From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 05BF1C433F5 for ; Tue, 15 Feb 2022 04:47:17 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S233921AbiBOErY (ORCPT ); Mon, 14 Feb 2022 23:47:24 -0500 Received: from mxb-00190b01.gslb.pphosted.com ([23.128.96.19]:43046 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229498AbiBOErW (ORCPT ); Mon, 14 Feb 2022 23:47:22 -0500 Received: from foss.arm.com (foss.arm.com [217.140.110.172]) by lindbergh.monkeyblade.net (Postfix) with ESMTP id 5E21390266 for ; Mon, 14 Feb 2022 20:47:13 -0800 (PST) Received: from usa-sjc-imap-foss1.foss.arm.com (unknown [10.121.207.14]) by usa-sjc-mx-foss1.foss.arm.com (Postfix) with ESMTP id 07FD01063; Mon, 14 Feb 2022 20:47:13 -0800 (PST) Received: from [10.163.48.145] (unknown [10.163.48.145]) by usa-sjc-imap-foss1.foss.arm.com (Postfix) with ESMTPSA id BFE7E3F718; Mon, 14 Feb 2022 20:47:10 -0800 (PST) Subject: Re: [PATCH] mm/pages_alloc.c: Don't create ZONE_MOVABLE beyond the end of a node To: Alistair Popple , akpm@linux-foundation.org Cc: jhubbard@nvidia.com, linux-mm@kvack.org, linux-kernel@vger.kernel.org, ziy@nvidia.com References: <20220215025831.2113067-1-apopple@nvidia.com> From: Anshuman Khandual Message-ID: <7b752e06-f345-cbb2-d05c-57e5fc5d8e5a@arm.com> Date: Tue, 15 Feb 2022 10:17:09 +0530 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101 Thunderbird/68.10.0 MIME-Version: 1.0 In-Reply-To: <20220215025831.2113067-1-apopple@nvidia.com> Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 7bit Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi Alistair, On 2/15/22 8:28 AM, Alistair Popple wrote: > ZONE_MOVABLE uses the remaining memory in each node. It's starting pfn > is also aligned to MAX_ORDER_NR_PAGES. It is possible for the remaining > memory in a node to be less than MAX_ORDER_NR_PAGES, meaning there is > not enough room for ZONE_MOVABLE on that node. How plausible is this scenario on normal systems ? Should not the node always contain MAX_ORDER_NR_PAGES aligned pages ? Also all zones which get created from that node should also be MAX_ORDER_NR_PAGES aligned ? I am just curious how a node could end up being like this. - Anshuman