From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757416AbcJXTgN (ORCPT ); Mon, 24 Oct 2016 15:36:13 -0400 Received: from mga07.intel.com ([134.134.136.100]:33958 "EHLO mga07.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757257AbcJXTgM (ORCPT ); Mon, 24 Oct 2016 15:36:12 -0400 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.31,543,1473145200"; d="scan'208";a="893448354" Subject: Re: [RFC 0/8] Define coherent device memory node To: David Nellans , Anshuman Khandual , linux-kernel@vger.kernel.org, linux-mm@kvack.org References: <1477283517-2504-1-git-send-email-khandual@linux.vnet.ibm.com> <580E4D2D.2070408@intel.com> <6f96676c-c1cb-c08b-1dea-8d6e6c6c3c68@nvidia.com> Cc: mhocko@suse.com, js1304@gmail.com, vbabka@suse.cz, mgorman@suse.de, minchan@kernel.org, akpm@linux-foundation.org, aneesh.kumar@linux.vnet.ibm.com, bsingharora@gmail.com From: Dave Hansen Message-ID: <580E62AB.8040303@intel.com> Date: Mon, 24 Oct 2016 12:36:11 -0700 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101 Thunderbird/38.8.0 MIME-Version: 1.0 In-Reply-To: <6f96676c-c1cb-c08b-1dea-8d6e6c6c3c68@nvidia.com> Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 10/24/2016 11:32 AM, David Nellans wrote: > On 10/24/2016 01:04 PM, Dave Hansen wrote: >> If you *really* don't want a "cdm" page to be migrated, then why isn't >> that policy set on the VMA in the first place? That would keep "cdm" >> pages from being made non-cdm. And, why would autonuma ever make a >> non-cdm page and migrate it in to cdm? There will be no NUMA access >> faults caused by the devices that are fed to autonuma. >> > Pages are desired to be migrateable, both into (starting cpu zone > movable->cdm) and out of (starting cdm->cpu zone movable) but only > through explicit migration, not via autonuma. OK, and is there a reason that the existing mbind code plus NUMA policies fails to give you this behavior? Does autonuma somehow override strict NUMA binding? > other pages in the same > VMA should still be migrateable between CPU nodes via autonuma however. That's not the way the implementation here works, as I understand it. See the VM_CDM patch and my responses to it. > Its expected a lot of these allocations are going to end up in THPs. > I'm not sure we need to explicitly disallow hugetlbfs support but the > identified use case is definitely via THPs not tlbfs. I think THP and hugetlbfs are implementations, not use cases. :) Is it too hard to support hugetlbfs that we should complicate its code to exclude it from this type of memory? Why? From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-pf0-f199.google.com (mail-pf0-f199.google.com [209.85.192.199]) by kanga.kvack.org (Postfix) with ESMTP id CAD566B0262 for ; Mon, 24 Oct 2016 15:36:12 -0400 (EDT) Received: by mail-pf0-f199.google.com with SMTP id n18so20929603pfe.7 for ; Mon, 24 Oct 2016 12:36:12 -0700 (PDT) Received: from mga09.intel.com (mga09.intel.com. [134.134.136.24]) by mx.google.com with ESMTPS id k17si784873pgh.279.2016.10.24.12.36.11 for (version=TLS1 cipher=AES128-SHA bits=128/128); Mon, 24 Oct 2016 12:36:12 -0700 (PDT) Subject: Re: [RFC 0/8] Define coherent device memory node References: <1477283517-2504-1-git-send-email-khandual@linux.vnet.ibm.com> <580E4D2D.2070408@intel.com> <6f96676c-c1cb-c08b-1dea-8d6e6c6c3c68@nvidia.com> From: Dave Hansen Message-ID: <580E62AB.8040303@intel.com> Date: Mon, 24 Oct 2016 12:36:11 -0700 MIME-Version: 1.0 In-Reply-To: <6f96676c-c1cb-c08b-1dea-8d6e6c6c3c68@nvidia.com> Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 7bit Sender: owner-linux-mm@kvack.org List-ID: To: David Nellans , Anshuman Khandual , linux-kernel@vger.kernel.org, linux-mm@kvack.org Cc: mhocko@suse.com, js1304@gmail.com, vbabka@suse.cz, mgorman@suse.de, minchan@kernel.org, akpm@linux-foundation.org, aneesh.kumar@linux.vnet.ibm.com, bsingharora@gmail.com On 10/24/2016 11:32 AM, David Nellans wrote: > On 10/24/2016 01:04 PM, Dave Hansen wrote: >> If you *really* don't want a "cdm" page to be migrated, then why isn't >> that policy set on the VMA in the first place? That would keep "cdm" >> pages from being made non-cdm. And, why would autonuma ever make a >> non-cdm page and migrate it in to cdm? There will be no NUMA access >> faults caused by the devices that are fed to autonuma. >> > Pages are desired to be migrateable, both into (starting cpu zone > movable->cdm) and out of (starting cdm->cpu zone movable) but only > through explicit migration, not via autonuma. OK, and is there a reason that the existing mbind code plus NUMA policies fails to give you this behavior? Does autonuma somehow override strict NUMA binding? > other pages in the same > VMA should still be migrateable between CPU nodes via autonuma however. That's not the way the implementation here works, as I understand it. See the VM_CDM patch and my responses to it. > Its expected a lot of these allocations are going to end up in THPs. > I'm not sure we need to explicitly disallow hugetlbfs support but the > identified use case is definitely via THPs not tlbfs. I think THP and hugetlbfs are implementations, not use cases. :) Is it too hard to support hugetlbfs that we should complicate its code to exclude it from this type of memory? Why? -- 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