From owner-freebsd-smp Sun Sep 28 07:15:00 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id HAA03134 for smp-outgoing; Sun, 28 Sep 1997 07:15:00 -0700 (PDT) Received: from sinbin.demos.su (sinbin.demos.su [194.87.0.31]) by hub.freebsd.org (8.8.7/8.8.7) with SMTP id HAA02866 for ; Sun, 28 Sep 1997 07:14:04 -0700 (PDT) Received: by sinbin.demos.su id SAA23424; (8.6.12/D) Sun, 28 Sep 1997 18:13:25 +0400 From: bag@sinbin.demos.su (Alex G. Bulushev) Message-Id: <199709281413.SAA23424@sinbin.demos.su> Subject: Re: intel dual PII DK440LX In-Reply-To: <19970927123904.39335@klemm.gtn.com> from "Andreas Klemm" at "Sep 27, 97 12:39:04 pm" X-ELM-OSV: (Our standard violations) no-mime=1; no-hdr-encoding=1 To: andreas@klemm.gtn.com (Andreas Klemm) Date: Sun, 28 Sep 1997 18:13:25 +0400 (MSD) Cc: smp@freebsd.org X-Mailer: ELM [version 2.4ME+ PL32 (25)] Content-Type: text Sender: owner-freebsd-smp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk > On Fri, Sep 26, 1997 at 11:59:36AM +0400, Alex G. Bulushev wrote: > > any information about -carrent on new intel dual PII 440LX motherboard > > and adaptec 7895 ... ? > > All I have to say ... WoW ! If this board should work fine, then > it will certainly be a good choice, because of the huge feature > list. > > You don't need an Ethernet card, don't need a SCSI controller, > sound card and so on ... currently we use ASUS and intel PR440FX on several servers with dual PPro 200, new intel mb very intresting for dual PII/300MHz, but: Q1: fbsd/SMP 440LX support Q2: fbsd adaptec 7895 support Alex. > > Price ? > > -- > Andreas Klemm | klemm.gtn.com - powered by > Symmetric MultiProcessor FreeBSD > http://www.freebsd.org/~fsmp/SMP/SMP.html > http://www.freebsd.org/~fsmp/SMP/benches.html > From owner-freebsd-smp Sun Sep 28 13:09:42 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id NAA20287 for smp-outgoing; Sun, 28 Sep 1997 13:09:42 -0700 (PDT) Received: from Ilsa.StevesCafe.com (Ilsa.StevesCafe.com [205.168.119.129]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id NAA20279 for ; Sun, 28 Sep 1997 13:09:38 -0700 (PDT) Received: from Ilsa.StevesCafe.com (localhost [127.0.0.1]) by Ilsa.StevesCafe.com (8.8.7/8.8.5) with ESMTP id OAA06017 for ; Sun, 28 Sep 1997 14:09:34 -0600 (MDT) Message-Id: <199709282009.OAA06017@Ilsa.StevesCafe.com> X-Mailer: exmh version 2.0gamma 1/27/96 From: Steve Passe To: smp@freebsd.org Subject: deadlock in PUSHDOWN_LEVEL_3 Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Sun, 28 Sep 1997 14:09:33 -0600 Sender: owner-freebsd-smp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk Hi, I have been fighting a deadlock in the "giant lock" PUSHDOWN_LEVEL_3 code. I would like help from any adverturous souls out there. I just committed new versions of ipl.s and ipl_funcs.c that have breakpoint()s at the appropriate places. To activate the code edit smptests.h, changing from: #define PUSHDOWN_LEVEL_3_NOT to: #define PUSHDOWN_LEVEL_3 then rebuild and install the test SMP kernel. It should deadlock anywhere from several minutes to several hours later. Start alot of processes of some sort to hasten the lockup. The DDB screen will look like: ------------------------------- cut --------------------------------- cil: 0x00000100Stopped at _breakpoint+0x1: ret db> ------------------------------- cut --------------------------------- a trace shows: ------------------------------- cut --------------------------------- db> trace _breakpoint(f4df0f74,5f5e101,ffffffff,80000000,f0112d21) at _breakpoint+0x1 _splhigh(80000000,f0220010,f0930010,0,0) at _splhigh+0xc4 doreti_swi() at doreti_swi+0x2a ------------------------------- cut --------------------------------- It always occurs with doreti_swi() at the bottom of the stack. The value of cil is USUALLY 0x00000100, but I once saw another value (disk INT) 'cil' is the 'current interrupt level'. It denotes the interrupt currently active. The value of 0100 specifies the rtc clock interrupt. The fact that cil is non-zero claims that we are currently inside an ISR, but this should be impossible when at point doreti_swi. 'inside_intr' is also always 0 when this occurs. I have found that you can clear cil (db> write cil 0) tnen continue without problem. This tends to confirm that we really are no longer inside the ISR specific code, and that somehow we are leaving cil set when we shouldn't. Another possibility is that the code in ipl.s, doreti_unpend sets cil in preparation for the call to XresumeNN, but somehow doesn't go down that path. Note that the giant lock still protects all the interrupt code for this test. It isn't removed till LEVEL_4 is enabled(don't bother, NOT ready yet) Hopefully a fresh set of eyes can find this! -- Steve Passe | powered by smp@csn.net | Symmetric MultiProcessor FreeBSD From owner-freebsd-smp Mon Sep 29 14:08:04 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id OAA10003 for smp-outgoing; Mon, 29 Sep 1997 14:08:04 -0700 (PDT) Received: from csla.csl.sri.com (csla.csl.sri.com [192.12.33.2]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id OAA09970 for ; Mon, 29 Sep 1997 14:07:52 -0700 (PDT) Received: from japonica.csl.sri.com (japonica.csl.sri.com [130.107.15.17]) by csla.csl.sri.com (8.8.6/8.8.5) with ESMTP id OAA28285 for ; Mon, 29 Sep 1997 14:03:31 -0700 (PDT) Received: from japonica.csl.sri.com (localhost [127.0.0.1]) by japonica.csl.sri.com (8.8.7/8.7.3) with ESMTP id OAA01141 for ; Mon, 29 Sep 1997 14:07:12 -0700 (PDT) Message-Id: <199709292107.OAA01141@japonica.csl.sri.com> To: smp@freebsd.org Subject: SMP-related automounter problem Date: Mon, 29 Sep 1997 14:07:12 -0700 From: Fred Gilham Sender: owner-freebsd-smp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk I'm having a strange problem that seems to come up with any snapshots after the end of August. It seems that automounting takes two tries. The first time I'll get a stale NFS handle; the second time everything is fine. It isn't a big problem. It's just annoying. For example, when reading mail in emacs I have to try at least two times to reply to some mail; once to remount the mh directory and once to remount my home directory. I have another non-smp machine that doesn't exhibit this problem. I can use the 970812 snapshot with no problem. I've used later August snapshots as well but I don't have them available any more. My system is a dual-PPro 200 with a super-micro P6DNF board. (I was experiencing the same problem with an older super-micro dual-PPro board which I upgraded for other reasons). -Fred Gilham gilham@csl.sri.com From owner-freebsd-smp Wed Oct 1 20:01:36 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id UAA26713 for smp-outgoing; Wed, 1 Oct 1997 20:01:36 -0700 (PDT) Received: from csvax.cs.caltech.edu (csvax.cs.caltech.edu [131.215.131.131]) by hub.freebsd.org (8.8.7/8.8.7) with SMTP id UAA26707 for ; Wed, 1 Oct 1997 20:01:31 -0700 (PDT) Received: from stun4p.cs.caltech.edu by csvax.cs.caltech.edu (4.1/1.34.1) id AA25497; Wed, 1 Oct 97 20:01:29 PDT Received: from localhost.cs.caltech.edu (localhost.cs.caltech.edu [127.0.0.1]) by stun4p.cs.caltech.edu (8.7.4/8.7.3) with SMTP id UAA22646; Wed, 1 Oct 1997 20:01:31 -0700 (PDT) Message-Id: <199710020301.UAA22646@stun4p.cs.caltech.edu> X-Authentication-Warning: stun4p.cs.caltech.edu: Host localhost.cs.caltech.edu [127.0.0.1] didn't use HELO protocol To: Fred Gilham Cc: smp@freebsd.org Subject: Re: SMP-related automounter problem In-Reply-To: Your message of "Mon, 29 Sep 1997 14:07:12 PDT." <199709292107.OAA01141@japonica.csl.sri.com> Date: Wed, 01 Oct 1997 20:01:30 -0700 From: Mika Nystrom Sender: owner-freebsd-smp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk I have been having the same problem---it is very annoying since almost all our filesystems are mounted via AMD. However, for us, the problem did not go away when we shut off one cpu (perhaps it got a little better, but it's still there). We are using dual and quad NeTpower/intel ppros. Mika Fred Gilham writes: > >I'm having a strange problem that seems to come up with any snapshots >after the end of August. It seems that automounting takes two tries. >The first time I'll get a stale NFS handle; the second time everything >is fine. > >It isn't a big problem. It's just annoying. For example, when >reading mail in emacs I have to try at least two times to reply to >some mail; once to remount the mh directory and once to remount my >home directory. > >I have another non-smp machine that doesn't exhibit this problem. > >I can use the 970812 snapshot with no problem. I've used later >August snapshots as well but I don't have them available any more. > >My system is a dual-PPro 200 with a super-micro P6DNF board. (I was >experiencing the same problem with an older super-micro dual-PPro >board which I upgraded for other reasons). > >-Fred Gilham gilham@csl.sri.com From owner-freebsd-smp Thu Oct 2 11:52:03 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id LAA15133 for smp-outgoing; Thu, 2 Oct 1997 11:52:03 -0700 (PDT) Received: from gratis.grondar.za (gratis.grondar.za [196.7.18.133]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id LAA15071; Thu, 2 Oct 1997 11:51:17 -0700 (PDT) Received: from greenpeace.grondar.za (7h04ybTPcaEk5KGfAxPP35M1C8x3An/w@greenpeace.grondar.za [196.7.18.132]) by gratis.grondar.za (8.8.7/8.8.7) with ESMTP id UAA24623; Thu, 2 Oct 1997 20:51:08 +0200 (SAT) Received: from greenpeace.grondar.za (2yLit/m0Bj4PFdVML6YwU/RoVb9i4qpN@localhost [127.0.0.1]) by greenpeace.grondar.za (8.8.7/8.8.7) with ESMTP id UAA18569; Thu, 2 Oct 1997 20:51:15 +0200 (SAT) Message-Id: <199710021851.UAA18569@greenpeace.grondar.za> To: phk@freebsd.org cc: smp@freebsd.org Subject: Making world on an SMP system. Date: Thu, 02 Oct 1997 20:51:14 +0200 From: Mark Murray Sender: owner-freebsd-smp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk Hi I am a recent convert to the world of SMP. Much of what I do on this board (a Giga-Byte GA586DX w 2xP5/200) is a _helluva_ lot easier, now that the load average is a bit more reasonable. (I do real number- crunching here :-) ). One thing that I have found to be frought with danger is large makes and "make -j N" (N an int). EG: make -j2 world usually falls over in a couple of minutes. It seems 1) either the dependancies are not SMP-friendly, or 2) make is broken. I suspect 1). Is this a known "thing"? Is anyone working on it? ISTR that PHK was into massively parallel makes a few years ago (?). Are you (PHK) still doing stuff there? M -- Mark Murray Join the anti-SPAM movement: http://www.cauce.org From owner-freebsd-smp Thu Oct 2 12:00:23 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id MAA15511 for smp-outgoing; Thu, 2 Oct 1997 12:00:23 -0700 (PDT) Received: from critter.freebsd.dk (critter.freebsd.dk [195.8.129.26]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id MAA15499 for ; Thu, 2 Oct 1997 12:00:15 -0700 (PDT) Received: from critter.freebsd.dk (localhost.cybercity.dk [127.0.0.1]) by critter.freebsd.dk (8.8.7/8.8.7) with ESMTP id UAA05090; Thu, 2 Oct 1997 20:55:56 +0200 (CEST) To: Mark Murray cc: smp@freebsd.org Subject: Re: Making world on an SMP system. In-reply-to: Your message of "Thu, 02 Oct 1997 20:51:14 +0200." <199710021851.UAA18569@greenpeace.grondar.za> Date: Thu, 02 Oct 1997 20:55:56 +0200 Message-ID: <5088.875818556@critter.freebsd.dk> From: Poul-Henning Kamp Sender: owner-freebsd-smp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk In message <199710021851.UAA18569@greenpeace.grondar.za>, Mark Murray writes: >1) either the dependancies are not SMP-friendly, or indeed. >I suspect 1). Is this a known "thing"? Is anyone working on it? ISTR >that PHK was into massively parallel makes a few years ago (?). > >Are you (PHK) still doing stuff there? No, at that time I "cheated". I frobbed the makefiles to generate a list of commands to execute, inserted some synchronization markers by hand, and used a traditional "farming" setup for execution. -- Poul-Henning Kamp FreeBSD coreteam member phk@FreeBSD.ORG "Real hackers run -current on their laptop." From owner-freebsd-smp Thu Oct 2 12:14:04 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id MAA16278 for smp-outgoing; Thu, 2 Oct 1997 12:14:04 -0700 (PDT) Received: from gratis.grondar.za (gratis.grondar.za [196.7.18.133]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id MAA16258 for ; Thu, 2 Oct 1997 12:13:58 -0700 (PDT) Received: from greenpeace.grondar.za (QrWslIahhdLBQ6ZgbvTRkK4z0rG+TCNa@greenpeace.grondar.za [196.7.18.132]) by gratis.grondar.za (8.8.7/8.8.7) with ESMTP id VAA24669; Thu, 2 Oct 1997 21:14:18 +0200 (SAT) Received: from greenpeace.grondar.za (5veK3gw7oJ3gjufxSsvhUaiajB5E77jd@localhost [127.0.0.1]) by greenpeace.grondar.za (8.8.7/8.8.7) with ESMTP id VAA07271; Thu, 2 Oct 1997 21:14:24 +0200 (SAT) Message-Id: <199710021914.VAA07271@greenpeace.grondar.za> To: Poul-Henning Kamp cc: smp@freebsd.org Subject: Re: Making world on an SMP system. Date: Thu, 02 Oct 1997 21:14:24 +0200 From: Mark Murray Sender: owner-freebsd-smp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk Poul-Henning Kamp wrote: > >1) either the dependancies are not SMP-friendly, or > > indeed. Is there any literature on the subject? My interest is now piqued. M -- Mark Murray Join the anti-SPAM movement: http://www.cauce.org From owner-freebsd-smp Sat Oct 4 15:55:44 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id PAA11641 for smp-outgoing; Sat, 4 Oct 1997 15:55:44 -0700 (PDT) Received: from dcarmich.pr.mcs.net (dcarmich.pr.mcs.net [204.95.63.202]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id PAA11630; Sat, 4 Oct 1997 15:55:40 -0700 (PDT) Received: (from dcarmich@localhost) by dcarmich.pr.mcs.net (8.8.5/8.8.5) id RAA00924; Sat, 4 Oct 1997 17:58:56 -0500 (CDT) From: Douglas Carmichael Message-Id: <199710042258.RAA00924@dcarmich.pr.mcs.net> Subject: Building parallel "Beowulf-style" supercomputers with FreeBSD To: freebsd-hackers@freebsd.org Date: Sat, 4 Oct 1997 17:58:56 -0500 (CDT) Cc: freebsd-smp@freebsd.org X-Mailer: ELM [version 2.4ME+ PL31H (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk Is there any way one can build large "Beowulf-like" clusters with FreeBSD and its SMP functionality? Anyone built such a beast? From owner-freebsd-smp Sat Oct 4 17:55:31 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id RAA16857 for smp-outgoing; Sat, 4 Oct 1997 17:55:31 -0700 (PDT) Received: from news.quick.net (donegan@news.quick.net [207.212.170.1]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id RAA16837; Sat, 4 Oct 1997 17:55:26 -0700 (PDT) Received: (from donegan@localhost) by news.quick.net (8.8.5/8.6.9) id RAA09114; Sat, 4 Oct 1997 17:55:19 -0700 (PDT) Date: Sat, 4 Oct 1997 17:55:18 -0700 (PDT) From: "Steven P. Donegan" To: Douglas Carmichael cc: freebsd-hackers@freebsd.org, freebsd-smp@freebsd.org Subject: Re: Building parallel "Beowulf-style" supercomputers with FreeBSD In-Reply-To: <199710042258.RAA00924@dcarmich.pr.mcs.net> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk On Sat, 4 Oct 1997, Douglas Carmichael wrote: > Is there any way one can build large "Beowulf-like" clusters with FreeBSD and its SMP functionality? > Anyone built such a beast? > > I too am very interested in SMP NOW toy's From owner-freebsd-smp Sat Oct 4 18:16:25 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id SAA17538 for smp-outgoing; Sat, 4 Oct 1997 18:16:25 -0700 (PDT) Received: from consys.com (consys.com [209.60.202.194]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id SAA17515; Sat, 4 Oct 1997 18:16:14 -0700 (PDT) Received: from cssunix.conceptual.com (cssunix.conceptual.com [10.0.2.5]) by consys.com (8.8.6/8.8.6) with ESMTP id SAA00917; Sat, 4 Oct 1997 18:16:13 -0700 (MST) Received: (from rcarter@localhost) by cssunix.conceptual.com (8.8.5/8.8.6) id SAA29007; Sat, 4 Oct 1997 18:16:13 -0700 (MST) Date: Sat, 4 Oct 1997 18:16:13 -0700 (MST) From: "Russell L. Carter" Message-Id: <199710050116.SAA29007@cssunix.conceptual.com> To: dcarmich@mcs.com, freebsd-hackers@freebsd.org Subject: Re: Building parallel "Beowulf-style" supercomputers with FreeBSD Cc: freebsd-smp@freebsd.org Sender: owner-freebsd-smp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk Yes. Now it is a curious thing that NFS server perf. is not considered, but such is life. In fact, there are lots of interesting questions to be asked, but, who am I to ask them? :-) Cheers, Russell