From owner-freebsd-fs  Mon Aug 24 12:04:30 1998
Return-Path: <owner-freebsd-fs@FreeBSD.ORG>
Received: (from majordom@localhost)
          by hub.freebsd.org (8.8.8/8.8.8) id MAA18649
          for freebsd-fs-outgoing; Mon, 24 Aug 1998 12:04:30 -0700 (PDT)
          (envelope-from owner-freebsd-fs@FreeBSD.ORG)
Received: from renko.ucs.ed.ac.uk (renko.ucs.ed.ac.uk [129.215.13.3])
          by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id MAA18608
          for <freebsd-fs@FreeBSD.ORG>; Mon, 24 Aug 1998 12:04:22 -0700 (PDT)
          (envelope-from sct@redhat.com)
Received: from dax.dcs.ed.ac.uk (root@dialup-119.publab.ed.ac.uk [129.215.38.119])
	by renko.ucs.ed.ac.uk (8.8.7/8.8.7) with ESMTP id UAA27811;
	Mon, 24 Aug 1998 20:03:29 +0100 (BST)
Received: (from sct@localhost)
	by dax.dcs.ed.ac.uk (8.8.7/8.8.7) id SAA25353;
	Mon, 24 Aug 1998 18:57:08 +0100
Date: Mon, 24 Aug 1998 18:57:08 +0100
Message-Id: <199808241757.SAA25353@dax.dcs.ed.ac.uk>
From: "Stephen C. Tweedie" <sct@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
To: MOLNAR Ingo <mingo@valerie.inf.elte.hu>
Cc: "Stephen C. Tweedie" <sct@redhat.com>, poipoi@famipow.com,
        freebsd-fs@FreeBSD.ORG, linux-fsdevel@vger.rutgers.edu
Subject: Re: name cache wiping in readdir ?
In-Reply-To: <Pine.GSO.3.96.980824155142.9210B-100000@valerie.inf.elte.hu>
References: <199808240934.KAA21958@dax.dcs.ed.ac.uk>
	<Pine.GSO.3.96.980824155142.9210B-100000@valerie.inf.elte.hu>
Sender: owner-freebsd-fs@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.org

Hi,

On Mon, 24 Aug 1998 15:53:51 +0200 (MET DST), MOLNAR Ingo
<mingo@valerie.inf.elte.hu> said:

> On Mon, 24 Aug 1998, Stephen C. Tweedie wrote:

>> One thing which I would like to try (we did it in earlier versions of
>> the primitive name cache in 2.0) is a two-level name cache, where we
>> promote names up to the next level on a hit and demote them when we want
>> to emply a list. [...]

> we have dentry aging which should take care of these issues. Whenever we
> reply to memory pressure, we are more likely to free unused dentries.

Yes.  The multi-level architecture is just an optimisation which lets us
deal with the common "find" special case without flushing _any_ of our
established dcache entries.  With an aging mechanism, the best that
happens is that a find leaves active entries alone but still ages and
eventually flushes out the initial contents of the cache.

--Stephen


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-fs" in the body of the message