From owner-freebsd-current Sun Apr 14 09:25:27 1996 Return-Path: owner-current Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id JAA16161 for current-outgoing; Sun, 14 Apr 1996 09:25:27 -0700 (PDT) Received: from DeepCore.dk (aalb10.pip.dknet.dk [194.192.0.170]) by freefall.freebsd.org (8.7.3/8.7.3) with ESMTP id JAA16114 for ; Sun, 14 Apr 1996 09:25:21 -0700 (PDT) Received: (from sos@localhost) by DeepCore.dk (8.7.5/8.7.3) id JAA03788; Sun, 14 Apr 1996 09:57:39 +0200 (MET DST) Message-Id: <199604140757.JAA03788@DeepCore.dk> Subject: Re: File System Corruption *sigh* To: scrappy@ki.net (Marc G. Fournier) Date: Sun, 14 Apr 1996 09:57:39 +0200 (MET DST) From: "Søren Schmidt" Cc: current@FreeBSD.org In-Reply-To: from "Marc G. Fournier" at Apr 14, 96 02:24:40 am From: sos@FreeBSD.org Reply-to: sos@FreeBSD.org X-Mailer: ELM [version 2.4 PL25 ME8b] MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Sender: owner-current@FreeBSD.org X-Loop: FreeBSD.org Precedence: bulk In reply to Marc G. Fournier who wrote: > > > I'm not having much luck with this -current machine, hardware > or software yet to be determined... :( ----- > For warnings under /usr/src/sys/compile/freebsd. Figuring them > for mere warnings, I kind of ignored them, but then got curious, since I've > never seen that warning before, so checked out the file, and found this > at line 52: > > #define _BSD_SIZE_T_ unsigned int /* sizeof ) * > #d`fin` _BSD_SSIZE_T_ int /* byte count or error */ > #define _BSD_TIME_T_ long /* time() */ > > Notice, as well, the corruption at the end of the BSD_SIZE_T > define... > > I am using nfs mounts for various things, but my /usr, /usr/src > and /usr/include directories are all local, so I would presume that the > nfs mounts wouldn't affect this. > > Other then the memory being in question, this machine is a > 486DX2-66 w/ 16Meg of RAM, an Adaptec 1542CF SCSI controller and a > SCSI hard drive: > > (aha0:0:0): "UNISYS U0531 ST3600N 8374" type 0 fixed SCSI 2 > sd0(aha0:0:0): Direct-Access 500MB (1025920 512 byte sectors) I've seen this type of error MANY times over the last half year on many different hw setups, the only common factor is that they all uses SCSI disk, but I think that is coincidence. A surefire way for me to reproduce this is to checkout a complete src-tree while a kernel-build is running, that way I get 2-3 files corrupted like this in the whole tree :( Oh, and none of the systems I see failing uses nfs.... I have searched high and low for explanations to this, but havn't been able to come up with any hard evidence. My gut feeling though is that its VM related, some buffer mess up of sorts.... -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Søren Schmidt (sos@FreeBSD.org) FreeBSD Core Team Even more code to hack -- will it ever end ..