From mboxrd@z Thu Jan  1 00:00:00 1970
From: Nick Piggin <npiggin@suse.de>
Subject: Re: [patch] cifs: use add_to_page_cache_lru
Date: Fri, 19 Mar 2010 12:13:10 +1100
Message-ID: <20100319011310.GG25636@laptop>
References: <20100317062316.GC2869@laptop>
 <524f69651003181143i67d2fbbdw148619b6065bca88@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: linux-fsdevel@vger.kernel.org, linux-cifs-client@lists.samba.org,
	Steve French <sfrench@samba.org>
To: Steve French <smfrench@gmail.com>
Return-path: <linux-fsdevel-owner@vger.kernel.org>
Received: from cantor.suse.de ([195.135.220.2]:51456 "EHLO mx1.suse.de"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1751606Ab0CSBNY (ORCPT <rfc822;linux-fsdevel@vger.kernel.org>);
	Thu, 18 Mar 2010 21:13:24 -0400
Content-Disposition: inline
In-Reply-To: <524f69651003181143i67d2fbbdw148619b6065bca88@mail.gmail.com>
Sender: linux-fsdevel-owner@vger.kernel.org
List-ID: <linux-fsdevel.vger.kernel.org>

On Thu, Mar 18, 2010 at 01:43:17PM -0500, Steve French wrote:
> Seems reasonable to me (also Shaggy acked this).  Any idea how likely
> it would affect performance (and would thus need to run through some
> simple iozone and cthon (timing mode) or similar perf testing)

It should positively affect LRU ordering, and CPU efficiency, but both
such a tiny amount it should be in the noise. It wouldn't affect IO
patterns (except as a side effect of slightly different pagecache
reclaim, but again in the noise).

I wouldn't worry about benchmarking it.