From mboxrd@z Thu Jan 1 00:00:00 1970 From: Paul Brook Subject: Re: [Qemu-devel] [PATCH 4/4] Reallocate dma buffers in read/write path if needed Date: Sun, 5 Oct 2008 00:00:27 +0100 Message-ID: <200810050000.28154.paul@codesourcery.com> References: <1223071531-31817-1-git-send-email-ryanh@us.ibm.com> <20081004135749.pphehrhuw9w4gwsc@imap.linux.ibm.com> <20081004214700.GH31395@us.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: aliguori@linux.ibm.com, Avi Kivity , qemu-devel@nongnu.org, Anthony Liguori , kvm@vger.kernel.org To: Ryan Harper Return-path: Received: from mail.codesourcery.com ([65.74.133.4]:59658 "EHLO mail.codesourcery.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754449AbYJDXAa convert rfc822-to-8bit (ORCPT ); Sat, 4 Oct 2008 19:00:30 -0400 In-Reply-To: <20081004214700.GH31395@us.ibm.com> Content-Disposition: inline Sender: kvm-owner@vger.kernel.org List-ID: On Saturday 04 October 2008, Ryan Harper wrote: > In all, it seems silly to worry about this sort of thing since the > entire process could be contained with process ulimits if this is rea= lly > a concern. =A0Are we any more concerned that by splitting the request= s > into many smaller requests that we're wasting cpu, pegging the > processor to 100% in some cases? Using small requests may be a bit inefficient, but it still works and a= llows=20 the guest to make progress. Allocating very large quantities of memory is very likely to kill the V= M one=20 way or another. This is not acceptable, especially when the guest hasn'= t even=20 done anything wrong. There are legitimate circumstances where the size = of the=20 outstanding IO requests may be comparable to the guest ram size. Paul