From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S964825AbVHaPES (ORCPT ); Wed, 31 Aug 2005 11:04:18 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S932496AbVHaPES (ORCPT ); Wed, 31 Aug 2005 11:04:18 -0400 Received: from silver.veritas.com ([143.127.12.111]:6920 "EHLO silver.veritas.com") by vger.kernel.org with ESMTP id S932380AbVHaPES (ORCPT ); Wed, 31 Aug 2005 11:04:18 -0400 Date: Wed, 31 Aug 2005 16:06:12 +0100 (BST) From: Hugh Dickins X-X-Sender: hugh@goblin.wat.veritas.com To: "Martin J. Bligh" cc: Arjan van de Ven , Dave McCracken , Andrew Morton , Linux Kernel , Linux Memory Management Subject: Re: [PATCH 1/1] Implement shared page tables In-Reply-To: <16640000.1125498711@[10.10.2.4]> Message-ID: References: <7C49DFF721CB4E671DB260F9@[10.1.1.4]> <1125489077.3213.12.camel@laptopd505.fenrus.org> <16640000.1125498711@[10.10.2.4]> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-OriginalArrivalTime: 31 Aug 2005 15:04:17.0844 (UTC) FILETIME=[424BF740:01C5AE3D] Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Wed, 31 Aug 2005, Martin J. Bligh wrote: > --Hugh Dickins wrote (on Wednesday, August 31, 2005 14:42:38 +0100): > > > > Which is indeed a further disincentive against shared page tables. > > Or shared pagetables a disincentive to randomizing the mmap space ;-) Fair point! > They're incompatible, but you could be left to choose one or the other > via config option. Wouldn't need config option: there's /proc/sys/kernel/randomize_va_space for the whole running system, compatibility check on the ELFs run, and the infinite stack rlimit: enough ways to suppress randomization if it doesn't suit you. > 3% on "a certain industry-standard database benchmark" (cough) is huge, > and we expect the benefit for PPC64 will be larger as we can share the > underlying hardware PTEs without TLB flushing as well. Okay - and you're implying that 3% comes from _using_ the shared page tables, rather than from avoiding the fork/exit overhead of setting them up and tearing them down. And it can't use huge TLB pages because... fragmentation? Hugh