From owner-freebsd-current Fri Aug 23 7:40:33 2002 Delivered-To: freebsd-current@freebsd.org Received: from mx1.FreeBSD.org (mx1.FreeBSD.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id CB4D737B401 for ; Fri, 23 Aug 2002 07:40:26 -0700 (PDT) Received: from tesla.distributel.net (nat.MTL.distributel.NET [66.38.181.24]) by mx1.FreeBSD.org (Postfix) with ESMTP id EFBBD43E8A for ; Fri, 23 Aug 2002 07:40:25 -0700 (PDT) (envelope-from bmilekic@unixdaemons.com) Received: (from bmilekic@localhost) by tesla.distributel.net (8.11.6/8.11.6) id g7NEcGe38522; Fri, 23 Aug 2002 10:38:16 -0400 (EDT) (envelope-from bmilekic@unixdaemons.com) Date: Fri, 23 Aug 2002 10:38:16 -0400 From: Bosko Milekic To: Terry Lambert Cc: Martin Blapp , Julian Elischer , Don Lewis , sos@freebsd.dk, marks@ripe.net, ktsin@acm.org, freebsd-current@FreeBSD.ORG Subject: Re: Memory corruption in CURRENT Message-ID: <20020823103816.A38472@unixdaemons.com> References: <20020823155725.T50084-100000@levais.imp.ch> <3D6645A4.24593DA1@mindspring.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5.1i In-Reply-To: <3D6645A4.24593DA1@mindspring.com>; from tlambert2@mindspring.com on Fri, Aug 23, 2002 at 07:24:36AM -0700 Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG On Fri, Aug 23, 2002 at 07:24:36AM -0700, Terry Lambert wrote: > Martin Blapp wrote: > > > Julian just wants a full set of tests. > > > > Ok, I'll build another row. > > > > > Julian: DISABLE_PG_G alone *might* fix a particular symptom on > > > a particular persons' machine, but that would not be definitive, > > > only anecdotal. > > > > Do you think this is a software or a hardware bug ? Are PIV's buggy ;-) ? > > I believe it's a general Pentium-class and above hardware bug, > which is currently most easily repeatable on Pentium 4's and AMD > Athlons. > > In the limit, though, hardware people will tell you that failure > to workaround hardware bugs *is* a software bug (hardware guys do > not understand "recursion" or "completeness" very well, I think > 8-) 8-)). > > FWIW: I think your machine will crash just the same without the > DISABLE_PSE but with the DISABLE_PG_G. FWIW, there are several problems going on that may be the cause of the fairly random sig4 and 11s you've been seeing. 1) race conditions in the code peter has mentionned 2) undocumented hardware bug #1 (the one peter mentionned and I described) regarding, seemingly, a 4M PDE mapping for the first 4M of memory. 3) undocumented hardware bugs #2 and #3 that Terry mentions. > -- Terry -- Bosko Milekic * bmilekic@unixdaemons.com * bmilekic@FreeBSD.org To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message