From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752694AbaBGUlm (ORCPT ); Fri, 7 Feb 2014 15:41:42 -0500 Received: from mail-pd0-f173.google.com ([209.85.192.173]:60212 "EHLO mail-pd0-f173.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751526AbaBGUll (ORCPT ); Fri, 7 Feb 2014 15:41:41 -0500 Date: Fri, 7 Feb 2014 12:41:38 -0800 (PST) From: David Rientjes X-X-Sender: rientjes@chino.kir.corp.google.com To: Raghavendra K T cc: Andrew Morton , Fengguang Wu , David Cohen , Al Viro , Damien Ramonda , Jan Kara , Linus , linux-mm@kvack.org, linux-kernel@vger.kernel.org Subject: Re: [RFC PATCH V5] mm readahead: Fix readahead fail for no local memory and limit readahead pages In-Reply-To: <52F4B8A4.70405@linux.vnet.ibm.com> Message-ID: References: <1390388025-1418-1-git-send-email-raghavendra.kt@linux.vnet.ibm.com> <20140206145105.27dec37b16f24e4ac5fd90ce@linux-foundation.org> <20140206152219.45c2039e5092c8ea1c31fd38@linux-foundation.org> <52F4B8A4.70405@linux.vnet.ibm.com> User-Agent: Alpine 2.02 (DEB 1266 2009-07-14) MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, 7 Feb 2014, Raghavendra K T wrote: > So following discussion TODO for my patch is: > > 1) Update the changelog with user visible impact of the patch. > (Andrew's suggestion) > 2) Add ACCESS_ONCE to numa_node_id(). > 3) Change the "readahead into remote memory" part of the documentation > which is misleading. > > ( I feel no need to add numa_mem_id() since we would specifically limit > the readahead with MAX_REMOTE_READAHEAD in memoryless cpu cases). > I don't understand what you're saying, numa_mem_id() is local memory to current's cpu. When a node consists only of cpus and not memory it is not true that all memory is then considered remote, you won't find that in any specification that defines memory affinity including the ACPI spec. I can trivially define all cpus on my system to be on memoryless nodes and having that affect readahead behavior when, in fact, there is affinity would be ridiculous. From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-pa0-f41.google.com (mail-pa0-f41.google.com [209.85.220.41]) by kanga.kvack.org (Postfix) with ESMTP id 298A86B0031 for ; Fri, 7 Feb 2014 15:41:43 -0500 (EST) Received: by mail-pa0-f41.google.com with SMTP id fa1so3655807pad.14 for ; Fri, 07 Feb 2014 12:41:42 -0800 (PST) Received: from mail-pb0-x22e.google.com (mail-pb0-x22e.google.com [2607:f8b0:400e:c01::22e]) by mx.google.com with ESMTPS id ez5si6324658pab.77.2014.02.07.12.41.40 for (version=TLSv1 cipher=ECDHE-RSA-RC4-SHA bits=128/128); Fri, 07 Feb 2014 12:41:41 -0800 (PST) Received: by mail-pb0-f46.google.com with SMTP id um1so3749044pbc.33 for ; Fri, 07 Feb 2014 12:41:40 -0800 (PST) Date: Fri, 7 Feb 2014 12:41:38 -0800 (PST) From: David Rientjes Subject: Re: [RFC PATCH V5] mm readahead: Fix readahead fail for no local memory and limit readahead pages In-Reply-To: <52F4B8A4.70405@linux.vnet.ibm.com> Message-ID: References: <1390388025-1418-1-git-send-email-raghavendra.kt@linux.vnet.ibm.com> <20140206145105.27dec37b16f24e4ac5fd90ce@linux-foundation.org> <20140206152219.45c2039e5092c8ea1c31fd38@linux-foundation.org> <52F4B8A4.70405@linux.vnet.ibm.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-mm@kvack.org List-ID: To: Raghavendra K T Cc: Andrew Morton , Fengguang Wu , David Cohen , Al Viro , Damien Ramonda , Jan Kara , Linus , linux-mm@kvack.org, linux-kernel@vger.kernel.org On Fri, 7 Feb 2014, Raghavendra K T wrote: > So following discussion TODO for my patch is: > > 1) Update the changelog with user visible impact of the patch. > (Andrew's suggestion) > 2) Add ACCESS_ONCE to numa_node_id(). > 3) Change the "readahead into remote memory" part of the documentation > which is misleading. > > ( I feel no need to add numa_mem_id() since we would specifically limit > the readahead with MAX_REMOTE_READAHEAD in memoryless cpu cases). > I don't understand what you're saying, numa_mem_id() is local memory to current's cpu. When a node consists only of cpus and not memory it is not true that all memory is then considered remote, you won't find that in any specification that defines memory affinity including the ACPI spec. I can trivially define all cpus on my system to be on memoryless nodes and having that affect readahead behavior when, in fact, there is affinity would be ridiculous. -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: email@kvack.org