From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261824AbVACSj5 (ORCPT ); Mon, 3 Jan 2005 13:39:57 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261784AbVACSf6 (ORCPT ); Mon, 3 Jan 2005 13:35:58 -0500 Received: from dhcp93115068.columbus.rr.com ([24.93.115.68]:6411 "EHLO nineveh.rivenstone.net") by vger.kernel.org with ESMTP id S261775AbVACSbk convert rfc822-to-8bit (ORCPT ); Mon, 3 Jan 2005 13:31:40 -0500 Date: Mon, 3 Jan 2005 13:31:34 -0500 From: Joseph Fannin To: Luke Kenneth Casson Leighton Cc: linux-kernel@vger.kernel.org, xen-devel@lists.sf.net Subject: Re: [XEN] using shmfs for swapspace Message-ID: <20050103183133.GA19081@samarkand.rivenstone.net> Mail-Followup-To: Luke Kenneth Casson Leighton , linux-kernel@vger.kernel.org, xen-devel@lists.sf.net References: <20050102162652.GA12268@lkcl.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: 8BIT In-Reply-To: <20050102162652.GA12268@lkcl.net> User-Agent: Mutt/1.4i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Sun, Jan 02, 2005 at 04:26:52PM +0000, Luke Kenneth Casson Leighton wrote: [...] > this is presumed to be infinitely better than forcing the swapspace to > be always on disk, especially with the guests only being allocated > 32mbyte of physical RAM. I'd be interested in knowing how a tmpfs that's gone far into swap performs compared to a more normal on-disk fs. I don't know if anyone has ever looked into it. Is it comparable, or is tmpfs's ability to swap more a last-resort escape hatch? This is the part where I would add something valuable to this conversation, if I were going to do that. (But no.) -- Joseph Fannin jhf@rivenstone.net