From mboxrd@z Thu Jan 1 00:00:00 1970 From: Rik van Riel Subject: Re: [PATCH 0/4] VM deadlock prevention -v5 Date: Fri, 25 Aug 2006 12:04:10 -0400 Message-ID: <44EF1F7A.3080001@redhat.com> References: <20060825153946.24271.42758.sendpatchset@twins> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Cc: Peter Zijlstra , linux-mm@kvack.org, linux-kernel@vger.kernel.org, netdev@vger.kernel.org, Indan Zupancic , Evgeniy Polyakov , Daniel Phillips , David Miller Return-path: Received: from mx1.redhat.com ([66.187.233.31]:711 "EHLO mx1.redhat.com") by vger.kernel.org with ESMTP id S1030200AbWHYQEe (ORCPT ); Fri, 25 Aug 2006 12:04:34 -0400 To: Christoph Lameter In-Reply-To: Sender: netdev-owner@vger.kernel.org List-Id: netdev.vger.kernel.org Christoph Lameter wrote: > On Fri, 25 Aug 2006, Peter Zijlstra wrote: > >> The basic premises is that network sockets serving the VM need undisturbed >> functionality in the face of severe memory shortage. >> >> This patch-set provides the framework to provide this. > > Hmmm.. Is it not possible to avoid the memory pools by > guaranteeing that a certain number of page is easily reclaimable? No. You need to guarantee that the memory is not gobbled up by another subsystem, but remains available for use by *this* subsystem. Otherwise you could still deadlock. -- What is important? What you want to be true, or what is true?