From owner-freebsd-current  Fri Sep 22 06:53:46 1995
Return-Path: owner-current
Received: (from root@localhost)
          by freefall.freebsd.org (8.6.12/8.6.6) id GAA01938
          for current-outgoing; Fri, 22 Sep 1995 06:53:46 -0700
Received: (from dyson@localhost)
          by freefall.freebsd.org (8.6.12/8.6.6) id GAA01931
          for current@freebsd.org; Fri, 22 Sep 1995 06:53:44 -0700
Date: Fri, 22 Sep 1995 06:53:44 -0700
From: John Dyson <dyson>
Message-Id: <199509221353.GAA01931@freefall.freebsd.org>
To: current@freebsd.org
Subject: The Sig-11/10 saga & Thanks DG!!!
Sender: owner-current@freebsd.org
Precedence: bulk

David has come up with a work-around, and last night (at about 4:00 AM
my time) finally I found my bug.  The bug was NOT in the new VFS layered
GETPAGES routine where I kept looking :-(, but in some of the mods to the
clustering code for EXT2FS!!!  In hindsight, the code was pathetically
broken, but could not see it.  I am going to wait until at least 16:00 EST
to commit the fix (and another one that DG found in the ufs_bmap code from
Lite and Lite-2) so that everyone can grab a working system.  The work-around
does impact read perf significantly, but of course that is better than
Sig-11s all over the place :-).  The performance with the complete fix
will be better again.

I am so glad that DG pointed me in the right direction on this problem,
otherwise I would have spent more time spinning my wheels!!!  Thanks
David!!!  Also, look forward to the fix being commited today...  I'll
email to -current when it happens....  Now, I get to sleep tonight...

John
dyson@freebsd.org