From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S933539AbdAJXiA (ORCPT ); Tue, 10 Jan 2017 18:38:00 -0500 Received: from mailapp01.imgtec.com ([195.59.15.196]:20260 "EHLO mailapp01.imgtec.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S933465AbdAJXhz (ORCPT ); Tue, 10 Jan 2017 18:37:55 -0500 Subject: Re: [PATCH] mm: page_alloc: Skip over regions of invalid pfns where possible To: Andrew Morton , Paul Burton References: <20161125185518.29885-1-paul.burton@imgtec.com> <20170106144348.f7d207baa7b3190a95aaeb2e@linux-foundation.org> CC: , From: James Hartley Message-ID: <0f03d5c6-182c-d30f-68ef-8d1a767bfcf8@imgtec.com> Date: Tue, 10 Jan 2017 23:37:53 +0000 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Thunderbird/45.5.1 MIME-Version: 1.0 In-Reply-To: <20170106144348.f7d207baa7b3190a95aaeb2e@linux-foundation.org> Content-Type: text/plain; charset="windows-1252" Content-Transfer-Encoding: 7bit X-Originating-IP: [10.40.8.58] Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 06/01/17 22:43, Andrew Morton wrote: > On Fri, 25 Nov 2016 18:55:18 +0000 Paul Burton wrote: > >> When using a sparse memory model memmap_init_zone() when invoked with >> the MEMMAP_EARLY context will skip over pages which aren't valid - ie. >> which aren't in a populated region of the sparse memory map. However if >> the memory map is extremely sparse then it can spend a long time >> linearly checking each PFN in a large non-populated region of the memory >> map & skipping it in turn. >> >> When CONFIG_HAVE_MEMBLOCK_NODE_MAP is enabled, we have sufficient >> information to quickly discover the next valid PFN given an invalid one >> by searching through the list of memory regions & skipping forwards to >> the first PFN covered by the memory region to the right of the >> non-populated region. Implement this in order to speed up >> memmap_init_zone() for systems with extremely sparse memory maps. > Could we have a changelog which includes some timing measurements? > That permits others to understand the value of this patch. > I have tested this patch on a virtual model of a Samurai CPU with a sparse memory map. The kernel boot time drops from 109 to 62 seconds. James