From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932078AbWFKQJK (ORCPT ); Sun, 11 Jun 2006 12:09:10 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751321AbWFKQJK (ORCPT ); Sun, 11 Jun 2006 12:09:10 -0400 Received: from vbn.0050556.lodgenet.net ([216.142.194.234]:33960 "EHLO vbn.0050556.lodgenet.net") by vger.kernel.org with ESMTP id S1751637AbWFKQJJ (ORCPT ); Sun, 11 Jun 2006 12:09:09 -0400 Subject: Re: [PATCH]: Adding a counter in vma to indicate the number of physical pages backing it From: Arjan van de Ven To: rohitseth@google.com Cc: Andrew Morton , Linux-mm@kvack.org, Linux-kernel@vger.kernel.org In-Reply-To: <1149903235.31417.84.camel@galaxy.corp.google.com> References: <1149903235.31417.84.camel@galaxy.corp.google.com> Content-Type: text/plain Date: Sun, 11 Jun 2006 18:09:02 +0200 Message-Id: <1150042142.3131.82.camel@laptopd505.fenrus.org> Mime-Version: 1.0 X-Mailer: Evolution 2.2.3 (2.2.3-2.fc4) Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Fri, 2006-06-09 at 18:33 -0700, Rohit Seth wrote: > Below is a patch that adds number of physical pages that each vma is > using in a process. Exporting this information to user space > using /proc//maps interface. is it really worth bloating the vma struct for this? there are quite a few workloads that have a gazilion vma's, and this patch adds both memory usage and cache pressure to those workloads...