From owner-freebsd-hackers Fri Dec 23 20:44:14 1994 Return-Path: hackers-owner Received: (from root@localhost) by freefall.cdrom.com (8.6.9/8.6.6) id UAA19078 for hackers-outgoing; Fri, 23 Dec 1994 20:44:14 -0800 Received: from frig.mt.cs.keio.ac.jp (frig.mt.cs.keio.ac.jp [131.113.32.7]) by freefall.cdrom.com (8.6.9/8.6.6) with ESMTP id EAA19072 for ; Sat, 24 Dec 1994 04:44:10 GMT Received: (from hosokawa@localhost) by frig.mt.cs.keio.ac.jp (8.6.9+2.4Wb/3.3Wb) id NAA07797; Sat, 24 Dec 1994 13:43:44 +0900 Date: Sat, 24 Dec 1994 13:43:44 +0900 Message-Id: <199412240443.NAA07797@frig.mt.cs.keio.ac.jp> To: davidg@Root.COM Cc: freebsd-hackers@freebsd.org Subject: Re: DX4/100 BUG In-Reply-To: Your message of Fri, 23 Dec 1994 20:01:22 -0800. <199412240401.UAA04733@corbin.Root.COM> From: hosokawa@mt.cs.keio.ac.jp (HOSOKAWA Tatsumi) X-Mailer: mnews [version 1.18PL3+] 1994-08/01(Mon) Sender: hackers-owner@freebsd.org Precedence: bulk In article <199412240401.UAA04733@corbin.Root.COM> davidg@Root.COM writes: >> What bug? I have a DX4/100 sitting next to me that works just fine with 2.0. I heard that not every DX4 has this bug. Some DX4 has a bug on SMM and hlt instruction. I found news articles about it at comp.sys.intel and comp.sys.ibm.pc.hardware.chips, but most of them have expired on our site. hosokawa (from comp.sys.ibm.pc.hardware.chips) -(BEGIN)-------------------------------------------------------------------- From: alhaz@xmission.com (Alfred Hazelworth) Newsgroups: comp.sys.intel,comp.sys.ibm.pc.hardware.chips,comp.os.os2.misc Subject: Re: **>>DX4/100 RECALLED<<** Date: 10 Dec 1994 19:36:25 -0700 Organization: XMission Public Access Internet (801 539 0900) Lines: 18 Message-ID: <3cdoj9$g1k@xmission.xmission.com> References: <1994Nov30.154328.25469@mlb.semi.harris.com> <3bkrdo$ej0@inewssc.intel.com> <1994Dec5.204344.15135@mlb.semi.harris.com> NNTP-Posting-Host: xmission Keywords: DX4/100 Recall CHIPLIST X-Newsreader: NN version 6.5.0 #2 (NOV) Xref: kogwy comp.sys.intel:25522 comp.sys.ibm.pc.hardware.chips:15441 comp.os.os2.misc:124116 s106275@cc.tut.fi (Anssi Saari) writes: >In <1994Dec5.204344.15135@mlb.semi.harris.com> bj@krusty.mlb.semi.harris.com (Bobby J. Fralic) writes: >:> What about the problem, detailed in the Chiplist quote below, is it for >:>real? Does it still exist on some/all DX4/100's? Does it mean there will >:>be a problem running WindowsNT,Windows95,OS2, etc. >Well, the current version of the Chiplist, 6.0, makes no mention of a bug in >the DX4. This might be in refrence to the problem OS/2 2.1 had with some pentiums and 486s. IT had something to do with the way Intel SMM works. Since the DX4 allways had SMM, the bug would crop up from time to time. The patch is refered to as the "pentium fix" because that's what people found it most usefull for or something. AlHaz Date: 11 Dec 1994 00:00:00 +0000 From: holger_f@holgi.escape.de (Holger Foerstemann) Newsgroups: comp.os.os2.misc,comp.sys.ibm.pc.hardware.chips,comp.sys.intel Message-ID: <5bd2TvA42aB@holgi.escape.de> References: Subject: Re: **>>DX4/100 RECALLED<<** X-Newsreader: XP v3.02 R/C1307 Reply-To: holger_f@xmp.escape.de Lines: 23 Xref: kogwy comp.os.os2.misc:124515 comp.sys.ibm.pc.hardware.chips:15572 comp.sys.intel:26142 mike wrotes in "Re: **>>DX4/100 RECALLED<<**": > > :> What about the problem, detailed in the Chiplist quote below, is it for > > :>real? Does it still exist on some/all DX4/100's? Does it mean there > > :>will be a problem running WindowsNT,Windows95,OS2, etc. > > > > Well, the current version of the Chiplist, 6.0, makes no mention of a bug > > in the DX4. > I have a DX4 and run NT 3.5 and have not had any problems yet. What should I > look for? Now I have my the fourth DX4/100 on my Board. The first three Proz have had big problems with OS/2. When I run Windows, the Proz works with no errors. But OS/2 is a 32-Bit-System! But at a speed of 75Mhz all DX4 work fine!!!! Holgi --------------------------------------------------------------------------- FIDO: Holger Foerstemann@2:241/530.12 & Holger Foerstemann@2:2437/430.1 ZConnect: HOLGI@EARTHBOX.ESCAPE.DE / UUCP: holger_f@xmp.escape.de --------------------------------------------------------------------------- ## CrossPoint v3.02 R ## From: bj@krusty.mlb.semi.harris.com (Bobby J. Fralic) Newsgroups: comp.sys.intel,comp.sys.ibm.pc.hardware.chips,comp.os.os2.misc Subject: Re: **>>DX4/100 RECALLED?<<** Date: 14 Dec 1994 16:22:04 GMT Organization: Harris Semiconductor, Melbourne, Florida Lines: 44 Message-ID: <3cn63c$13r@hearye.mlb.semi.harris.com> References: NNTP-Posting-Host: krusty.mlb.semi.harris.com Xref: kogwy comp.sys.intel:26835 comp.sys.ibm.pc.hardware.chips:15703 comp.os.os2.misc:124813 In article , >Mike Whitacre wrote: >>In article , writes: >> :> bj@krusty.mlb.semi.harris.com posted for Lou: >> :> What about the problem, detailed in the Chiplist quote below, is it for >> :>real? Does it still exist on some/all DX4/100's? Does it mean there will >> :>be a problem running WindowsNT,Windows95,OS2, etc. >> Well, the current version of the Chiplist, 6.0, makes no mention of a bug in >> the DX4. >I have a DX4 and run NT 3.5 and have not had any problems yet. What should I >look for? Mike Here is a summary of some of the replies I saved. From: ig25@fg70.rz.uni-karlsruhe.de (Thomas Koenig) Linux uses the HLT instruction for its idle loop since were 1.1.18 or so. (People were really astonished when the temperature dropped by 40 Kelvin or so :-) However, some chips were reported as hanging, and Linux now checks this as part of its bootup sequence (as it does for the FDIV bug since 1.1.66). There were a few reports of machines hanging as a result of this; my guess would be this resulted in the recall. -- Thomas Koenig, Thomas.Koenig@ciw.uni-karlsruhe.de, ig25@dkauni2.bitnet. ********************************************************************* From: alhaz@xmission.com (Alfred Hazelworth) Subject: Re: **>>DX4/100 RECALLED<<** This might be in refrence to the problem OS/2 2.1 had with some pentiums and 486s. IT had something to do with the way Intel SMM works. Since the DX4 allways had SMM, the bug would crop up from time to time. The patch is refered to as the "pentium fix" because that's what people found it most usefull for or something. AlHaz ********************************************************************* Where is this patch available? Lou -(END)---------------------------------------------------------------------- --- ----------------------------------------------------------- HOSOKAWA, Tatsumi (hosokawa@mt.cs.keio.ac.jp) Dept. of Computer Science, Keio University, Yokohama, Japan -----------------------------------------------------------