -0300 (ART) (envelope-from fernando@mecon.gov.ar) Received: (from fpscha@localhost) by bal740r0.mecon.gov.ar (8.12.6/8.12.6/Submit) id h3FCclMC000500; Tue, 15 Apr 2003 09:38:47 -0300 (ART) (envelope-from fernando@mecon.gov.ar) X-Authentication-Warning: bal740r0.mecon.gov.ar: fpscha set sender to fernando@mecon.gov.ar using -f Date: Tue, 15 Apr 2003 09:38:47 -0300 From: Fernando Schapachnik To: freebsd-performance@freebsd.org Message-ID: <20030415123847.GC318@bal740r0.mecon.gov.ar> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.4.1i X-OS: FreeBSD 4.7 - http://www.freebsd.org Subject: No buffer space available X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 15 Apr 2003 12:38:57 -0000 Hi, First of all, I hope this is on-topic. Please let me know otherwise. Scenario: 4.7p10 machine, 128 Mb RAM, running 3 squid processes (each one with a different ACL set on a different port). Everything fine. Now I'm trying to end up with one process with a merge of the ACLs, receiving the load of the previous 3. ACLs are in files and account for 168 Kb. After few seconds of running the unified process, squid reports comm_open: socket failure: (55) No buffer space available which maps to: if ((new_socket = socket(AF_INET, sock_type, proto)) < 0) The system doesn't seem to be lucking mbufs (according to netstat -b), nor memory, nor file descriptors. I recompiled kernel with MAXSSIZ="(256*1024*1024)" to get the stack size lifted, but no change. Top shows: ---------------------------------------------------------------------------- last pid: 453; load averages: 0.48, 0.30, 0.13 up 0+00:02:32 08:48:00 36 processes: 1 running, 35 sleeping CPU states: 0.4% user, 0.0% nice, 0.4% system, 0.0% interrupt, 99.2% idle Mem: 52M Active, 4608K Inact, 14M Wired, 3912K Cache, 22M Buf, 45M Free Swap: 500M Total, 13M Used, 487M Free, 2% Inuse PID USERNAME PRI NICE SIZE RES STATE TIME WCPU CPU COMMAND 187 nobody 2 0 46440K 45624K poll 0:07 0.59% 0.59% squid 453 root 28 0 1900K 956K RUN 0:00 0.56% 0.15% top ---------------------------------------------------------------------------- # more /proc/187/rlimit cpu -1 -1 fsize -1 -1 data 536870912 536870912 stack 268435456 268435456 core -1 -1 rss -1 -1 memlock -1 -1 nproc 896 896 nofile 1792 1792 sbsize -1 -1 vmem -1 -1 Each squid individual process has more or less the same memory footprint of the merged one. Squid mailing lists show post with similar problems but no solutions. Any ideas? Thanks in advance. Fernando Schapachnik From owner-freebsd-performance@FreeBSD.ORG Tue Apr 15 06:03:25 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id C991537B401 for ; Tue, 15 Apr 2003 06:03:25 -0700 (PDT) Received: from otter3.centtech.com (moat3.centtech.com [207.200.51.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id BE75143FA3 for ; Tue, 15 Apr 2003 06:03:22 -0700 (PDT) (envelope-from anderson@centtech.com) Received: from centtech.com ([204.177.173.226]) by otter3.centtech.com (8.12.3/8.12.3) with ESMTP id h3FD3L56017330; Tue, 15 Apr 2003 08:03:22 -0500 (CDT) (envelope-from anderson@centtech.com) Message-ID: <3E9C035B.3000301@centtech.com> Date: Tue, 15 Apr 2003 08:04:27 -0500 From: Eric Anderson User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.0.1) Gecko/20020823 Netscape/7.0 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Fernando Schapachnik References: <20030415123847.GC318@bal740r0.mecon.gov.ar> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit cc: freebsd-performance@freebsd.org Subject: Re: No buffer space available X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 15 Apr 2003 13:03:26 -0000 Can you do a netstat -s and send that? Also, have you done any sysctl tweaks? Like maxsoconn? Eric Fernando Schapachnik wrote: >Hi, > First of all, I hope this is on-topic. Please let me know otherwise. > > Scenario: 4.7p10 machine, 128 Mb RAM, running 3 squid processes (each >one with a different ACL set on a different port). Everything fine. Now I'm >trying to end up with one process with a merge of the ACLs, receiving the load >of the previous 3. ACLs are in files and account for 168 Kb. > > After few seconds of running the unified process, squid reports > >comm_open: socket failure: (55) No buffer space available > > which maps to: > >if ((new_socket = socket(AF_INET, sock_type, proto)) < 0) > > The system doesn't seem to be lucking mbufs (according to netstat -b), >nor memory, nor file descriptors. I recompiled kernel with >MAXSSIZ="(256*1024*1024)" to get the stack size lifted, but no change. > > Top shows: > >---------------------------------------------------------------------------- >last pid: 453; load averages: 0.48, 0.30, 0.13 up >0+00:02:32 08:48:00 >36 processes: 1 running, 35 sleeping >CPU states: 0.4% user, 0.0% nice, 0.4% system, 0.0% interrupt, 99.2% idle >Mem: 52M Active, 4608K Inact, 14M Wired, 3912K Cache, 22M Buf, 45M Free >Swap: 500M Total, 13M Used, 487M Free, 2% Inuse > > PID USERNAME PRI NICE SIZE RES STATE TIME WCPU CPU COMMAND > 187 nobody 2 0 46440K 45624K poll 0:07 0.59% 0.59% squid > 453 root 28 0 1900K 956K RUN 0:00 0.56% 0.15% top >---------------------------------------------------------------------------- > ># more /proc/187/rlimit >cpu -1 -1 >fsize -1 -1 >data 536870912 536870912 >stack 268435456 268435456 >core -1 -1 >rss -1 -1 >memlock -1 -1 >nproc 896 896 >nofile 1792 1792 >sbsize -1 -1 >vmem -1 -1 > > Each squid individual process has more or less the same memory >footprint of the merged one. > > Squid mailing lists show post with similar problems but no solutions. > > Any ideas? Thanks in advance. > > > >Fernando Schapachnik >_______________________________________________ >freebsd-performance@freebsd.org mailing list >http://lists.freebsd.org/mailman/listinfo/freebsd-performance >To unsubscribe, send any mail to "freebsd-performance-unsubscribe@freebsd.org" > > From owner-freebsd-performance@FreeBSD.ORG Tue Apr 15 06:57:57 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 2631237B401 for ; Tue, 15 Apr 2003 06:57:57 -0700 (PDT) Received: from relay2.mecon.ar (relay2.mecon.ar [168.101.16.11]) by mx1.FreeBSD.org (Postfix) with ESMTP id 6111E43F93 for ; Tue, 15 Apr 2003 06:57:55 -0700 (PDT) (envelope-from fernando@mecon.gov.ar) Received: from racing.mecon.ar (racing.mecon.ar [168.101.133.15]) by relay2.mecon.ar (8.12.6/8.12.6) with ESMTP id h3FDvoMp073181; Tue, 15 Apr 2003 10:57:50 -0300 (ART) (envelope-from fernando@mecon.gov.ar) Received: from racing.mecon.ar (meyosp.mecon.gov.ar [10.11.0.149]) by racing.mecon.ar (8.12.6/8.12.6) with ESMTP id h3FDviba025266; Tue, 15 Apr 2003 10:57:44 -0300 (ART) (envelope-from fernando@mecon.gov.ar) Received: from bal740r0.mecon.gov.ar (bal740r0.mecon.ar [10.11.1.11]) by racing.mecon.ar (8.12.6/8.12.6) with ESMTP id h3FDvhk1025255; Tue, 15 Apr 2003 10:57:43 -0300 (ART) (envelope-from fernando@mecon.gov.ar) Received: from bal740r0.mecon.gov.ar (localhost [127.0.0.1]) by bal740r0.mecon.gov.ar (8.12.6/8.12.6) with ESMTP id h3FDvhVk000665; Tue, 15 Apr 2003 10:57:43 -0300 (ART) (envelope-from fernando@mecon.gov.ar) Received: (from fpscha@localhost) by bal740r0.mecon.gov.ar (8.12.6/8.12.6/Submit) id h3FDvg3m000664; Tue, 15 Apr 2003 10:57:42 -0300 (ART) (envelope-from fernando@mecon.gov.ar) X-Authentication-Warning: bal740r0.mecon.gov.ar: fpscha set sender to fernando@mecon.gov.ar using -f Date: Tue, 15 Apr 2003 10:57:42 -0300 From: Fernando Schapachnik To: Eric Anderson Message-ID: <20030415135742.GE318@bal740r0.mecon.gov.ar> References: <20030415123847.GC318@bal740r0.mecon.gov.ar> <3E9C035B.3000301@centtech.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: <3E9C035B.3000301@centtech.com> User-Agent: Mutt/1.4.1i X-OS: FreeBSD 4.7 - http://www.freebsd.org cc: freebsd-performance@freebsd.org Subject: Re: No buffer space available X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 15 Apr 2003 13:57:57 -0000 En un mensaje anterior, Eric Anderson escribió: > Can you do a netstat -s and send that? Also, have you done any sysctl > tweaks? Like maxsoconn? That was it! I looked at netstat -a but grepped it to get only squid ports. Looking at the full netstat -a revelead that for some yet unknown reason my external autentication process is leaving thousands of TIME_WAIT sockets against the authentication DB. Why this happens is something I have to figure out, but certainly doesn't have to do with -performance anymore :) Thanks for the help and sorry about the finally uninteresting issue! Regards. Fernando. From owner-freebsd-performance@FreeBSD.ORG Tue Apr 15 16:27:11 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id BE11037B401 for ; Tue, 15 Apr 2003 16:27:11 -0700 (PDT) Received: from mail.2u2.nu (ns1.2u2.nu [62.59.31.66]) by mx1.FreeBSD.org (Postfix) with ESMTP id B128343F3F for ; Tue, 15 Apr 2003 16:27:10 -0700 (PDT) (envelope-from lucio.jankok@2u2.nu) Received: from mail.2u2.nu (localhost [127.0.0.1]) by localhost (Postfix) with ESMTP id 1DD86643 for ; Tue, 15 Apr 2003 21:54:32 +0200 (CEST) Received: from localhost (atlantis.local.net [10.23.31.10]) by mail.2u2.nu (Postfix) with ESMTP id D106462A for ; Tue, 15 Apr 2003 21:54:31 +0200 (CEST) Received: (qmail 29898 invoked by uid 109); 15 Apr 2003 19:57:37 -0000 Date: Tue, 15 Apr 2003 21:57:37 +0200 From: "L. Jankok" To: Frank Bonnet Message-ID: <20030415195737.GA29876@atlantis.local.net> References: <20030415124035.B25027@bart.esiee.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20030415124035.B25027@bart.esiee.fr> User-Agent: Mutt/1.4i X-Operating-System: SunOS atlantis 5.9 sparc X-Sender: lj@2u2.nu X-message-flag: Outlook : A program to spread viri, but it can do mail too. cc: freebsd-performance@freebsd.org Subject: Re: tweaking FreeBSD for Squid using X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: lj@2u2.nu List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 15 Apr 2003 23:27:12 -0000 Hi on Squid's website you will find some pointers. which version of FreeBSD do you want to tweak ?.. the values are version dependant.. regards, Lucio Jankok On Tue, Apr 15, 2003 at 12:40:35PM +0200, Frank Bonnet wrote: :Hi : :I need infos in order to tweak a FreeBSD box kernel :to use the Squid cache/proxy software on a dedicated :machine. : :The hardware is OK P4 2.2 Ghz with 2.5 Gb RAM :and a 36 Gb / 15k rpm disk for caching. : :thank for any infos : :-- :Frank Bonnet :_______________________________________________ :freebsd-performance@freebsd.org mailing list :http://lists.freebsd.org/mailman/listinfo/freebsd-performance :To unsubscribe, send any mail to "freebsd-performance-unsubscribe@freebsd.org" -- The whole problem with the world is that fools and fanatics are always so certain of themselves, and wiser people so full of doubts. --Bertrand Russell From owner-freebsd-performance@FreeBSD.ORG Tue Apr 15 19:10:14 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 1414F37B401 for ; Tue, 15 Apr 2003 19:10:14 -0700 (PDT) Received: from mailserv.ait.ac.th (mailserv.ait.ac.th [203.159.5.10]) by mx1.FreeBSD.org (Postfix) with ESMTP id 02AD543F3F for ; Tue, 15 Apr 2003 19:10:13 -0700 (PDT) (envelope-from alain@ait.ac.th) Received: from localhost (mailserv.ait.ac.th [127.0.0.1]) by mailserv.ait.ac.th (Postfix) with ESMTP id 13387113F66; Wed, 16 Apr 2003 09:10:11 +0700 (ICT) Received: from alain.ait.ac.th (alain.itserv.ait.ac.th [203.159.28.4]) by mailserv.ait.ac.th (Postfix) with ESMTP id 41EAE113F97; Wed, 16 Apr 2003 09:10:07 +0700 (ICT) Received: (from alain@localhost) by alain.ait.ac.th (8.12.4/8.12.4) id h3G2A5LF007881; Wed, 16 Apr 2003 09:10:05 +0700 Date: Wed, 16 Apr 2003 09:10:04 +0700 From: Alain Fauconnet To: Frank Bonnet Message-ID: <20030416021004.GA7867@ait.ac.th> References: <20030415124035.B25027@bart.esiee.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20030415124035.B25027@bart.esiee.fr> User-Agent: Mutt/1.4i X-Virus-Scanned: by AMaViS perl-11 cc: freebsd-performance@freebsd.org Subject: Re: tweaking FreeBSD for Squid using X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Apr 2003 02:10:14 -0000 On Tue, Apr 15, 2003 at 12:40:35PM +0200, Frank Bonnet wrote: > Hi > > I need infos in order to tweak a FreeBSD box kernel > to use the Squid cache/proxy software on a dedicated > machine. > > The hardware is OK P4 2.2 Ghz with 2.5 Gb RAM > and a 36 Gb / 15k rpm disk for caching. To my experience, heavly loaded Squid boxes run into I/O bottlenecks faster than anything else (apart from memory maybe, but you have ample memory). My advice would be: get more disks. A single spindle is just not a good configuration for Squid, even a fast disk. You didn't mention whether is it SCSI or IDE. I usually go for SCSI for anything I/O intensive, although that's more a cultural thing than based on any hard evidence I've seen. Don't do RAID. Bring up a filesystem on each disk (with soft updates of course), mount them "-o noatime" and configure Squid to use multiple cache dirs. Sorry if I'm just telling the obvious here. I'm sure someone will come up with some finer tuning advice. Bonne chance, -- Alain Fauconnet IT Security Specialist & CISO -- ITServ Asian Institute of Technology From owner-freebsd-performance@FreeBSD.ORG Tue Apr 15 19:31:26 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 36CBC37B401 for ; Tue, 15 Apr 2003 19:31:26 -0700 (PDT) Received: from walter.dfmm.org (walter.dfmm.org [209.151.233.240]) by mx1.FreeBSD.org (Postfix) with ESMTP id B1E7243F93 for ; Tue, 15 Apr 2003 19:31:25 -0700 (PDT) (envelope-from jason@shalott.net) Received: (qmail 60278 invoked by uid 1000); 16 Apr 2003 02:31:25 -0000 Received: from localhost (sendmail-bs@127.0.0.1) by localhost with SMTP; 16 Apr 2003 02:31:25 -0000 Date: Tue, 15 Apr 2003 19:31:24 -0700 (PDT) From: Jason Stone X-X-Sender: To: In-Reply-To: <20030416021004.GA7867@ait.ac.th> Message-ID: <20030415192851.G4074-100000@walter> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Subject: Re: tweaking FreeBSD for Squid using X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Apr 2003 02:31:26 -0000 -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 > You didn't mention whether is it SCSI or IDE. Based on the fact that it's 15k rpms, I'm guessing scsi - do ide disks faster than 7.2k rpms exist? > Don't do RAID. Bring up a filesystem on each disk (with soft updates > of course), mount them "-o noatime" and configure Squid to use > multiple cache dirs. Why is this preferrable to striping with raid-0? -Jason -------------------------------------------------------------------------- Freud himself was a bit of a cold fish, and one cannot avoid the suspicion that he was insufficiently fondled when he was an infant. -- Ashley Montagu -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (FreeBSD) Comment: See https://private.idealab.com/public/jason/jason.gpg iD8DBQE+nMB9swXMWWtptckRAvo2AJ0exlPHsPvOhn/BFzQhR8wYYBFoUQCfYg6Y 6x7sGEeGNs0YgDbCS8itklM= =AiEv -----END PGP SIGNATURE----- From owner-freebsd-performance@FreeBSD.ORG Tue Apr 15 19:48:48 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 62CEB37B401 for ; Tue, 15 Apr 2003 19:48:48 -0700 (PDT) Received: from mailserv.ait.ac.th (mailserv.ait.ac.th [203.159.5.10]) by mx1.FreeBSD.org (Postfix) with ESMTP id 993E643F85 for ; Tue, 15 Apr 2003 19:48:47 -0700 (PDT) (envelope-from alain@ait.ac.th) Received: from localhost (mailserv.ait.ac.th [127.0.0.1]) by mailserv.ait.ac.th (Postfix) with ESMTP id 64D66113FBB; Wed, 16 Apr 2003 09:48:46 +0700 (ICT) Received: from alain.ait.ac.th (alain.itserv.ait.ac.th [203.159.28.4]) by mailserv.ait.ac.th (Postfix) with ESMTP id 846F0113F36; Wed, 16 Apr 2003 09:48:45 +0700 (ICT) Received: (from alain@localhost) by alain.ait.ac.th (8.12.4/8.12.4) id h3G2miK7028348; Wed, 16 Apr 2003 09:48:44 +0700 Date: Wed, 16 Apr 2003 09:48:44 +0700 From: Alain Fauconnet To: Jason Stone Message-ID: <20030416024844.GC7867@ait.ac.th> References: <20030416021004.GA7867@ait.ac.th> <20030415192851.G4074-100000@walter> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20030415192851.G4074-100000@walter> User-Agent: Mutt/1.4i X-Virus-Scanned: by AMaViS perl-11 cc: freebsd-performance@freebsd.org Subject: Re: tweaking FreeBSD for Squid using X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Apr 2003 02:48:48 -0000 On Tue, Apr 15, 2003 at 07:31:24PM -0700, Jason Stone wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > > > You didn't mention whether is it SCSI or IDE. > > Based on the fact that it's 15k rpms, I'm guessing scsi - do ide disks > faster than 7.2k rpms exist? I don't know. Seems that IDE disks evolve too fast for me nowadays. That's also why I was writing that I'm not even sure that the old stance "don't use IDE for servers" is still valid. OOTH, I've had a lot of trouble with busy IDE-based (ASUS P4* m/b) FreeBSD servers lately (hard hangs, see bug kern/44867). > > > > Don't do RAID. Bring up a filesystem on each disk (with soft updates > > of course), mount them "-o noatime" and configure Squid to use > > multiple cache dirs. > > Why is this preferrable to striping with raid-0? Well, because Squid does load balancing over multiple cache dirs quite well by itself, and (presumably) in a smarter way than just spreading raw disk blocks, so adding another layer of software for RAID-0 doesn't bring anything, and wastes CPU cycles. I'm not even sure that hardware RAID-0 is a good idea. According to my own experience (admittedly on a Linux box), removing software striping and using multiple cache dirs on physical volumes gave me a significant performance boost and lowered the load average of the server by about 30%. Since then, I've always stayed away from RAID on Squid boxes, whatever the O/S. Er. This is getting off topic maybe. I'll follow up privately if needed. _Alain_ From owner-freebsd-performance@FreeBSD.ORG Tue Apr 15 21:52:24 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id BBFCF37B401 for ; Tue, 15 Apr 2003 21:52:24 -0700 (PDT) Received: from rms21.rommon.net (rms21.rommon.net [193.64.42.200]) by mx1.FreeBSD.org (Postfix) with ESMTP id 498D743FBF for ; Tue, 15 Apr 2003 21:52:23 -0700 (PDT) (envelope-from pete@he.iki.fi) Received: from PHE (h93.vuokselantie10.fi [193.64.42.147]) by rms21.rommon.net (8.12.6p2/8.12.6) with SMTP id h3G4q2qo084486; Wed, 16 Apr 2003 07:52:11 +0300 (EEST) (envelope-from pete@he.iki.fi) Message-ID: <06a501c303d4$05c78de0$932a40c1@PHE> From: "Petri Helenius" To: "Alain Fauconnet" , "Jason Stone" References: <20030416021004.GA7867@ait.ac.th><20030415192851.G4074-100000@walter> <20030416024844.GC7867@ait.ac.th> Date: Wed, 16 Apr 2003 07:52:34 +0300 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 8bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1106 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 cc: freebsd-performance@freebsd.org Subject: Re: tweaking FreeBSD for Squid using X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Apr 2003 04:52:25 -0000 > > I don't know. Seems that IDE disks evolve too fast for me nowadays. > That's also why I was writing that I'm not even sure that the old > stance "don't use IDE for servers" is still valid. > OOTH, I've had a lot of trouble with busy IDE-based (ASUS P4* m/b) > FreeBSD servers lately (hard hangs, see bug kern/44867). > Western digital Raptor´s spin at 10000rpm though they only come with 37 gig at the moment. So good for database applications but not for large scale storage. Pete From owner-freebsd-performance@FreeBSD.ORG Tue Apr 15 22:14:18 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 314E637B404 for ; Tue, 15 Apr 2003 22:14:18 -0700 (PDT) Received: from mail.2u2.nu (ns1.2u2.nu [62.59.31.66]) by mx1.FreeBSD.org (Postfix) with ESMTP id A760843FA3 for ; Tue, 15 Apr 2003 22:14:16 -0700 (PDT) (envelope-from lucio.jankok@2u2.nu) Received: from mail.2u2.nu (localhost [127.0.0.1]) by localhost (Postfix) with ESMTP id 95ED3646 for ; Wed, 16 Apr 2003 07:11:07 +0200 (CEST) Received: from localhost (atlantis.local.net [10.23.31.10]) by mail.2u2.nu (Postfix) with ESMTP id 2D209638 for ; Wed, 16 Apr 2003 07:11:07 +0200 (CEST) Received: (qmail 692 invoked by uid 109); 16 Apr 2003 05:14:13 -0000 Date: Wed, 16 Apr 2003 07:14:13 +0200 From: "L. Jankok" To: Petri Helenius Message-ID: <20030416051413.GB653@atlantis.local.net> References: <20030416024844.GC7867@ait.ac.th> <06a501c303d4$05c78de0$932a40c1@PHE> Mime-Version: 1.0 Content-Type: text/plain; charset=unknown-8bit Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: <06a501c303d4$05c78de0$932a40c1@PHE> User-Agent: Mutt/1.4i X-Operating-System: SunOS atlantis 5.9 sparc X-Sender: lj@2u2.nu X-message-flag: Outlook : A program to spread viri, but it can do mail too. cc: Jason Stone cc: freebsd-performance@freebsd.org cc: Alain Fauconnet Subject: Re: tweaking FreeBSD for Squid using X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: lj@2u2.nu List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Apr 2003 05:14:18 -0000 The last time I checked IDE drives were still sequential in their access.. so by design a bottleneck for high performance databases or squid.. and a burden for your cpu Lucio Jankok On Wed, Apr 16, 2003 at 07:52:34AM +0300, Petri Helenius wrote: :> :> I don't know. Seems that IDE disks evolve too fast for me nowadays. :> That's also why I was writing that I'm not even sure that the old :> stance "don't use IDE for servers" is still valid. :> OOTH, I've had a lot of trouble with busy IDE-based (ASUS P4* m/b) :> FreeBSD servers lately (hard hangs, see bug kern/44867). :> :Western digital Raptor´s spin at 10000rpm though they only come with 37 gig :at the moment. So good for database applications but not for large scale storage. : :Pete : :_______________________________________________ :freebsd-performance@freebsd.org mailing list :http://lists.freebsd.org/mailman/listinfo/freebsd-performance :To unsubscribe, send any mail to "freebsd-performance-unsubscribe@freebsd.org" -- The whole problem with the world is that fools and fanatics are always so certain of themselves, and wiser people so full of doubts. --Bertrand Russell From owner-freebsd-performance@FreeBSD.ORG Tue Apr 15 23:19:00 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 8DFB337B401 for ; Tue, 15 Apr 2003 23:19:00 -0700 (PDT) Received: from rms21.rommon.net (rms21.rommon.net [193.64.42.200]) by mx1.FreeBSD.org (Postfix) with ESMTP id 4FBD843FD7 for ; Tue, 15 Apr 2003 23:18:59 -0700 (PDT) (envelope-from pete@he.iki.fi) Received: from PHE (h93.vuokselantie10.fi [193.64.42.147]) by rms21.rommon.net (8.12.6p2/8.12.6) with SMTP id h3G6Idqo085701; Wed, 16 Apr 2003 09:18:39 +0300 (EEST) (envelope-from pete@he.iki.fi) Message-ID: <072601c303e0$19a50a70$932a40c1@PHE> From: "Petri Helenius" To: References: <20030416024844.GC7867@ait.ac.th> <06a501c303d4$05c78de0$932a40c1@PHE> <20030416051413.GB653@atlantis.local.net> Date: Wed, 16 Apr 2003 09:19:12 +0300 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 8bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1106 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 cc: Jason Stone cc: freebsd-performance@freebsd.org cc: Alain Fauconnet Subject: Re: tweaking FreeBSD for Squid using X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Apr 2003 06:19:00 -0000 You mean ATA OVERLAP support? It seems to be there only with the former IBM drives. Does not seem to change that much. Your interrupt latency is usually three to four orders of magnitude smaller than disk latency anyway. I used to buy IBM drives but gave up because of reliability issues and changed over to WD. Pete > The last time I checked IDE drives were still sequential in their > access.. so by design a bottleneck for high performance databases > or squid.. and a burden for your cpu > > Lucio Jankok > > On Wed, Apr 16, 2003 at 07:52:34AM +0300, Petri Helenius wrote: > :> > :> I don't know. Seems that IDE disks evolve too fast for me nowadays. > :> That's also why I was writing that I'm not even sure that the old > :> stance "don't use IDE for servers" is still valid. > :> OOTH, I've had a lot of trouble with busy IDE-based (ASUS P4* m/b) > :> FreeBSD servers lately (hard hangs, see bug kern/44867). > :> > :Western digital Raptor´s spin at 10000rpm though they only come with 37 gig > :at the moment. So good for database applications but not for large scale storage. > : > :Pete > : > :_______________________________________________ > :freebsd-performance@freebsd.org mailing list > :http://lists.freebsd.org/mailman/listinfo/freebsd-performance > :To unsubscribe, send any mail to "freebsd-performance-unsubscribe@freebsd.org" > > -- > The whole problem with the world is that fools and fanatics are always > so certain of themselves, and wiser people so full of doubts. > --Bertrand Russell > > From owner-freebsd-performance@FreeBSD.ORG Tue Apr 15 23:42:57 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id E403D37B401 for ; Tue, 15 Apr 2003 23:42:57 -0700 (PDT) Received: from stork.mail.pas.earthlink.net (stork.mail.pas.earthlink.net [207.217.120.188]) by mx1.FreeBSD.org (Postfix) with ESMTP id F233E43FBF for ; Tue, 15 Apr 2003 23:42:56 -0700 (PDT) (envelope-from tlambert2@mindspring.com) Received: from pool0135.cvx40-bradley.dialup.earthlink.net ([216.244.42.135] helo=mindspring.com) by stork.mail.pas.earthlink.net with asmtp (SSLv3:RC4-MD5:128) (Exim 3.33 #1) id 195gd8-0000sd-00; Tue, 15 Apr 2003 23:42:51 -0700 Message-ID: <3E9CFB18.8ECD3502@mindspring.com> Date: Tue, 15 Apr 2003 23:41:28 -0700 From: Terry Lambert X-Mailer: Mozilla 4.79 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 To: Alain Fauconnet References: <20030416021004.GA7867@ait.ac.th> <20030416024844.GC7867@ait.ac.th> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-ELNK-Trace: b1a02af9316fbb217a47c185c03b154d40683398e744b8a4f9a4ca659417c43c30646f17c74876d7350badd9bab72f9c350badd9bab72f9c350badd9bab72f9c cc: freebsd-performance@freebsd.org cc: Jason Stone Subject: Re: tweaking FreeBSD for Squid using X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Apr 2003 06:42:58 -0000 Alain Fauconnet wrote: > I don't know. Seems that IDE disks evolve too fast for me nowadays. > That's also why I was writing that I'm not even sure that the old > stance "don't use IDE for servers" is still valid. > OOTH, I've had a lot of trouble with busy IDE-based (ASUS P4* m/b) > FreeBSD servers lately (hard hangs, see bug kern/44867). FWIW: IDE disks do not support disconnected writes, which is a significant performance bottleneck when you are writing to the disk; only disconnected reads are supported. What this means is that, depending on your content, you could take a big performance hit for anything you write to cache. It's a pretty common practice these days for active content to permit itself to be cached, but have a short expiration; even if that weren't the case, there's plenty of cacheable content out there (think "pretty much every image"). What's worse is that this is precisely the larger content, which is exactly what you would most want to write to disk. I guess it really depends on why you are proxy-caching the stuff in the first place. If it's because you have a lot of users behind a big pipe, then you are probably going to end up being I/O bound on the disk. I definitely agree with you that you are better off with multiple cache dirs than with RAID-0... on the other hand, it could be worth it to have multiple disks, if you are using IDE, since you get one concurrent outstanding write per IDE disk you add. To get 8 concurrent writes, you pretty much need 8 disks, and the only way to get something like that would be an IDE RAID controller. Personally, I'd just spring the extra hundred bucks for the SCSI, though, and get 128 or more concurrent writes, instead (tagged command queue depth). -- Terry From owner-freebsd-performance@FreeBSD.ORG Wed Apr 16 00:27:06 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id D067F37B401 for ; Wed, 16 Apr 2003 00:27:06 -0700 (PDT) Received: from bart.esiee.fr (bart.esiee.fr [147.215.1.20]) by mx1.FreeBSD.org (Postfix) with ESMTP id BD97643FB1 for ; Wed, 16 Apr 2003 00:27:05 -0700 (PDT) (envelope-from bonnetf@bart.esiee.fr) Received: from bart.esiee.fr (localhost.esiee.fr [127.0.0.1]) by bart.esiee.fr (8.12.9/8.12.7) with ESMTP id h3G7QuiI001450; Wed, 16 Apr 2003 09:26:56 +0200 (METDST) Received: (from bonnetf@localhost) by bart.esiee.fr (8.12.9/8.12.7/Submit) id h3G7Qtk8001449; Wed, 16 Apr 2003 09:26:55 +0200 (METDST) Date: Wed, 16 Apr 2003 09:26:55 +0200 From: Frank Bonnet To: Alain Fauconnet Message-ID: <20030416092655.B1396@bart.esiee.fr> References: <20030416021004.GA7867@ait.ac.th> <20030415192851.G4074-100000@walter> <20030416024844.GC7867@ait.ac.th> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <20030416024844.GC7867@ait.ac.th>; from alain@ait.ac.th on Wed, Apr 16, 2003 at 09:48:44AM +0700 cc: freebsd-performance@freebsd.org cc: Jason Stone Subject: Re: tweaking FreeBSD for Squid using X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Apr 2003 07:27:07 -0000 On Wed, Apr 16, 2003 at 09:48:44AM +0700, Alain Fauconnet wrote: > On Tue, Apr 15, 2003 at 07:31:24PM -0700, Jason Stone wrote: > > -----BEGIN PGP SIGNED MESSAGE----- > > Hash: SHA1 > > > > > > > You didn't mention whether is it SCSI or IDE. > > > > Based on the fact that it's 15k rpms, I'm guessing scsi - do ide disks > > faster than 7.2k rpms exist? > disk is SCSI of course -- Frank Bonnet From owner-freebsd-performance@FreeBSD.ORG Wed Apr 16 07:23:26 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id E5E4237B401 for ; Wed, 16 Apr 2003 07:23:25 -0700 (PDT) Received: from svaha.com (svaha.com [64.46.156.67]) by mx1.FreeBSD.org (Postfix) with ESMTP id F042143F75 for ; Wed, 16 Apr 2003 07:23:24 -0700 (PDT) (envelope-from meconlen@obfuscated.net) Received: from FUSIONBOX ([64.156.25.5]) (AUTH: LOGIN meconlen, TLS: TLSv1/SSLv3,128bits,RC4-MD5) by svaha.com with esmtp; Wed, 16 Apr 2003 10:23:23 -0400 From: "Michael Conlen" To: "'Alain Fauconnet'" , "'Jason Stone'" Date: Wed, 16 Apr 2003 10:23:17 -0400 Message-ID: <000a01c30423$bd0f2a40$2b038c0a@corp.neutelligent.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook, Build 10.0.4510 Importance: Normal In-Reply-To: <20030416024844.GC7867@ait.ac.th> X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 cc: freebsd-performance@freebsd.org Subject: RE: tweaking FreeBSD for Squid using X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Apr 2003 14:23:26 -0000 A lot of what you face doing a Squid server is backplane and other bus issues, though it's dependant on what you call "high performance" A pair of Sun E220R's (2 SPARC II processors) for example handled 1 = million requests a day on a pair of mirrored 72 GB drives each. (Granted they = were very nice 72GB drives). The thing about the Sun boxes was that they = could get information out of memory really really fast, and the NIC cards = could work to their full potential. Every device that did IO was on it's own = PCI bus.=20 It used to be that IDE drives took more processing power from the host = to perform it's operations, where as SCSI does not. If that's still true = I'd use that as a reason to stay away from IDE.=20 The other advantage of SCSI, if you need great disk IO, is that you can = have a lot of spindles. On a large SCSI system in a Sun for example I can get = a single drive array to look like one SCSI device (with 14 disks in it) = and put a lot of arrays on a channel. If I buy small, fast SCSI disks I can = take full advantage of the 160 MB/sec array, where as I've seen a big fast = IDE disk push no more than 10 MB/sec. The arrays can do RAID before it gets = to the controller card, so you don't need the RAID in the box at all.=20 Speaking of which, does anyone know of SCSI disk arrays with hardware = RAID that work with FreeBSD?=20 I've moved out of the Sun world and in to the FreeBSD world = professionally and have no idea what's out there for PC hardware. -- Michael Conlen -----Original Message----- From: owner-freebsd-performance@freebsd.org [mailto:owner-freebsd-performance@freebsd.org] On Behalf Of Alain = Fauconnet Sent: Tuesday, April 15, 2003 10:49 PM To: Jason Stone Cc: freebsd-performance@freebsd.org Subject: Re: tweaking FreeBSD for Squid using On Tue, Apr 15, 2003 at 07:31:24PM -0700, Jason Stone wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 >=20 >=20 > > You didn't mention whether is it SCSI or IDE. >=20 > Based on the fact that it's 15k rpms, I'm guessing scsi - do ide disks > faster than 7.2k rpms exist? I don't know. Seems that IDE disks evolve too fast for me nowadays. That's also why I was writing that I'm not even sure that the old stance "don't use IDE for servers" is still valid. OOTH, I've had a lot of trouble with busy IDE-based (ASUS P4* = m/b) FreeBSD servers lately (hard hangs, see bug kern/44867). >=20 >=20 > > Don't do RAID. Bring up a filesystem on each disk (with soft = updates > > of course), mount them "-o noatime" and configure Squid to = use > > multiple cache dirs. >=20 > Why is this preferrable to striping with raid-0? Well, because Squid does load balancing over multiple cache dirs quite well by itself, and (presumably) in a smarter way than just spreading raw disk blocks, so adding another layer of software for RAID-0 doesn't bring anything, and wastes CPU cycles. I'm not even sure that hardware RAID-0 is a good idea. According to my own experience (admittedly on a Linux box), removing software striping and using multiple cache dirs on physical volumes gave me a significant performance boost and lowered the load average of the server by about 30%. Since then, I've always stayed away from RAID on Squid boxes, whatever the O/S. Er. This is getting off topic maybe. I'll follow up privately if needed. _Alain_ _______________________________________________ freebsd-performance@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-performance To unsubscribe, send any mail to "freebsd-performance-unsubscribe@freebsd.org" From owner-freebsd-performance@FreeBSD.ORG Wed Apr 16 07:52:15 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id B131637B401 for ; Wed, 16 Apr 2003 07:52:15 -0700 (PDT) Received: from sabre.velocet.net (sabre.velocet.net [216.138.209.205]) by mx1.FreeBSD.org (Postfix) with ESMTP id CDE2343FCB for ; Wed, 16 Apr 2003 07:52:14 -0700 (PDT) (envelope-from dgilbert@velocet.ca) Received: from trooper.velocet.ca (trooper.velocet.net [216.138.242.2]) by sabre.velocet.net (Postfix) with ESMTP id BB9D0138153; Wed, 16 Apr 2003 10:52:11 -0400 (EDT) Received: by trooper.velocet.ca (Postfix, from userid 66) id CAF2974D72; Wed, 16 Apr 2003 10:52:10 -0400 (EDT) Received: by canoe.velocet.net (Postfix, from userid 101) id 1788756791B; Wed, 16 Apr 2003 10:52:08 -0400 (EDT) From: David Gilbert MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-ID: <16029.28184.11825.471228@canoe.velocet.net> Date: Wed, 16 Apr 2003 10:52:08 -0400 To: "Michael Conlen" In-Reply-To: <000a01c30423$bd0f2a40$2b038c0a@corp.neutelligent.com> References: <20030416024844.GC7867@ait.ac.th> <000a01c30423$bd0f2a40$2b038c0a@corp.neutelligent.com> X-Mailer: VM 7.07 under 21.1 (patch 14) "Cuyahoga Valley" XEmacs Lucid cc: 'Jason Stone' cc: freebsd-performance@freebsd.org cc: 'Alain Fauconnet' Subject: RE: tweaking FreeBSD for Squid using X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Apr 2003 14:52:16 -0000 >>>>> "Michael" == Michael Conlen writes: Michael> A lot of what you face doing a Squid server is backplane and Michael> other bus issues, though it's dependant on what you call Michael> "high performance" Michael> A pair of Sun E220R's (2 SPARC II processors) for example Michael> handled 1 million requests a day on a pair of mirrored 72 GB Michael> drives each. (Granted they were very nice 72GB drives). The Michael> thing about the Sun boxes was that they could get information Michael> out of memory really really fast, and the NIC cards could Michael> work to their full potential. Every device that did IO was on Michael> it's own PCI bus. There are several orders of magnitude in difference between motherboards (even of the same chipset) for PCI performance. PCI seems to be a bus that can be implemented well ... or very, very poorly. If you're planning to serve up 100Mbit plus from a PC, test several good (ie: expensive) motherboards in a bakeoff. Motherboards change so often that I can't even give you recomendations ... you can't buy them anymore. Ironically, many of the best motherboards for performance have also been high DOA. The K7S5A, for instance, had a DOA rate of 50% for us (50% crashed on memory stress tests, etc), but the good ones culled from the litter are among the best boards we have in production. Michael> It used to be that IDE drives took more processing power from Michael> the host to perform it's operations, where as SCSI does Michael> not. If that's still true I'd use that as a reason to stay Michael> away from IDE. The real advantage of SCSI (for large request rates) is tagged command queueing. Many spindles + tagged queueing = fast. Michael> The other advantage of SCSI, if you need great disk IO, is Michael> that you can have a lot of spindles. On a large SCSI system Michael> in a Sun for example I can get a single drive array to look Michael> like one SCSI device (with 14 disks in it) and put a lot of Michael> arrays on a channel. If I buy small, fast SCSI disks I can Michael> take full advantage of the 160 MB/sec array, where as I've Michael> seen a big fast IDE disk push no more than 10 MB/sec. The Michael> arrays can do RAID before it gets to the controller card, so Michael> you don't need the RAID in the box at all. RAID isn't always a win with Squid. Michael> Speaking of which, does anyone know of SCSI disk arrays with Michael> hardware RAID that work with FreeBSD? Michael> I've moved out of the Sun world and in to the FreeBSD world Michael> professionally and have no idea what's out there for PC Michael> hardware. As I've said before, in the category of non-silly-expensive RAID, vinum is faster than any I've tested. that said, SCSI<-->SCSI raid systems should all work with FreeBSD. Look in the hardware release notes for PCI raid devices, but dis-recomend them. Dave. -- ============================================================================ |David Gilbert, Velocet Communications. | Two things can only be | |Mail: dgilbert@velocet.net | equal if and only if they | |http://daveg.ca | are precisely opposite. | =========================================================GLO================ From owner-freebsd-performance@FreeBSD.ORG Wed Apr 16 15:20:57 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 6379A37B401 for ; Wed, 16 Apr 2003 15:20:57 -0700 (PDT) Received: from kurdistan.ath.cx (adsl-64-169-155-173.dsl.chic01.pacbell.net [64.169.155.173]) by mx1.FreeBSD.org (Postfix) with ESMTP id 338C843F3F for ; Wed, 16 Apr 2003 15:20:56 -0700 (PDT) (envelope-from sereciya@kurdistan.ath.cx) Received: from kurdistan.ath.cx (ns1 [127.0.0.1]) by kurdistan.ath.cx (8.12.8/8.12.6) with ESMTP id h3GMKv04058322 for ; Wed, 16 Apr 2003 15:20:57 -0700 (PDT) (envelope-from sereciya@kurdistan.ath.cx) Received: (from sereciya@localhost) by kurdistan.ath.cx (8.12.8/8.12.6/Submit) id h3GMKvkS058321 for freebsd-performance@freebsd.org; Wed, 16 Apr 2003 15:20:57 -0700 (PDT) Date: Wed, 16 Apr 2003 15:20:57 -0700 From: =?unknown-8bit?Q?S=EAr=EAciya_Kurdistan=EE?= To: freebsd-performance@freebsd.org Message-ID: <20030416222057.GC57404@kurdistan.ath.cx> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="CblX+4bnyfN0pR09" Content-Disposition: inline Subject: FreeBSD Memory Pages Not Locked? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Apr 2003 22:20:57 -0000 --CblX+4bnyfN0pR09 Content-Type: text/plain; charset=unknown-8bit Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Hello, I recently installed "gpnupg" from the ports collection and upon running it (after the key generation), I found myself seeing the following error: gpg: Warning: using insecure memory! ... Those of you who are impatient and think that this is a gpg &| port specific problem, please be patient and read on ... from: http://www.gnupg.org/documentation/faqs.html#q6.1 Here's what the GPG FAQ says:=20 "6.1) Why do I get "gpg: Warning: using insecure memory!" On many systems this program should be installed as setuid(root). This is necessary to lock memory pages. Locking memory pages prevents the operating system from writing them to disk and thereby keeping your secret keys really secret. If you get no warning message about insecure memory your operating system supports locking without being root. The program drops root privileges as soon as locked memory is allocated." So my question is: does FreeBSD really not have support for locking memory pages? if this is true, then what is the reason that this has not yet been implemented, is this not an important security feature? otherwise... if FreeBSD does in fact have support for locking memory pages, then why am I getting this error message? If any of you have encountered this problem, and would like to offer some help &| advice, you have a captive audience of at least one, me! Thanks in advance, --$=EAr=EAciya Kurdistan=EE +--------------------------------------------------------------+ | Welat xwe ava nake, dest bidin hevdu, pist nedin tu dijmin=EE | | Riya azadiy=EA ne h=EAsan e, h=EAviya xwe bernedin, dema me | | n=EAz=EEk e. | | | | Hevalt=EE bi kes=EAn du r=FB nekin, hevalt=EE bi hevdu ra bikin | | Ne ji hevaltiya wan kes=EAn p=EAxwas =FB r=FB dir=EAj, ne bi wan | | kes=EAn xw=EEnperest, ne j=EE ji y=EAn din. | | | | -$=EAr=EAciya Kurdistan=EE | +--------------------------------------------------------------+ translation provided on request: sereciya at kurdistan.ath.cx --CblX+4bnyfN0pR09 Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (FreeBSD) iD8DBQE+nddJ2rX+NTHoe2QRAj3UAKCY2ucSxuMUWwF15CxOw4LLCQRAvACeL+qo loWcvlLzmlsTy6h3lzqsdo8= =badu -----END PGP SIGNATURE----- --CblX+4bnyfN0pR09-- From owner-freebsd-performance@FreeBSD.ORG Wed Apr 16 16:17:57 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id DF01F37B401 for ; Wed, 16 Apr 2003 16:17:57 -0700 (PDT) Received: from perrin.int.nxad.com (internal.ext.nxad.com [69.1.70.251]) by mx1.FreeBSD.org (Postfix) with ESMTP id 696FB43FBD for ; Wed, 16 Apr 2003 16:17:57 -0700 (PDT) (envelope-from sean@perrin.int.nxad.com) Received: by perrin.int.nxad.com (Postfix, from userid 1001) id 5D10921061; Wed, 16 Apr 2003 16:17:56 -0700 (PDT) Date: Wed, 16 Apr 2003 16:17:56 -0700 From: Sean Chittenden To: =?unknown-8bit?Q?S=EAr=EAciya_Kurdistan=EE?= Message-ID: <20030416231756.GI79923@perrin.int.nxad.com> References: <20030416222057.GC57404@kurdistan.ath.cx> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="U+NfgObvpQT1Q9Yq" Content-Disposition: inline In-Reply-To: <20030416222057.GC57404@kurdistan.ath.cx> User-Agent: Mutt/1.4i X-PGP-Key: finger seanc@FreeBSD.org X-PGP-Fingerprint: 3849 3760 1AFE 7B17 11A0 83A6 DD99 E31F BC84 B341 X-Web-Homepage: http://sean.chittenden.org/ cc: freebsd-performance@freebsd.org Subject: Re: FreeBSD Memory Pages Not Locked? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Apr 2003 23:17:58 -0000 --U+NfgObvpQT1Q9Yq Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable > I recently installed "gpnupg" from the ports collection and > upon running it (after the key generation), I found myself > seeing the following error: >=20 > gpg: Warning: using insecure memory! >=20 > ... Those of you who are impatient and think that this is a > gpg &| port specific problem, please be patient and read on ... >=20 > from: http://www.gnupg.org/documentation/faqs.html#q6.1 > Here's what the GPG FAQ says:=20 I believe it's because some of the pages required to use this program are on disk in swap someplace. This is very off topic for performance@, please ask on ports@ or questions@ next time. -sc --=20 Sean Chittenden --U+NfgObvpQT1Q9Yq Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Comment: Sean Chittenden iD8DBQE+neSj3ZnjH7yEs0ERAgS/AJ9leljIfguEsvNc25ngM8JkB0L8zgCgnm7M TNHUlXbUN+DFPb82ulleJDA= =UB4g -----END PGP SIGNATURE----- --U+NfgObvpQT1Q9Yq-- From owner-freebsd-performance@FreeBSD.ORG Wed Apr 16 16:23:35 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id A048D37B401 for ; Wed, 16 Apr 2003 16:23:35 -0700 (PDT) Received: from kurdistan.ath.cx (adsl-64-169-155-173.dsl.chic01.pacbell.net [64.169.155.173]) by mx1.FreeBSD.org (Postfix) with ESMTP id B747F43FAF for ; Wed, 16 Apr 2003 16:23:34 -0700 (PDT) (envelope-from sereciya@kurdistan.ath.cx) Received: from kurdistan.ath.cx (ns1 [127.0.0.1]) by kurdistan.ath.cx (8.12.8/8.12.6) with ESMTP id h3GNNa04059933 for ; Wed, 16 Apr 2003 16:23:36 -0700 (PDT) (envelope-from sereciya@kurdistan.ath.cx) Received: (from sereciya@localhost) by kurdistan.ath.cx (8.12.8/8.12.6/Submit) id h3GNNZ6e059932 for freebsd-performance@freebsd.org; Wed, 16 Apr 2003 16:23:35 -0700 (PDT) Date: Wed, 16 Apr 2003 16:23:35 -0700 From: =?unknown-8bit?Q?S=EAr=EAciya_Kurdistan=EE?= To: freebsd-performance@freebsd.org Message-ID: <20030416232335.GE58471@kurdistan.ath.cx> References: <20030416222057.GC57404@kurdistan.ath.cx> <20030416231756.GI79923@perrin.int.nxad.com> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="brEuL7wsLY8+TuWz" Content-Disposition: inline In-Reply-To: <20030416231756.GI79923@perrin.int.nxad.com> Subject: Re: FreeBSD Memory Pages Not Locked? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Apr 2003 23:23:35 -0000 --brEuL7wsLY8+TuWz Content-Type: text/plain; charset=unknown-8bit Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Di Wed, Apr 16, 2003 at 04:17:56PM -0700 da Sean Chittenden niv=EEsand: > > I recently installed "gpnupg" from the ports collection and > > upon running it (after the key generation), I found myself > > seeing the following error: > >=20 > > gpg: Warning: using insecure memory! > >=20 > > ... Those of you who are impatient and think that this is a > > gpg &| port specific problem, please be patient and read on ... > >=20 > > from: http://www.gnupg.org/documentation/faqs.html#q6.1 > > Here's what the GPG FAQ says:=20 >=20 > I believe it's because some of the pages required to use this program > are on disk in swap someplace. This is very off topic for > performance@, please ask on ports@ or questions@ next time. -sc Thank you. I shall re-post to questions. --$=EAr=EAciya Kurdistan=EE +--------------------------------------------------------------+ | Welat xwe ava nake, dest bidin hevdu, pist nedin tu dijmin=EE | | Riya azadiy=EA ne h=EAsan e, h=EAviya xwe bernedin, dema me | | n=EAz=EEk e. | | | | Hevalt=EE bi kes=EAn du r=FB nekin, hevalt=EE bi hevdu ra bikin | | Ne ji hevaltiya wan kes=EAn p=EAxwas =FB r=FB dir=EAj, ne bi wan | | kes=EAn xw=EEnperest, ne j=EE ji y=EAn din. | | | | -$=EAr=EAciya Kurdistan=EE | +--------------------------------------------------------------+ translation provided on request: sereciya at kurdistan.ath.cx --brEuL7wsLY8+TuWz Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (FreeBSD) iD8DBQE+neX32rX+NTHoe2QRAk4GAJ49TTct3g+dzdFQSsfmOl5ZsxDrDwCfW5uJ Ge/zonS/u3xzNKtbNBkVtXM= =Xdd7 -----END PGP SIGNATURE----- --brEuL7wsLY8+TuWz-- From owner-freebsd-performance@FreeBSD.ORG Wed Apr 16 16:37:26 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 5254A37B404 for ; Wed, 16 Apr 2003 16:37:26 -0700 (PDT) Received: from mail.yadt.co.uk (yadt.demon.co.uk [158.152.4.134]) by mx1.FreeBSD.org (Postfix) with SMTP id E802843FB1 for ; Wed, 16 Apr 2003 16:37:22 -0700 (PDT) (envelope-from davidt@yadt.co.uk) Received: (qmail 11680 invoked from network); 16 Apr 2003 23:37:20 -0000 Received: from gattaca.local.yadt.co.uk (HELO mail.gattaca.yadt.co.uk) (qmailr@10.0.0.2) by xfiles.yadt.co.uk with SMTP; 16 Apr 2003 23:37:20 -0000 Received: (qmail 50740 invoked by uid 1000); 16 Apr 2003 23:37:20 -0000 Date: Thu, 17 Apr 2003 00:37:20 +0100 From: David Taylor To: =?iso-8859-15?Q?S=EAr=EAciya_Kurdistan=EE?= Message-ID: <20030416233719.GA49658@gattaca.yadt.co.uk> References: <20030416222057.GC57404@kurdistan.ath.cx> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="gBBFr7Ir9EOA20Yy" Content-Disposition: inline In-Reply-To: <20030416222057.GC57404@kurdistan.ath.cx> User-Agent: Mutt/1.4.1i X-Mailman-Approved-At: Wed, 16 Apr 2003 16:47:50 -0700 cc: freebsd-questions@freebsd.org Subject: Re: FreeBSD Memory Pages Not Locked? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 16 Apr 2003 23:37:26 -0000 --gBBFr7Ir9EOA20Yy Content-Type: text/plain; charset=iso-8859-15 Content-Disposition: inline Content-Transfer-Encoding: quoted-printable [moved to -questions, and unsnipped for that reason] [note, I'm not ON -questions, so please CC me in any replies] On Wed, 16 Apr 2003, S=EAr=EAciya Kurdistan=EE wrote: > Hello, >=20 > I recently installed "gpnupg" from the ports collection and > upon running it (after the key generation), I found myself > seeing the following error: >=20 > gpg: Warning: using insecure memory! >=20 > ... Those of you who are impatient and think that this is a > gpg &| port specific problem, please be patient and read on ... >=20 > from: http://www.gnupg.org/documentation/faqs.html#q6.1 > Here's what the GPG FAQ says:=20 >=20 > "6.1) Why do I get "gpg: Warning: using insecure memory!" >=20 > On many systems this program should be installed as setuid(root). > This is necessary to lock memory pages. Locking memory pages > prevents the operating system from writing them to disk and > thereby keeping your secret keys really secret. If you get no > warning message about insecure memory your operating system > supports locking without being root. The program drops root > privileges as soon as locked memory is allocated." >=20 >=20 > So my question is: does FreeBSD really not have support for > locking memory pages? Not by non root users. =20 > if this is true, then what is the reason > that this has not yet been implemented, > is this not an important security feature? (I assume) because if any user could lock pages in memory, so that it could not be swapped, they could cause the system to run low on physical memory, resulting in a DoS (Denial of service) attack. =20 > otherwise... if FreeBSD does in fact have > support for locking memory pages, then > why am I getting this error message? Because you haven't made gpg setuid root (chmod u+s /usr/local/bin/gpg should achieve that -- but there are security considerations). You should either: accept that your passphrase/private key may end up on swap at some point; or set the program set-uid root, and accept that any security problems in gpg (before the point where it drops privileges) could result in your root account being comprimised (and the gpg binary being replaced with another one that e-mails your passphrase around the globe). The correct solution depends on how paranoid you are, who has access to your box, etc. =20 > If any of you have encountered this problem, and would like > to offer some help &| advice, you have a captive audience > of at least one, me! Most of this was explained in the FAQ that you posted, I'm not entirely sure how you didn't understand it, but possibly it's badly worded and i just intuitively understand it because I know the answer already. --=20 David Taylor davidt@yadt.co.uk "The future just ain't what it used to be" --gBBFr7Ir9EOA20Yy Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (FreeBSD) iD8DBQE+nekvfIqKXSsJ/xERAsvoAJwKLwq6bupPzD3z28V4HQIIxQlkPQCdEO7C W2V7oqTrnNBUgNlYhvVDPvY= =UkKE -----END PGP SIGNATURE----- --gBBFr7Ir9EOA20Yy-- From owner-freebsd-performance@FreeBSD.ORG Wed Apr 16 17:10:14 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 8D51837B401 for ; Wed, 16 Apr 2003 17:10:14 -0700 (PDT) Received: from walter.dfmm.org (walter.dfmm.org [209.151.233.240]) by mx1.FreeBSD.org (Postfix) with ESMTP id 01ADA43FDD for ; Wed, 16 Apr 2003 17:10:14 -0700 (PDT) (envelope-from jason@shalott.net) Received: (qmail 1559 invoked by uid 1000); 17 Apr 2003 00:10:13 -0000 Received: from localhost (sendmail-bs@127.0.0.1) by localhost with SMTP; 17 Apr 2003 00:10:13 -0000 Date: Wed, 16 Apr 2003 17:10:13 -0700 (PDT) From: Jason Stone X-X-Sender: To: =?unknown-8bit?Q?S=EAr=EAciya_Kurdistan=EE?= In-Reply-To: <20030416222057.GC57404@kurdistan.ath.cx> Message-ID: <20030416165844.A4074-100000@walter> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII cc: freebsd-performance@freebsd.org Subject: Re: FreeBSD Memory Pages Not Locked? X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 17 Apr 2003 00:10:14 -0000 -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 > I recently installed "gpnupg" from the ports collection and > upon running it (after the key generation), I found myself > seeing the following error: > > gpg: Warning: using insecure memory! 1) This is a question for freebsd-security, not freebsd-performance 2) Yes, freebsd does support locking pages in memory with mlock, but only root can call mlock. If you make gpg setuid root (chmod 4111 `which gpg`) then it will be able to mlock and the warning will go away. However, you must decide if that is a good security practice, because now bugs in gpg can be used to gain root on that machine, and if an attacker gains root, he gain just sniff your tty and get your passphrase next time you enter it. Additionally, other programs on the machine do not mlock sensitive data into core (think login, sshd, ssh-agent, etc), so you're already vulnerable to having sensitive data retrieved from swap. If having sensitive data retrieved from swap is really a concern for you, run freebsd-5 and use gbde to encrypt your whole swap partition. 3) Or, just add "no-secmem-warning" to your ~/.gnupg/options to silence the warning. It's really unlikely that an attack on unencrypted data in swap will ever affect you. -Jason -------------------------------------------------------------------------- Freud himself was a bit of a cold fish, and one cannot avoid the suspicion that he was insufficiently fondled when he was an infant. -- Ashley Montagu -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (FreeBSD) Comment: See https://private.idealab.com/public/jason/jason.gpg iD8DBQE+nfDlswXMWWtptckRAupFAKDtyHf26X3TsAJ6qh67rQHPqXIT6gCguXmA A5immbQ9tsm+aN40DXbCxek= =hllG -----END PGP SIGNATURE----- From owner-freebsd-performance@FreeBSD.ORG Fri Apr 18 05:44:35 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 3F9E737B401; Fri, 18 Apr 2003 05:44:35 -0700 (PDT) Received: from ns.purk.ee (ns.purk.ee [213.35.131.222]) by mx1.FreeBSD.org (Postfix) with ESMTP id 4717543FBF; Fri, 18 Apr 2003 05:44:34 -0700 (PDT) (envelope-from hostmaster@ns.purk.ee) Received: from [192.168.3.179] (unknown [192.168.3.179]) by ns.purk.ee (Postfix) with ESMTP id 9FFEAADE3; Fri, 18 Apr 2003 15:44:31 +0300 (EEST) From: argo To: aic7xxx@freebsd.org Content-Type: text/plain Organization: Message-Id: <1050669889.575.12.camel@station.purk.ee> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.2.0 Date: 18 Apr 2003 15:44:50 +0300 Content-Transfer-Encoding: 7bit cc: freebsd-performance@freebsd.org Subject: aic7892 trouble X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 18 Apr 2003 12:44:36 -0000 Hi folks, I have Adaptec U160 Host-Adapter with 2x Atlas 10k disks.When i copy large files (~700MB) then performance is pretty bad.I got only about 35MB/sec between those 2 disks.Theres no difference 4.x or 5 -current. I made couple of custom kernels with different options but no luck. Could that be related with bad cable issue? Thanks in advance. ahc0: port 0xa000-0xa0ff mem 0xe2024000-0xe2024fff irq 11 at device 9.0 on pci0 aic7892: Ultra160 Wide Channel A, SCSI Id=7, 32/253 SCBs da0 at ahc0 bus 0 target 0 lun 0 da0: Fixed Direct Access SCSI-3 device da0: 160.000MB/s transfers (80.000MHz, offset 127, 16bit), Tagged Queueing Enabled da0: 17537MB (35916548 512 byte sectors: 255H 63S/T 2235C) da1 at ahc0 bus 0 target 1 lun 0 da1: Fixed Direct Access SCSI-3 device da1: 160.000MB/s transfers (80.000MHz, offset 127, 16bit), Tagged Queueing Enabled da1: 17537MB (35916548 512 byte sectors: 255H 63S/T 2235C) From owner-freebsd-performance@FreeBSD.ORG Fri Apr 18 06:44:34 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 2CA5737B401; Fri, 18 Apr 2003 06:44:34 -0700 (PDT) Received: from sabre.velocet.net (sabre.velocet.net [216.138.209.205]) by mx1.FreeBSD.org (Postfix) with ESMTP id 9BCEF43FD7; Fri, 18 Apr 2003 06:44:33 -0700 (PDT) (envelope-from dgilbert@velocet.ca) Received: from trooper.velocet.ca (trooper.velocet.net [216.138.242.2]) by sabre.velocet.net (Postfix) with ESMTP id BC67713850F; Fri, 18 Apr 2003 09:44:32 -0400 (EDT) Received: by trooper.velocet.ca (Postfix, from userid 66) id 50A2474A74; Fri, 18 Apr 2003 09:44:32 -0400 (EDT) Received: by canoe.velocet.net (Postfix, from userid 101) id 55D1156762E; Fri, 18 Apr 2003 09:44:28 -0400 (EDT) From: David Gilbert MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-ID: <16032.316.217010.480864@canoe.velocet.net> Date: Fri, 18 Apr 2003 09:44:28 -0400 To: argo In-Reply-To: <1050669889.575.12.camel@station.purk.ee> References: <1050669889.575.12.camel@station.purk.ee> X-Mailer: VM 7.07 under 21.1 (patch 14) "Cuyahoga Valley" XEmacs Lucid cc: aic7xxx@freebsd.org cc: freebsd-performance@freebsd.org Subject: aic7892 trouble X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 18 Apr 2003 13:44:34 -0000 >>>>> "argo" == argo writes: argo> Hi folks, I have Adaptec U160 Host-Adapter with 2x Atlas 10k argo> disks.When i copy large files (~700MB) then performance is argo> pretty bad.I got only about 35MB/sec between those 2 argo> disks.Theres no difference 4.x or 5 -current. I made couple of argo> custom kernels with different options but no luck. Could that argo> be related with bad cable issue? That's about expected thruput for 18G 10k disks. If you buy 36G 10k disks, you'll likely get 70-ish meg/s ... +- how good your PCI bus is. Dave. -- ============================================================================ |David Gilbert, Velocet Communications. | Two things can only be | |Mail: dgilbert@velocet.net | equal if and only if they | |http://daveg.ca | are precisely opposite. | =========================================================GLO================ From owner-freebsd-performance@FreeBSD.ORG Fri Apr 18 07:41:59 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 2E2F637B401; Fri, 18 Apr 2003 07:41:59 -0700 (PDT) Received: from adsl-63-198-35-122.dsl.snfc21.pacbell.net (adsl-63-198-35-122.dsl.snfc21.pacbell.net [63.198.35.122]) by mx1.FreeBSD.org (Postfix) with ESMTP id 5073943FE5; Fri, 18 Apr 2003 07:41:58 -0700 (PDT) (envelope-from j_guojun@lbl.gov) Received: from lbl.gov (localhost.pacbell.net [127.0.0.1]) by spode.pacbell.net (8.12.8p1/8.12.8) with ESMTP id h3I5C2u6000430; Thu, 17 Apr 2003 22:12:02 -0700 (PDT) (envelope-from j_guojun@lbl.gov) Sender: jin@adsl-63-198-35-122.dsl.snfc21.pacbell.net Message-ID: <3E9F8922.1166ABA8@lbl.gov> Date: Thu, 17 Apr 2003 22:12:02 -0700 From: "Jin Guojun [NCS]" X-Mailer: Mozilla 4.76 [en] (X11; U; FreeBSD 4.8-RELEASE i386) X-Accept-Language: zh, zh-CN, en-US, en MIME-Version: 1.0 To: freebsd-performance@freebsd.org Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Content-Filtered-By: Mailman/MimeDel 2.1.1 cc: freebsd-hackers@freebsd.org Subject: patch for test (Was: tcp_output starving -- is due to mbuf get delay?) X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 18 Apr 2003 14:41:59 -0000 I have modified the sockbuf and mbuf operation to double the throughput over high bandwidth delay product path. The patch is available at: http://www-didc.lbl.gov/~jin/network/lion/content.html#FreeBSD_Patches The current modification is for tcp transmission only. I have adapted some code of uipc_socket2.c from Sam Leffler http://www.freebsd.org/~sam/thorpe-stable.patch for tcp receiver, but it has not been tested yet, so the tcp_input.p is empty. I ignored all record chain (m_nextpkt) related code. The details is explained at http://www-didc.lbl.gov/~jin/network/lion/content.html#BSDMbuf Once the tcp_input code is tested, I will submit the patch to bugs@freebsd.org. I may submit the patch regardless if tcp_input code works or not, because the tcp sender (server) is more important in high-speed network than the receiver (client). It is appreciated if any one can verify the patch and provide feedback. -- ------------ Jin Guojun ----------- v --- j_guojun@lbl.gov --- Distributed Systems Department http://www.itg.lbl.gov/~jin M/S 50B-2239 Ph#:(510) 486-7531 Fax: 486-6363 Lawrence Berkeley National Laboratory, Berkeley, CA 94720 From owner-freebsd-performance@FreeBSD.ORG Fri Apr 18 08:29:32 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id D55A637B404 for ; Fri, 18 Apr 2003 08:29:32 -0700 (PDT) Received: from relay.pair.com (relay.pair.com [209.68.1.20]) by mx1.FreeBSD.org (Postfix) with SMTP id B7FD443FBD for ; Fri, 18 Apr 2003 08:29:30 -0700 (PDT) (envelope-from silby@silby.com) Received: (qmail 13113 invoked from network); 18 Apr 2003 15:29:29 -0000 Received: from niwun.pair.com (HELO localhost) (209.68.2.70) by relay.pair.com with SMTP; 18 Apr 2003 15:29:29 -0000 X-pair-Authenticated: 209.68.2.70 Date: Thu, 17 Apr 2003 22:25:30 -0500 (CDT) From: Mike Silbersack To: "Jin Guojun [NCS]" In-Reply-To: <3E9F8922.1166ABA8@lbl.gov> Message-ID: <20030417222124.D3757@odysseus.silby.com> References: <3E9F8922.1166ABA8@lbl.gov> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII cc: freebsd-hackers@freebsd.org cc: freebsd-performance@freebsd.org Subject: Re: patch for test (Was: tcp_output starving -- is due to mbuf get delay?) X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 18 Apr 2003 15:29:33 -0000 On Thu, 17 Apr 2003, Jin Guojun [NCS] wrote: > I have modified the sockbuf and mbuf operation to double the throughput over > high bandwidth delay product path. > > The patch is available at: > http://www-didc.lbl.gov/~jin/network/lion/content.html#FreeBSD_Patches > > The current modification is for tcp transmission only. Impressive results, it sounds like we definitely need this improvement (in some form or another.) > Once the tcp_input code is tested, I will submit the patch to bugs@freebsd.org. > I may submit the patch regardless if tcp_input code works or not, because the > tcp > sender (server) is more important in high-speed network than the receiver > (client). > > It is appreciated if any one can verify the patch and provide feedback. > > -- > ------------ Jin Guojun ----------- v --- j_guojun@lbl.gov --- But I don't have time to look at it. :) Keep the patches and benchmarks coming, Mike "Silby" Silbersack From owner-freebsd-performance@FreeBSD.ORG Fri Apr 18 09:58:12 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 5A83B37B405; Fri, 18 Apr 2003 09:58:12 -0700 (PDT) Received: from samson.dc.luth.se (samson.dc.luth.se [130.240.112.30]) by mx1.FreeBSD.org (Postfix) with ESMTP id DE59D43FE9; Fri, 18 Apr 2003 09:58:09 -0700 (PDT) (envelope-from bj@dc.luth.se) Received: from dc.luth.se (root@bompe.dc.luth.se [130.240.60.42]) by samson.dc.luth.se (8.12.5/8.12.5) with ESMTP id h3IGw8LG012691; Fri, 18 Apr 2003 18:58:08 +0200 (MET DST) Received: from bompe.dc.luth.se (bj@localhost.dc.luth.se [127.0.0.1]) by dc.luth.se (8.12.6/8.11.3) with ESMTP id h3IGw72F021961; Fri, 18 Apr 2003 18:58:07 +0200 (CEST) (envelope-from bj@bompe.dc.luth.se) Message-Id: <200304181658.h3IGw72F021961@dc.luth.se> X-Mailer: exmh version 2.5 07/13/2001 with nmh-1.0.4 To: "Jin Guojun [NCS]" In-reply-to: Your message of Thu, 17 Apr 2003 22:12:02 PDT. <3E9F8922.1166ABA8@lbl.gov> Dcc: From: Borje Josefsson X-Disposition-notification-to: Borje.Josefsson@dc.luth.se X-uri: http://www.dc.luth.se/~bj/index.html Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable Date: Fri, 18 Apr 2003 18:58:07 +0200 Sender: bj@dc.luth.se cc: freebsd-hackers@freebsd.org cc: freebsd-performance@freebsd.org Subject: Re: patch for test (Was: tcp_output starving -- is due to mbuf get delay?) X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: bj@dc.luth.se List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 18 Apr 2003 16:58:12 -0000 On Thu, 17 Apr 2003 22:12:02 PDT "Jin Guojun [NCS]" wrote: > I have modified the sockbuf and mbuf operation to double the throughput= over > high bandwidth delay product path. > = > The patch is available at: > http://www-didc.lbl.gov/~jin/network/lion/content.html#FreeBSD_Patc= hes > = > The current modification is for tcp transmission only. > = > I have adapted some code of uipc_socket2.c from Sam Leffler > http://www.freebsd.org/~sam/thorpe-stable.patch > = > for tcp receiver, but it has not been tested yet, so the tcp_input.p is= empty. > = > I ignored all record chain (m_nextpkt) related code. The details is exp= lained at > = > http://www-didc.lbl.gov/~jin/network/lion/content.html#BSDMbuf > = > Once the tcp_input code is tested, I will submit the patch to bugs@free= bsd.org. > I may submit the patch regardless if tcp_input code works or not, becau= se the > tcp > sender (server) is more important in high-speed network than the receiv= er > (client). > = > It is appreciated if any one can verify the patch and provide feedback.= OK. I have now tried this patch on a newly-installed 4.8R. The patch = applied fine. When the sysctl net.inet.tcp.liondmask is unset, everything= = seems OK, but when setting it to 7 (as specified with the patch = instructions) i get: Fatal trap 12: page fault while in kernel mode. (I could write down all the stuff on addresses etc if it makes sense) when I run ttcp to test the performance. This is repeatable. I'm willing to test more, if someone provides me with some hints on what = to do. --B=F6rje From owner-freebsd-performance@FreeBSD.ORG Fri Apr 18 13:04:29 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 93C0D37B401; Fri, 18 Apr 2003 13:04:29 -0700 (PDT) Received: from postal2.lbl.gov (postal2.lbl.gov [131.243.248.26]) by mx1.FreeBSD.org (Postfix) with ESMTP id C122D43F93; Fri, 18 Apr 2003 13:04:28 -0700 (PDT) (envelope-from j_guojun@lbl.gov) Received: from postal2.lbl.gov (localhost [127.0.0.1]) by postal2.lbl.gov (8.12.9/8.12.9) with ESMTP id h3IK4RCY007628; Fri, 18 Apr 2003 13:04:28 -0700 (PDT) Received: from lbl.gov (gracie.lbl.gov [131.243.2.175]) by postal2.lbl.gov (8.12.9/8.12.9) with ESMTP id h3IK4O1b007621; Fri, 18 Apr 2003 13:04:27 -0700 (PDT) Sender: jin@lbl.gov Message-ID: <3EA05A48.D9BB85E7@lbl.gov> Date: Fri, 18 Apr 2003 13:04:24 -0700 From: "Jin Guojun [DSD]" X-Mailer: Mozilla 4.76 [en] (X11; U; FreeBSD 4.7-RELEASE i386) X-Accept-Language: zh, zh-CN, en MIME-Version: 1.0 To: bj@dc.luth.se References: <200304181658.h3IGw72F021961@dc.luth.se> Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit cc: freebsd-hackers@freebsd.org cc: freebsd-performance@freebsd.org Subject: Re: patch for test (Was: tcp_output starving -- is due to mbuf get delay?) X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 18 Apr 2003 20:04:29 -0000 Opps, there was a bad file -- tcp_input.p -- which is not working yet. Also, a patch file -- tcp_usrreq.p -- was missing. I will take the tcp_input.p out and put tcp_usrreq.p in. When it is finished, I will send another mail out. -Jin Borje Josefsson wrote: > On Thu, 17 Apr 2003 22:12:02 PDT "Jin Guojun [NCS]" wrote: > > > I have modified the sockbuf and mbuf operation to double the throughput over > > high bandwidth delay product path. > > > > The patch is available at: > > http://www-didc.lbl.gov/~jin/network/lion/content.html#FreeBSD_Patches > > > > The current modification is for tcp transmission only. > > > > I have adapted some code of uipc_socket2.c from Sam Leffler > > http://www.freebsd.org/~sam/thorpe-stable.patch > > > > for tcp receiver, but it has not been tested yet, so the tcp_input.p is empty. > > > > I ignored all record chain (m_nextpkt) related code. The details is explained at > > > > http://www-didc.lbl.gov/~jin/network/lion/content.html#BSDMbuf > > > > Once the tcp_input code is tested, I will submit the patch to bugs@freebsd.org. > > I may submit the patch regardless if tcp_input code works or not, because the > > tcp > > sender (server) is more important in high-speed network than the receiver > > (client). > > > > It is appreciated if any one can verify the patch and provide feedback. > > OK. I have now tried this patch on a newly-installed 4.8R. The patch > applied fine. When the sysctl net.inet.tcp.liondmask is unset, everything > seems OK, but when setting it to 7 (as specified with the patch > instructions) i get: > > Fatal trap 12: page fault while in kernel mode. > (I could write down all the stuff on addresses etc if it makes sense) > > when I run ttcp to test the performance. > > This is repeatable. > > I'm willing to test more, if someone provides me with some hints on what > to do. > > --Börje From owner-freebsd-performance@FreeBSD.ORG Sat Apr 19 06:05:32 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 006BB37B401; Sat, 19 Apr 2003 06:05:32 -0700 (PDT) Received: from samson.dc.luth.se (samson.dc.luth.se [130.240.112.30]) by mx1.FreeBSD.org (Postfix) with ESMTP id 34FCE43F75; Sat, 19 Apr 2003 06:05:30 -0700 (PDT) (envelope-from bj@dc.luth.se) Received: from dc.luth.se (root@bompe.dc.luth.se [130.240.60.42]) by samson.dc.luth.se (8.12.5/8.12.5) with ESMTP id h3JD5TLG019266; Sat, 19 Apr 2003 15:05:29 +0200 (MET DST) Received: from bompe.dc.luth.se (bj@localhost.dc.luth.se [127.0.0.1]) by dc.luth.se (8.12.6/8.11.3) with ESMTP id h3JD5S2F026929; Sat, 19 Apr 2003 15:05:28 +0200 (CEST) (envelope-from bj@bompe.dc.luth.se) Message-Id: <200304191305.h3JD5S2F026929@dc.luth.se> X-Mailer: exmh version 2.5 07/13/2001 with nmh-1.0.4 To: "Jin Guojun [DSD]" In-reply-to: Your message of Fri, 18 Apr 2003 13:04:24 PDT. <3EA05A48.D9BB85E7@lbl.gov> Dcc: From: Borje Josefsson X-Disposition-notification-to: Borje.Josefsson@dc.luth.se X-uri: http://www.dc.luth.se/~bj/index.html Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable Date: Sat, 19 Apr 2003 15:05:28 +0200 Sender: bj@dc.luth.se cc: freebsd-hackers@freebsd.org cc: freebsd-performance@freebsd.org Subject: Re: patch for test (Was: tcp_output starving -- is due to mbuf get delay?) X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: bj@dc.luth.se List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 19 Apr 2003 13:05:32 -0000 Hmm. I'm not sure if I misunderstood if this was ready for another test = run or not. Anyhow - I took the new patch .tgz (which, btw, still had = tcp_input.p in it). I applied the patches (except tcp_input) and tested. Now I get: Panic: bad cur_off 00000 m_p 0xc0a7f400 0xc0a7f400 my_off 0 1448 cc 3407144 As usual, I'm willing to test more when there are an update available. --B=F6rje On Fri, 18 Apr 2003 13:04:24 PDT "Jin Guojun [DSD]" wrote: > Opps, there was a bad file -- tcp_input.p -- which is not working yet. > Also, a patch file -- tcp_usrreq.p -- was missing. > = > I will take the tcp_input.p out and put tcp_usrreq.p in. > When it is finished, I will send another mail out. > = > -Jin > = > Borje Josefsson wrote: > = > > On Thu, 17 Apr 2003 22:12:02 PDT "Jin Guojun [NCS]" wrote: > > > > > I have modified the sockbuf and mbuf operation to double the throug= hput over > > > high bandwidth delay product path. > > > > > > The patch is available at: > > > http://www-didc.lbl.gov/~jin/network/lion/content.html#FreeBSD_= Patches > > > > > > The current modification is for tcp transmission only. > > > > > > I have adapted some code of uipc_socket2.c from Sam Leffler > > > http://www.freebsd.org/~sam/thorpe-stable.patch > > > > > > for tcp receiver, but it has not been tested yet, so the tcp_input.= p is empty. > > > > > > I ignored all record chain (m_nextpkt) related code. The details is= explained at > > > > > > http://www-didc.lbl.gov/~jin/network/lion/content.html#BSDMbuf > > > > > > Once the tcp_input code is tested, I will submit the patch to bugs@= freebsd.org. > > > I may submit the patch regardless if tcp_input code works or not, b= ecause the > > > tcp > > > sender (server) is more important in high-speed network than the re= ceiver > > > (client). > > > > > > It is appreciated if any one can verify the patch and provide feedb= ack. > > > > OK. I have now tried this patch on a newly-installed 4.8R. The patch > > applied fine. When the sysctl net.inet.tcp.liondmask is unset, everyt= hing > > seems OK, but when setting it to 7 (as specified with the patch > > instructions) i get: > > > > Fatal trap 12: page fault while in kernel mode. > > (I could write down all the stuff on addresses etc if it makes sens= e) > > > > when I run ttcp to test the performance. > > > > This is repeatable. > > > > I'm willing to test more, if someone provides me with some hints on w= hat > > to do. > > > > --B=F6rje > = From owner-freebsd-performance@FreeBSD.ORG Sat Apr 19 17:46:49 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 7E08437B401; Sat, 19 Apr 2003 17:46:49 -0700 (PDT) Received: from snark.ratmir.ru (snark.ratmir.ru [213.24.248.177]) by mx1.FreeBSD.org (Postfix) with ESMTP id B8D6443FBD; Sat, 19 Apr 2003 17:46:45 -0700 (PDT) (envelope-from freebsd@snark.ratmir.ru) Received: from snark.ratmir.ru (freebsd@localhost [127.0.0.1]) by snark.ratmir.ru (8.12.9/8.12.9) with ESMTP id h3K0kfC2052304; Sun, 20 Apr 2003 04:46:42 +0400 (MSD) (envelope-from freebsd@snark.ratmir.ru) Received: (from freebsd@localhost) by snark.ratmir.ru (8.12.9/8.12.9/Submit) id h3K0kdjN052299; Sun, 20 Apr 2003 04:46:39 +0400 (MSD) Date: Sun, 20 Apr 2003 04:46:39 +0400 From: Alex Semenyaka To: freebsd-performance@freebsd.org, freebsd-hackers@freebsd.org, freebsd-current@freebsd.org, freebsd-standards@freebsd.org Message-ID: <20030420004639.GA52081@snark.ratmir.ru> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable User-Agent: Mutt/1.5.4i X-Mailman-Approved-At: Sat, 19 Apr 2003 18:25:01 -0700 Subject: tjr@@freebsd.org, imp@freebsd.org, ru@freebsd.org X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 20 Apr 2003 00:46:50 -0000 Hi, I've heard a lot of recommendation about where I shoud show my patch and the results of performance changes because of it. So I gathered all advices into To: and CC: fields. Sorry, if there are unnecessary addresses (which then?). Also as I was told the final patch I've send as the PR. Brief description what was done: I've chanched the arithmitics in the /bin/= sh =66rom 32 bits to 64 bits. There are some doubts that it conforms to the standards: it does, I have send a quotations to -standards, there were no objections. Couple of people advuces me to use intmax_t and %jd - I've rewr= itten the patch, now there is those species instead of long long and %qd. The last question was performance, I will show the results of measurements below. The patch can be found in the attach to this message or in the PR bin/51171. Also I've added the overflow control to the addition, substraction and multiplication. There is pretty small overhead (see data below) so I decided it harmless. First, here is the processor description in the box I've run the tests. CPU: Intel Pentium III (1002.28-MHz 686-class CPU) Origin =3D "GenuineIntel" Id =3D 0x68a Stepping =3D 10 Features=3D0x387f9ff real memory =3D 268435456 (262144K bytes) avail memory =3D 256475136 (250464K bytes) The tests was a set of 12 scripts with the following structure: for x in `jot 300000` do done and operations were "" (just no operation, empty loop), i=3D$(($i+1)), i=3D$(($i+$m)) (here is _two_ variables), i=3D$(($i*1)), i=3D$(($i*$m)), i= =3D$(($i/1)), i=3D$(($i/$m)), i=3D$(($i<<1)), i=3D$(($i<<$m)), i=3D$(($i||1)), i=3D$(($i|= |$m)), i=3D$((~$i)). In the tests with two variables second one ($m) was always eq= ual to 1. Starting value of $i was also 1.` I've run those twelve scripts first with the 64-bit shell without overflow control and compared with 32-bit shell: Test Time Difference Arith. No Old New Abs %% operation ----+----------------+-------------------------------- 0 1.35 1.36 0.04 0.99% no op 1 5.12 5.36 0.72 4.69% i=3D$(($i+1)) 2 5.17 5.43 0.78 5.03% i=3D$(($i+$a)) 3 4.39 4.57 0.55 4.18% i=3D$(($i*1)) 4 4.43 4.64 0.64 4.82% i=3D$(($i*$m)) 5 4.40 4.62 0.67 5.08% i=3D$(($i/1)) 6 4.45 4.68 0.68 5.09% i=3D$(($i/$m)) 7 5.20 6.37 3.51 22.50% i=3D$(($i<<1)) 8 5.25 6.42 3.51 22.27% i=3D$(($i<<$m)) 9 4.52 4.67 0.45 3.32% i=3D$(($i||1)) 10 4.58 4.73 0.44 3.20% i=3D$(($i||$m)) 11 4.30 4.38 0.25 1.94% i=3D$((~$i)) As you can see, even for arithmetic-only script the overhead is not too big except with one case: shift operation. I decided to investigate is it usual script operation. I've went through all scripts I could find in my FreeBSD box. I've searched them with "locate .sh | grep '\.sh$'". There were a lot of them: $ locate .sh | grep '\.sh$' | wc -l 1637 But there was no any script that uses the shift operation. Good, but not enough. I've take the script that uses arithmetics and do some other job, ttfadmin.sh from the Abiword package. I've run in 10000 times in the loop with both (64-bit and 32-bit) shells. As an argument it received empty directory so no work has been done, just run, check pars, found no files, exit. It takes 65.35 seconds in the first case and 65.30 second in the seco= nd one. So the the time that arithmetics takes during the real script execution is too small in comparison to total running time (obviously: arithmetics is in-core calculations while any script usually run some external programs etc, and at least I/O is involved). Then I've run tests with the addition and multiplication to compare 64-bit shells compiled without the overview control and with it. In the last case I've turned that control on by setting corresponding command line option (-O in my patch). Here are results: 1 5.24 5.26 0.08 0.51% 2 5.31 5.34 0.08 0.50% 3 4.53 4.57 0.14 1.03% 4 4.60 4.65 0.14 1.01% You can see, the difference is just negligible.=20 So I hope I've answered all questions were asked about that 64-bit modifica= tion. All apprehensions are, I hope, dismissed. Probably, after the last look of the responsible person it can be committed, am I right? Thanks! SY, Alex From owner-freebsd-performance@FreeBSD.ORG Sat Apr 19 18:10:50 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 9572837B401; Sat, 19 Apr 2003 18:10:50 -0700 (PDT) Received: from snark.ratmir.ru (snark.ratmir.ru [213.24.248.177]) by mx1.FreeBSD.org (Postfix) with ESMTP id E3AC643FE1; Sat, 19 Apr 2003 18:10:46 -0700 (PDT) (envelope-from freebsd@snark.ratmir.ru) Received: from snark.ratmir.ru (freebsd@localhost [127.0.0.1]) by snark.ratmir.ru (8.12.9/8.12.9) with ESMTP id h3K1AeC2052397; Sun, 20 Apr 2003 05:10:41 +0400 (MSD) (envelope-from freebsd@snark.ratmir.ru) Received: (from freebsd@localhost) by snark.ratmir.ru (8.12.9/8.12.9/Submit) id h3K1Ad4h052396; Sun, 20 Apr 2003 05:10:39 +0400 (MSD) Date: Sun, 20 Apr 2003 05:10:39 +0400 From: Alex Semenyaka To: freebsd-performance@freebsd.org, freebsd-hackers@freebsd.org, freebsd-current@freebsd.org, freebsd-standards@freebsd.org Message-ID: <20030420011039.GC52081@snark.ratmir.ru> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable User-Agent: Mutt/1.5.4i X-Mailman-Approved-At: Sat, 19 Apr 2003 18:25:01 -0700 cc: ru@freebsd.org cc: imp@freebsd.org cc: tjr@freebsd.org Subject: /bin/sh and 32-bit arithmetics [CORRECTED] X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 20 Apr 2003 01:10:51 -0000 I'm EXTREMELY sorry about the previous wrong posting, just mixed the header fields. Sorry again! Here is the correct one. Especially sorry to those guys whose addresses get into the Subject line. J= ust a mistake, nothing personal :( ---------------------------------------------------------------------------= -- Hi, I've heard a lot of recommendation about where I shoud show my patch and the results of performance changes because of it. So I gathered all advices into To: and CC: fields. Sorry, if there are unnecessary addresses (which then?). Also as I was told the final patch I've send as the PR. Brief description what was done: I've chanched the arithmitics in the /bin/= sh =66rom 32 bits to 64 bits. There are some doubts that it conforms to the standards: it does, I have send a quotations to -standards, there were no objections. Couple of people advuces me to use intmax_t and %jd - I've rewr= itten the patch, now there is those species instead of long long and %qd. The last question was performance, I will show the results of measurements below. The patch can be found in the attach to this message or in the PR bin/51171. Also I've added the overflow control to the addition, substraction and multiplication. There is pretty small overhead (see data below) so I decided it harmless. First, here is the processor description in the box I've run the tests. CPU: Intel Pentium III (1002.28-MHz 686-class CPU) Origin =3D "GenuineIntel" Id =3D 0x68a Stepping =3D 10 Features=3D0x387f9ff real memory =3D 268435456 (262144K bytes) avail memory =3D 256475136 (250464K bytes) The tests was a set of 12 scripts with the following structure: for x in `jot 300000` do done and operations were "" (just no operation, empty loop), i=3D$(($i+1)), i=3D$(($i+$m)) (here is _two_ variables), i=3D$(($i*1)), i=3D$(($i*$m)), i= =3D$(($i/1)), i=3D$(($i/$m)), i=3D$(($i<<1)), i=3D$(($i<<$m)), i=3D$(($i||1)), i=3D$(($i|= |$m)), i=3D$((~$i)). In the tests with two variables second one ($m) was always eq= ual to 1. Starting value of $i was also 1.` I've run those twelve scripts first with the 64-bit shell without overflow control and compared with 32-bit shell: Test Time Difference Arith. No Old New Abs %% operation ----+----------------+-------------------------------- 0 1.35 1.36 0.04 0.99% no op 1 5.12 5.36 0.72 4.69% i=3D$(($i+1)) 2 5.17 5.43 0.78 5.03% i=3D$(($i+$a)) 3 4.39 4.57 0.55 4.18% i=3D$(($i*1)) 4 4.43 4.64 0.64 4.82% i=3D$(($i*$m)) 5 4.40 4.62 0.67 5.08% i=3D$(($i/1)) 6 4.45 4.68 0.68 5.09% i=3D$(($i/$m)) 7 5.20 6.37 3.51 22.50% i=3D$(($i<<1)) 8 5.25 6.42 3.51 22.27% i=3D$(($i<<$m)) 9 4.52 4.67 0.45 3.32% i=3D$(($i||1)) 10 4.58 4.73 0.44 3.20% i=3D$(($i||$m)) 11 4.30 4.38 0.25 1.94% i=3D$((~$i)) As you can see, even for arithmetic-only script the overhead is not too big except with one case: shift operation. I decided to investigate is it usual script operation. I've went through all scripts I could find in my FreeBSD box. I've searched them with "locate .sh | grep '\.sh$'". There were a lot of them: $ locate .sh | grep '\.sh$' | wc -l 1637 But there was no any script that uses the shift operation. Good, but not enough. I've take the script that uses arithmetics and do some other job, ttfadmin.sh from the Abiword package. I've run in 10000 times in the loop with both (64-bit and 32-bit) shells. As an argument it received empty directory so no work has been done, just run, check pars, found no files, exit. It takes 65.35 seconds in the first case and 65.30 second in the seco= nd one. So the the time that arithmetics takes during the real script execution is too small in comparison to total running time (obviously: arithmetics is in-core calculations while any script usually run some external programs etc, and at least I/O is involved). Then I've run tests with the addition and multiplication to compare 64-bit shells compiled without the overview control and with it. In the last case I've turned that control on by setting corresponding command line option (-O in my patch). Here are results: 1 5.24 5.26 0.08 0.51% 2 5.31 5.34 0.08 0.50% 3 4.53 4.57 0.14 1.03% 4 4.60 4.65 0.14 1.01% You can see, the difference is just negligible.=20 So I hope I've answered all questions were asked about that 64-bit modifica= tion. All apprehensions are, I hope, dismissed. Probably, after the last look of the responsible person it can be committed, am I right? Thanks! SY, Alex From owner-freebsd-performance@FreeBSD.ORG Sat Apr 19 18:34:05 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id DCE4537B401; Sat, 19 Apr 2003 18:34:05 -0700 (PDT) Received: from snark.ratmir.ru (snark.ratmir.ru [213.24.248.177]) by mx1.FreeBSD.org (Postfix) with ESMTP id B53A043F3F; Sat, 19 Apr 2003 18:34:03 -0700 (PDT) (envelope-from freebsd@snark.ratmir.ru) Received: from snark.ratmir.ru (freebsd@localhost [127.0.0.1]) by snark.ratmir.ru (8.12.9/8.12.9) with ESMTP id h3K1Y0C2073365; Sun, 20 Apr 2003 05:34:00 +0400 (MSD) (envelope-from freebsd@snark.ratmir.ru) Received: (from freebsd@localhost) by snark.ratmir.ru (8.12.9/8.12.9/Submit) id h3K1Y0gn073364; Sun, 20 Apr 2003 05:34:00 +0400 (MSD) Date: Sun, 20 Apr 2003 05:34:00 +0400 From: Alex Semenyaka To: Alex Semenyaka Message-ID: <20030420013400.GB52428@snark.ratmir.ru> References: <20030420011039.GC52081@snark.ratmir.ru> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20030420011039.GC52081@snark.ratmir.ru> User-Agent: Mutt/1.5.4i cc: freebsd-current@freebsd.org cc: ru@freebsd.org cc: tjr@freebsd.org cc: imp@freebsd.org cc: freebsd-hackers@freebsd.org cc: freebsd-performance@freebsd.org cc: freebsd-standards@freebsd.org Subject: Re: /bin/sh and 32-bit arithmetics [CORRECTED] X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 20 Apr 2003 01:34:06 -0000 ...ghmmm... sorry and sorry... and here the patch, it is pretty small: diff -u -r -U 2 -b ../sh.old/Makefile ./Makefile --- ../sh.old/Makefile Sat Apr 19 23:22:31 2003 +++ ./Makefile Sun Apr 20 02:47:41 2003 @@ -19,5 +19,5 @@ LFLAGS= -8 # 8-bit lex scanner for arithmetic -CFLAGS+=-DSHELL -I. -I${.CURDIR} +CFLAGS+=-DSHELL -I. -I${.CURDIR} -DOVERFLOW # for debug: #CFLAGS+= -g -DDEBUG=2 -pg diff -u -r -U 2 -b ../sh.old/arith.h ./arith.h --- ../sh.old/arith.h Fri Jul 19 08:38:51 2002 +++ ./arith.h Sun Apr 20 02:37:37 2003 @@ -35,4 +35,7 @@ */ -int arith(char *); +/* XXX some day probably should go to /usr/include/machine/_inttypes.h */ +#define MAXINT_LEN 20 + +intmax_t arith(char *); int expcmd(int , char **); diff -u -r -U 2 -b ../sh.old/arith.y ./arith.y --- ../sh.old/arith.y Fri Jul 19 08:38:51 2002 +++ ./arith.y Sun Apr 20 03:52:31 2003 @@ -1,2 +1,13 @@ +%{ +#include +#ifdef OVERFLOW +#include "options.h" +#endif + +#define YYSTYPE intmax_t + +static intmax_t arith_res; +%} + %token ARITH_NUM ARITH_LPAREN ARITH_RPAREN @@ -15,5 +26,6 @@ exp: expr = { - return ($1); + arith_res = $1; + return (0); } ; @@ -34,7 +46,25 @@ | expr ARITH_LSHIFT expr = { $$ = $1 << $3; } | expr ARITH_RSHIFT expr = { $$ = $1 >> $3; } - | expr ARITH_ADD expr = { $$ = $1 + $3; } - | expr ARITH_SUB expr = { $$ = $1 - $3; } - | expr ARITH_MUL expr = { $$ = $1 * $3; } + | expr ARITH_ADD expr = { + $$ = $1 + $3; +#ifdef OVERFLOW + if (Oflag && is_add_overflow($1, $3, $$)) + yyerror("overflow in"); +#endif + } + | expr ARITH_SUB expr = { + $$ = $1 - $3; +#ifdef OVERFLOW + if (Oflag && is_add_overflow($1, -$3, $$)) + yyerror("overflow in"); +#endif + } + | expr ARITH_MUL expr = { + $$ = $1 * $3; +#ifdef OVERFLOW + if (Oflag && $$/$1 != $3 ) + yyerror("overflow in"); +#endif + } | expr ARITH_DIV expr = { if ($3 == 0) @@ -110,16 +140,24 @@ int -arith(char *s) +is_add_overflow(intmax_t a, intmax_t b, intmax_t s) { - long result; + if (a > 0 && b > 0 && s <= 0) + return 1; + if (a < 0 && b < 0 && s >= 0) + return 1; + return 0; +} +intmax_t +arith(char *s) +{ arith_buf = arith_startbuf = s; INTOFF; - result = yyparse(); + yyparse(); arith_lex_reset(); /* reprime lex */ INTON; - return (result); + return (arith_res); } @@ -143,5 +181,5 @@ char *concat; char **ap; - long i; + intmax_t i; if (argc > 1) { @@ -168,5 +206,5 @@ i = arith(p); - out1fmt("%ld\n", i); + out1fmt("%jd\n", i); return (! i); } diff -u -r -U 2 -b ../sh.old/arith_lex.l ./arith_lex.l --- ../sh.old/arith_lex.l Fri Jul 19 08:38:51 2002 +++ ./arith_lex.l Sun Apr 20 02:37:37 2003 @@ -44,8 +44,9 @@ #endif /* not lint */ +#include #include "y.tab.h" #include "error.h" -extern int yylval; +extern intmax_t yylval; extern char *arith_buf, *arith_startbuf; #undef YY_INPUT @@ -57,5 +58,5 @@ %% [ \t\n] { ; } -[0-9]+ { yylval = atol(yytext); return(ARITH_NUM); } +[0-9]+ { yylval = strtoll(yytext, NULL, 10); return(ARITH_NUM); } "(" { return(ARITH_LPAREN); } ")" { return(ARITH_RPAREN); } diff -u -r -U 2 -b ../sh.old/expand.c ./expand.c --- ../sh.old/expand.c Fri Jan 17 14:37:03 2003 +++ ./expand.c Sun Apr 20 02:37:37 2003 @@ -46,4 +46,5 @@ #include #include +#include #include #include @@ -367,5 +368,5 @@ { char *p, *start; - int result; + intmax_t result; int begoff; int quotes = flag & (EXP_FULL | EXP_CASE | EXP_REDIR); @@ -383,8 +384,8 @@ * characters have to be processed left to right. */ -#if INT_MAX / 1000000000 >= 10 || INT_MIN / 1000000000 <= -10 -#error "integers with more than 10 digits are not supported" -#endif - CHECKSTRSPACE(12 - 2, expdest); +//#if INT_MAX / 1000000000 >= 10 || INT_MIN / 1000000000 <= -10 +//#error "integers with more than 10 digits are not supported" +//#endif + CHECKSTRSPACE(MAXINT_LEN, expdest); USTPUTC('\0', expdest); start = stackblock(); @@ -408,5 +409,5 @@ rmescapes(p+2); result = arith(p+2); - fmtstr(p, 12, "%d", result); + fmtstr(p, MAXINT_LEN + 2, "%qd", result); while (*p++) ; diff -u -r -U 2 -b ../sh.old/options.h ./options.h --- ../sh.old/options.h Tue Aug 27 05:36:28 2002 +++ ./options.h Sun Apr 20 02:24:46 2003 @@ -67,6 +67,13 @@ #define Tflag optlist[16].val #define Pflag optlist[17].val +#ifdef OVERFLOW +#define Oflag optlist[18].val +#endif +#ifdef OVERFLOW +#define NOPTS 19 +#else #define NOPTS 18 +#endif struct optent { @@ -96,4 +103,7 @@ { "trapsasync", 'T', 0 }, { "physical", 'P', 0 }, +#ifdef OVERFLOW + { "overflow", 'O', 0 }, +#endif }; #else diff -u -r -U 2 -b ../sh.old/sh.1 ./sh.1 --- ../sh.old/sh.1 Tue Feb 25 13:27:12 2003 +++ ./sh.1 Sun Apr 20 02:52:02 2003 @@ -44,5 +44,5 @@ .Sh SYNOPSIS .Nm -.Op Fl /+abCEefIimnPpsTuVvx +.Op Fl /+abCEefIimnOPpsTuVvx .Op Fl /+o Ar longname .Op Fl c Ar string @@ -226,4 +226,6 @@ execute them. This is useful for checking the syntax of shell scripts. +.It Fl O Li interactive +If compiled with the overflow checks, turn them during arithmetic operations on. .It Fl P Li physical Change the default for the From owner-freebsd-performance@FreeBSD.ORG Sat Apr 19 20:58:22 2003 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 8532037B401; Sat, 19 Apr 2003 20:58:22 -0700 (PDT) Received: from adsl-63-198-35-122.dsl.snfc21.pacbell.net (adsl-63-198-35-122.dsl.snfc21.pacbell.net [63.198.35.122]) by mx1.FreeBSD.org (Postfix) with ESMTP id 9522B43FBD; Sat, 19 Apr 2003 20:58:21 -0700 (PDT) (envelope-from j_guojun@lbl.gov) Received: from lbl.gov (localhost.pacbell.net [127.0.0.1]) ESMTP id h3K301o1000382; Sat, 19 Apr 2003 20:00:02 -0700 (PDT) (envelope-from j_guojun@lbl.gov) Sender: jin@adsl-63-198-35-122.dsl.snfc21.pacbell.net Message-ID: <3EA20D31.2606389B@lbl.gov> Date: Sat, 19 Apr 2003 20:00:01 -0700 From: "Jin Guojun [NCS]" X-Mailer: Mozilla 4.76 [en] (X11; U; FreeBSD 4.8-RELEASE i386) X-Accept-Language: zh, zh-CN, en-US, en MIME-Version: 1.0 To: bj@dc.luth.se References: <200304191305.h3JD5S2F026929@dc.luth.se> Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit cc: freebsd-hackers@freebsd.org cc: freebsd-performance@freebsd.org Subject: Re: patch for test (Was: tcp_output starving -- is due to mbuf get delay?) X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 20 Apr 2003 03:58:23 -0000 Not yet. I said I will send out email when it is ready. The problem is we have a significantly modified system based on 4.8-RC2. I tried to extract the only sockbut/mbuf code, but it apparently, the patch is in completed for pure 4.8 system. Somehow, it have some depenency of our new TCP stack. So, I am building two new systems, one pure 4.7 and one pure 4.8. I will extract correct patch for these system, test them, then send out another email. Thanks for the patient. -Jin Borje Josefsson wrote: > Hmm. I'm not sure if I misunderstood if this was ready for another test > run or not. Anyhow - I took the new patch .tgz (which, btw, still had > tcp_input.p in it). I applied the patches (except tcp_input) and tested. > > Now I get: > > Panic: bad cur_off > 00000 m_p 0xc0a7f400 0xc0a7f400 my_off 0 1448 cc 3407144 > > As usual, I'm willing to test more when there are an update available. > > --Börje > > On Fri, 18 Apr 2003 13:04:24 PDT "Jin Guojun [DSD]" wrote: > > > Opps, there was a bad file -- tcp_input.p -- which is not working yet. > > Also, a patch file -- tcp_usrreq.p -- was missing. > > > > I will take the tcp_input.p out and put tcp_usrreq.p in. > > When it is finished, I will send another mail out. > > > > -Jin > > > > Borje Josefsson wrote: > > > > > On Thu, 17 Apr 2003 22:12:02 PDT "Jin Guojun [NCS]" wrote: > > > > > > > I have modified the sockbuf and mbuf operation to double the throughput over > > > > high bandwidth delay product path. > > > > > > > > The patch is available at: > > > > http://www-didc.lbl.gov/~jin/network/lion/content.html#FreeBSD_Patches > > > > > > > > The current modification is for tcp transmission only. > > > > > > > > I have adapted some code of uipc_socket2.c from Sam Leffler > > > > http://www.freebsd.org/~sam/thorpe-stable.patch > > > > > > > > for tcp receiver, but it has not been tested yet, so the tcp_input.p is empty. > > > > > > > > I ignored all record chain (m_nextpkt) related code. The details is explained at > > > > > > > > http://www-didc.lbl.gov/~jin/network/lion/content.html#BSDMbuf > > > > > > > > Once the tcp_input code is tested, I will submit the patch to bugs@freebsd.org. > > > > I may submit the patch regardless if tcp_input code works or not, because the > > > > tcp > > > > sender (server) is more important in high-speed network than the receiver > > > > (client). > > > > > > > > It is appreciated if any one can verify the patch and provide feedback. > > > > > > OK. I have now tried this patch on a newly-installed 4.8R. The patch > > > applied fine. When the sysctl net.inet.tcp.liondmask is unset, everything > > > seems OK, but when setting it to 7 (as specified with the patch > > > instructions) i get: > > > > > > Fatal trap 12: page fault while in kernel mode. > > > (I could write down all the stuff on addresses etc if it makes sense) > > > > > > when I run ttcp to test the performance. > > > > > > This is repeatable. > > > > > > I'm willing to test more, if someone provides me with some hints on what > > > to do. > > > > > > --Börje