From owner-freebsd-current  Thu Dec  7 12:46:39 1995
Return-Path: owner-current
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id MAA17856
          for current-outgoing; Thu, 7 Dec 1995 12:46:39 -0800 (PST)
Received: from phaeton.artisoft.com (phaeton.Artisoft.COM [198.17.250.211])
          by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id MAA17842
          for <current@freebsd.org>; Thu, 7 Dec 1995 12:46:35 -0800 (PST)
Received: (from terry@localhost) by phaeton.artisoft.com (8.6.11/8.6.9) id NAA05505 for current@freebsd.org; Thu, 7 Dec 1995 13:45:03 -0700
From: Terry Lambert <terry@lambert.org>
Message-Id: <199512072045.NAA05505@phaeton.artisoft.com>
Subject: I hates leases to pieces
To: current@freebsd.org
Date: Thu, 7 Dec 1995 13:45:03 -0700 (MST)
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
Precedence: bulk

Now for page 2 on leases:

Block profiling shows that lock manipulation is a significant fraction
of the time in many types of access.

Why aren't vp's allowed to be locked when calling lease checking?

There doesn't seem to be an explicit reason for this, and it complicates
the code by changing:

	ndinit(&nd, ...|LOCKLEAF, ...);
	LEASE_CHECK(nd.vp, ...);

Into:

	ndinit(&nd, ..., ...);
	LEASE_CHECK(nd.vp, ...);
	VOP_LOCK(nd.vp);

And of course namei() will have locked the leaf and explicitly calls
VOP_UNLOCK().


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