From owner-freebsd-hackers Fri Jun 13 11:05:45 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.5/8.8.5) id LAA10860 for hackers-outgoing; Fri, 13 Jun 1997 11:05:45 -0700 (PDT) Received: from helbig.informatik.ba-stuttgart.de (helbig.informatik.ba-stuttgart.de [141.31.166.22]) by hub.freebsd.org (8.8.5/8.8.5) with ESMTP id LAA10852 for ; Fri, 13 Jun 1997 11:05:39 -0700 (PDT) Received: (from helbig@localhost) by helbig.informatik.ba-stuttgart.de (8.8.5/8.8.5) id UAA10160; Fri, 13 Jun 1997 20:03:48 +0200 (MET DST) From: Wolfgang Helbig Message-Id: <199706131803.UAA10160@helbig.informatik.ba-stuttgart.de> Subject: Re: Strange things afoot In-Reply-To: from Stephen Roome at "Jun 13, 97 01:28:55 pm" To: steve@visint.co.uk (Stephen Roome) Date: Fri, 13 Jun 1997 20:03:47 +0200 (MET DST) Cc: freebsd-hackers@FreeBSD.ORG X-Mailer: ELM [version 2.4ME+ PL30 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-hackers@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk Hi, > I'm installing bisdn-0.97 on FreeBSD 2.2.2, I've used the mkbisdntree > script and patches which are supposedly for current. > Anyway, this isn't the bit that's out of the ordinary (much). > > mkbisdntree makes a huge symbolic link tree to the normal /sys but with > some alterations. So I go into /bisdn/i386/conf and edit a configuration > file, run config on it and it makes a compile directory. > > So I then go into that directory and do a make depend. After I've then > installed all the extra bits of code it needs which mkbisdntree didn't put Which `extra bits of code' are you refering to? The only files that are needed for compilation and are not installed by mkbisdntree are /usr/sys.bisdn/i386/include/bisdn_ioctl.h and /usr/sys.bisdn/i386/include/bisdn_isdn.h Instead these files are copied to /usr/include/machine which in turn is a symbolic link to /sys/i386/include once you did a ``make world'' on your machine. So these files are `missing' only on a system that never was used for ``make world''. This is interesting, because ``make world'' is considered to be a very good hardware tester, and your problems might be hardware caused. So if you have about 250 Mbyte left on your /usr partition give it a try and see what happens. > in place it starts giving out > > (cpp) died with signal 6. > > After a few of these errors (about 20 in a row) make then dies with signal > 11. > > I get this a lot, this is because ffs_extern.h somehow seems to get > completely corrupted in the patching process, although I can't find > anything in the patches which actually touches this file. > > After fixing ffs_extern.h back to what it should be it compiles again, but > then halfway through I invariably end up with make exiting on signal 11. > > At this point I usually find it's best to reboot, as everything starts > dieing with segmentation faults or bus errors. (Last time I decided not > reboot I ended up with ls/cd not even working.) > > I've been using /usr mounted async for this, but I'm not sure that's the > problem. The problems only start occuring once cpp has crashed really > badly on ffs_extern.h, until that point everything is fine. > > Once I've rebooted, it seems to be okay again, (/usr is not mounted async > by default so it's back to normal there) and I'm now able to finish making > the kernel. > > So, my question is, does anyone know what might be causing this, I don't > think this is likely to be a problem with my hardware as everything works > fine up to this point and I ran this box with various 3.0 snapshots and X > and played quake on it for a few hours a day for about a month without > problems (well, it was my desktop machine for quite some time, and I'm > pretty sure I used every bit of hardware quite thoroughly.) > > I'm thinking this is perhaps being caused by async mounting and excessive > use of soft links, as that's the only thing that's actually different to > the normal everyday use that this box has been subjected too. > > Steve Roome > Vision Interactive Ltd. > > Wolfgang