From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Aneesh Kumar K.V" Subject: Re: [RFC][PATCH v2 10/21] mm: build separate zonelist for PMEM and DRAM node Date: Tue, 01 Jan 2019 14:44:41 +0530 Message-ID: <87sgyc7n9a.fsf@linux.ibm.com> References: <20181226131446.330864849@intel.com> <20181226133351.644607371@intel.com> Mime-Version: 1.0 Content-Type: text/plain Cc: Linux Memory Management List , Fan Du , Fengguang Wu , kvm@vger.kernel.org, LKML , Yao Yuan , Peng Dong , Huang Ying , Liu Jingqi , Dong Eddie , Dave Hansen , Zhang Yi , Dan Williams To: Fengguang Wu , Andrew Morton Return-path: In-Reply-To: <20181226133351.644607371@intel.com> Sender: linux-kernel-owner@vger.kernel.org List-Id: kvm.vger.kernel.org Fengguang Wu writes: > From: Fan Du > > When allocate page, DRAM and PMEM node should better not fall back to > each other. This allows migration code to explicitly control which type > of node to allocate pages from. > > With this patch, PMEM NUMA node can only be used in 2 ways: > - migrate in and out > - numactl Can we achieve this using nodemask? That way we don't tag nodes with different properties such as DRAM/PMEM. We can then give the flexibilility to the device init code to add the new memory nodes to the right nodemask -aneesh From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-qt1-f198.google.com (mail-qt1-f198.google.com [209.85.160.198]) by kanga.kvack.org (Postfix) with ESMTP id CF7A08E0002 for ; Tue, 1 Jan 2019 04:14:59 -0500 (EST) Received: by mail-qt1-f198.google.com with SMTP id w15so36510788qtk.19 for ; Tue, 01 Jan 2019 01:14:59 -0800 (PST) Received: from mx0a-001b2d01.pphosted.com (mx0b-001b2d01.pphosted.com. [148.163.158.5]) by mx.google.com with ESMTPS id q11si900217qvb.83.2019.01.01.01.14.58 for (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Tue, 01 Jan 2019 01:14:58 -0800 (PST) Received: from pps.filterd (m0098421.ppops.net [127.0.0.1]) by mx0a-001b2d01.pphosted.com (8.16.0.22/8.16.0.22) with SMTP id x019EAue034711 for ; Tue, 1 Jan 2019 04:14:58 -0500 Received: from e06smtp04.uk.ibm.com (e06smtp04.uk.ibm.com [195.75.94.100]) by mx0a-001b2d01.pphosted.com with ESMTP id 2pr417ttfh-1 (version=TLSv1.2 cipher=AES256-GCM-SHA384 bits=256 verify=NOT) for ; Tue, 01 Jan 2019 04:14:58 -0500 Received: from localhost by e06smtp04.uk.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Tue, 1 Jan 2019 09:14:56 -0000 From: "Aneesh Kumar K.V" Subject: Re: [RFC][PATCH v2 10/21] mm: build separate zonelist for PMEM and DRAM node In-Reply-To: <20181226133351.644607371@intel.com> References: <20181226131446.330864849@intel.com> <20181226133351.644607371@intel.com> Date: Tue, 01 Jan 2019 14:44:41 +0530 MIME-Version: 1.0 Content-Type: text/plain Message-Id: <87sgyc7n9a.fsf@linux.ibm.com> Sender: owner-linux-mm@kvack.org List-ID: To: Fengguang Wu , Andrew Morton Cc: Linux Memory Management List , Fan Du , kvm@vger.kernel.org, LKML , Yao Yuan , Peng Dong , Huang Ying , Liu Jingqi , Dong Eddie , Dave Hansen , Zhang Yi , Dan Williams Fengguang Wu writes: > From: Fan Du > > When allocate page, DRAM and PMEM node should better not fall back to > each other. This allows migration code to explicitly control which type > of node to allocate pages from. > > With this patch, PMEM NUMA node can only be used in 2 ways: > - migrate in and out > - numactl Can we achieve this using nodemask? That way we don't tag nodes with different properties such as DRAM/PMEM. We can then give the flexibilility to the device init code to add the new memory nodes to the right nodemask -aneesh