From owner-freebsd-hackers  Sun Feb  9 11:21:34 1997
Return-Path: <owner-hackers>
Received: (from root@localhost)
          by freefall.freebsd.org (8.8.5/8.8.5) id LAA01676
          for hackers-outgoing; Sun, 9 Feb 1997 11:21:34 -0800 (PST)
Received: from dg-rtp.dg.com (dg-rtp.rtp.dg.com [128.222.1.2])
          by freefall.freebsd.org (8.8.5/8.8.5) with SMTP id LAA01656
          for <freebsd-hackers@freefall.FreeBSD.org>; Sun, 9 Feb 1997 11:21:24 -0800 (PST)
Received: by dg-rtp.dg.com (5.4R3.10/dg-rtp-v02)
	id AA09105; Sun, 9 Feb 1997 14:20:52 -0500
Received: from ponds by dg-rtp.dg.com.rtp.dg.com; Sun,  9 Feb 1997 14:20 EST
Received: from lakes.water.net (lakes [10.0.0.3]) by ponds.water.net (8.8.3/8.7.3) with ESMTP id NAA02648; Sun, 9 Feb 1997 13:49:10 -0500 (EST)
Received: (from root@localhost) by lakes.water.net (8.8.3/8.6.9) id NAA24339; Sun, 9 Feb 1997 13:53:34 -0500 (EST)
Date: Sun, 9 Feb 1997 13:53:34 -0500 (EST)
From: Thomas David Rivers <ponds!rivers@dg-rtp.dg.com>
Message-Id: <199702091853.NAA24339@lakes.water.net>
To: ponds!root.com!dg, ponds!freefall.cdrom.com!freebsd-hackers,
        ponds!uriah.heep.sax.de!joerg_wunsch, ponds!lambert.org!terry
Subject: daily panics, ffs_valloc: dup alloc - Good news!
Content-Type: text
Sender: owner-hackers@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk


As I said before; I would test our Joerg's supposition.  I'm happy to
report it seems to be right on the money! (Good work!)

I built a new "newfs", with NTRACKS bumped to 2 and NSECTORS dropped 
to 2048.

It worked like a champ!  No more panic: ffs_valloc: dup alloc.

I'd say the problem is that the underlying code can't handle one
track (head).

We should probably go ahead and use this work-around in 2.1.7 and
2.2.  Perhaps, if we're so inclined, we can determine what a
better fix would be to keep the 1 track idea.  [It could possibly
be simply setting fs->fs_cgmask to 0 if the number of tracks is 1; but 
I'm not sure.]  That investigation could wait until after 2.2 and
2.1.7.

Thanks to everyone for sticking with me on this!

[Now, the question becomes how to adjust an existing file system; which
I don't think can be done :-) ]

Here's my (simple) patch to newfs.c...

	- Dave Rivers -

*** newfs.c.ori	Sun Sep 17 05:56:20 1995
--- newfs.c	Sun Feb  9 13:51:46 1997
***************
*** 150,160 ****
   * We do this by saying one head, with a lot of sectors on it.
   * The number of sectors are used to determine the size of a cyl-group.
   * Kirk suggested one or two meg per "cylinder" so we say two.
   */
  
! #define NTRACKS		1	/* number of heads */
  
! #define NSECTORS	4096	/* number of sectors */
  
  int	mfs;			/* run as the memory based filesystem */
  int	Nflag;			/* run without writing file system */
--- 150,166 ----
   * We do this by saying one head, with a lot of sectors on it.
   * The number of sectors are used to determine the size of a cyl-group.
   * Kirk suggested one or two meg per "cylinder" so we say two.
+  *
+  *  NB: Although it's tempting to make NTRACKS 1; the underlying file
+  *      system code will then receive an fs_cgmask of 0xffffffff which
+  *      will make for ino<->block calculations; causing "dup alloc" and
+  *      other panics.  Until that problem is addressed, we pretend to have
+  *      two heads (two heads are better than one :-) .)
   */
  
! #define NTRACKS		2	/* number of heads */
  
! #define NSECTORS	2048	/* number of sectors */
  
  int	mfs;			/* run as the memory based filesystem */
  int	Nflag;			/* run without writing file system */