From owner-freebsd-current  Wed Apr  2 12:23:14 1997
Return-Path: <owner-current>
Received: (from root@localhost)
          by freefall.freebsd.org (8.8.5/8.8.5) id MAA21415
          for current-outgoing; Wed, 2 Apr 1997 12:23:14 -0800 (PST)
Received: from phaeton.artisoft.com (phaeton.Artisoft.COM [198.17.250.50])
          by freefall.freebsd.org (8.8.5/8.8.5) with SMTP id MAA21374;
          Wed, 2 Apr 1997 12:22:43 -0800 (PST)
Received: (from terry@localhost) by phaeton.artisoft.com (8.6.11/8.6.9) id NAA14273; Wed, 2 Apr 1997 13:05:42 -0700
From: Terry Lambert <terry@lambert.org>
Message-Id: <199704022005.NAA14273@phaeton.artisoft.com>
Subject: Re: ufs lock panic in -current
To: phk@critter.dk.tfs.com (Poul-Henning Kamp)
Date: Wed, 2 Apr 1997 13:05:42 -0700 (MST)
Cc: terry@lambert.org, ache@nagual.ru, dyson@freebsd.org, bde@zeta.org.au,
        current@freebsd.org
In-Reply-To: <247.860009451@critter> from "Poul-Henning Kamp" at Apr 2, 97 09:30:51 pm
X-Mailer: ELM [version 2.4 PL24]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-current@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

> (Terry sent his standard reply #1:)
> >
> >The fix is conceptually non-trivial, and requires that the transitive
> >closure be calculated at the same layer for all FS's.  This implies a
> >veto, rather than a call-down, interface for VOP_LOCK.
> >
> >Physically, the fix is about 200 lines of code, including all FS
> >code changes.
> 
> (And I send my standard reply #1:)
> 
> Well, send us the patch Terry!


Have you guys integrated the patches I sent to Julian for the
namei/nameifree fixes and the redundant code reduction "EXCLUDE"
NDINIT() op flag yet?

I even broke them out as seperate from all my other patches, even
though it meant buying another disk and building a seperate source
tree to do it: I've spend a not inconsiderable amount of money
trying to appease you by hopping on the correct foot during the
chanting and incense burning.


If not, which will you integrate first, the lock veto fixes, or
the layering fixes?

I only ask so that when I hit the same function in the same file
for both of them, I get the dependency order right.


Or am I just supposed to send patches that work regardless of the
dependency order?

I'll remind you that dependencies aren't acyclic, unless we move to
one function per object module, which would be ridiculous.


I did send some order independent patches to you once (by stuffing
in all the patches at once), but you said that they touched too much
stuff at one time, so they couldn't be reviewed...

Please make up your mind, and let me know when you're done.


					Regards,
					Terry Lambert
					terry@lambert.org
---
Any opinions in this posting are my own and not those of my present
or previous employers.