From owner-freebsd-current  Sun Jun  2 06:58:54 1996
Return-Path: owner-current
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id GAA20029
          for current-outgoing; Sun, 2 Jun 1996 06:58:54 -0700 (PDT)
Received: from ihgw2.att.com (ihgw2.att.com [207.19.48.2])
          by freefall.freebsd.org (8.7.5/8.7.3) with SMTP id GAA20024;
          Sun, 2 Jun 1996 06:58:52 -0700 (PDT)
From: dob@nasvr1.cb.att.com
Received: from nasvr1.cb.att.com by ihig2.att.att.com (SMI-8.6/EMS-1.2 sol2)
	id IAA15365; Sun, 2 Jun 1996 08:56:02 -0500
Received: by nasvr1.cb.att.com (5.x/EMS-1.1 Sol2)
	id AA08319; Sun, 2 Jun 1996 09:56:19 -0400
Cc: dyson@freebsd.org, ejc@nasvr1.cb.att.com, freebsd-current@freebsd.org,
        scrappy@ki.net
Received: from cbsky.cb.att.com by nasvr1.cb.att.com (5.x/EMS-1.1 Sol2)
	id AA08300; Sun, 2 Jun 1996 09:56:09 -0400
Received: by cbsky.cb.att.com (5.x/EMS-1.1 Sol2)
	id AA20401; Sun, 2 Jun 1996 09:55:41 -0400
Date: Sun, 2 Jun 1996 09:55:41 -0400
Message-Id: <9606021355.AA20401@cbsky.cb.att.com>
To: grog@lemis.de, toor@dyson.iquest.net
Original-Cc: dyson@freebsd.org, ejc@nasvr1, freebsd-current@freebsd.org, scrappy@ki.net
Subject: Re: Vm fixes NG
Sender: owner-current@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

> >>> Anyway, panic was ``panic: freeing held page, count = 4...''
> >>>
> >> 	I'm getting the same thing, *but* the description on today's
> >> re-commit of pmap.c sounds like it fixes that...going to try it now...
> >>
> > I *wish* so, but it might not.  Just let me know!!!
> 
> I've just built a kernel with ctm 2065.  emacs still takes several
> attempts to start, but once it's running, there are no more problems
> (this is the way it's been all along--sorry I didn't mention that
> before).  Today the crashes were all clean SIGSEGVs.

John, Et.Al,

Make world took all night with three or so hand fixes for y.tab.h errors, but 
when I saw my monitor this morning, the ``freeing held page'' panic (count=12)
was on the console.  The recent pmap, vm_fault fixes much improved the problem
on my machine (8MB, 0KB L2), but there still seems to be a hole somewhere.  

Thanks,

Dan O'Brien