From owner-freebsd-hackers  Sun Jun  4 19:38:58 1995
Return-Path: hackers-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.10/8.6.6) id TAA05364
          for hackers-outgoing; Sun, 4 Jun 1995 19:38:58 -0700
Received: from mail.id.net (kilroy.id.net [152.160.9.10])
          by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id TAA05356
          ; Sun, 4 Jun 1995 19:38:54 -0700
Received: from hades.id.net (hades.id.net [152.160.9.12]) by mail.id.net (8.6.8/8.6.6) with ESMTP id WAA02143; Sun, 4 Jun 1995 22:39:05 -0400
From: Robert Shady <rls@kilroy.id.net>
Received: (rls@localhost) by hades.id.net (8.6.11/8.6.6) id GAA07533; Mon, 5 Jun 1995 06:41:26 -0400
Message-Id: <199506051041.GAA07533@hades.id.net>
Subject: Re: too many sig 11's
To: dzerkel@feephi.phofarm.com (Danny J. Zerkel)
Date: Mon, 5 Jun 1995 06:41:26 -0400 (EDT)
Cc: jkh@freefall.cdrom.com, freebsd-hackers@freebsd.org
In-Reply-To: <199506050152.VAA00775@feephi.phofarm.com> from "Danny J. Zerkel" at Jun 4, 95 09:52:01 pm
X-Mailer: ELM [version 2.4 PL24]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 1616      
Sender: hackers-owner@freebsd.org
Precedence: bulk

> > > I've had a lot of problems with sig 11s again in 2.0.5-ALPHA...  Some
> > > things won't even execute (e.g., ftp, troff).  Others die during execution
> > > sometimes (during make of kernel).  At this point, I'm so frustrated I
> > > don't even know what to look for.  Any help?
> >
> > You may have a trashed kernel.  During the bin install, how many pieces
> > does it say it's getting?
> 
> I've rebuilt the kernel (typing make MANY times and rm any *.o that resulted
> from a sig 11'd cc).  And it still happens.  The bin install complained BIG
> time.  Many "skipping to next header" messages.  It looks like it is having
> a problem with untarring.  "cat bin.* | tar tzf -" dies at different places
> each time I try it.  Same for sgames.* and sgnu.*...  It's possible that I
> have something set wrong in the kernel I built, but it is impossible for me
> to tell (the LINT file is like having only the even numbered pages of a book).
> And I can't tell if the differences between LINT and GENERIC mean anything.
> 
> As an example of the present wierdness: /usr/bin/ftp dumps core immediately
> and /stand/ftp works fine.

See, and this is the same exact problem we are having.  We run a MUD game
on one of our 2.0.5 machines, and it SEG_FAULT's at LEAST once an hour, 
sometimes as often as every 5 to 10 minutes.  This was both with the
distributed kernel, and our rebuilt kernels.  The same exact game ran
just fine for DAYS under 1.1.5.1, no problems what-so-ever.  I think
there is some serious problem stuck somewhere deep in the 2.0 conversion.
More than likely it's memory related. ;)

	-- Rob