From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Srivatsa S. Bhat" Subject: Re: [RFC PATCH 1/8] mm: Introduce memory regions data-structure to capture region boundaries within node Date: Thu, 08 Nov 2012 01:42:52 +0530 Message-ID: <509AC0C4.4030704@linux.vnet.ibm.com> References: <20121106195026.6941.24662.stgit@srivatsabhat.in.ibm.com> <20121106195225.6941.2868.stgit@srivatsabhat.in.ibm.com> <50999755.4000209@linux.vnet.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Return-path: Received: from e23smtp06.au.ibm.com ([202.81.31.148]:38558 "EHLO e23smtp06.au.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752458Ab2KGUOH (ORCPT ); Wed, 7 Nov 2012 15:14:07 -0500 Received: from /spool/local by e23smtp06.au.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Thu, 8 Nov 2012 06:11:41 +1000 In-Reply-To: <50999755.4000209@linux.vnet.ibm.com> Sender: linux-pm-owner@vger.kernel.org List-Id: linux-pm@vger.kernel.org To: Dave Hansen Cc: akpm@linux-foundation.org, mgorman@suse.de, mjg59@srcf.ucam.org, paulmck@linux.vnet.ibm.com, maxime.coquelin@stericsson.com, loic.pallardy@stericsson.com, arjan@linux.intel.com, kmpark@infradead.org, kamezawa.hiroyu@jp.fujitsu.com, lenb@kernel.org, rjw@sisk.pl, gargankita@gmail.com, amit.kachhap@linaro.org, svaidy@linux.vnet.ibm.com, thomas.abraham@linaro.org, santosh.shilimkar@ti.com, linux-pm@vger.kernel.org, linux-mm@kvack.org, linux-kernel@vger.kernel.org On 11/07/2012 04:33 AM, Dave Hansen wrote: > On 11/06/2012 11:52 AM, Srivatsa S. Bhat wrote: >> But of course, memory regions are sub-divisions *within* a node, so it makes >> sense to keep the data-structures in the node's struct pglist_data. (Thus >> this placement makes memory regions parallel to zones in that node). > > I think it's pretty silly to create *ANOTHER* subdivision of memory > separate from sparsemem. One that doesn't handle large amounts of > memory or scale with memory hotplug. As it stands, you can only support > 256*512MB=128GB of address space, which seems pretty puny. > > This node_regions[]: > >> @@ -687,6 +698,8 @@ typedef struct pglist_data { >> struct zone node_zones[MAX_NR_ZONES]; >> struct zonelist node_zonelists[MAX_ZONELISTS]; >> int nr_zones; >> + struct node_mem_region node_regions[MAX_NR_REGIONS]; >> + int nr_node_regions; >> #ifdef CONFIG_FLAT_NODE_MEM_MAP /* means !SPARSEMEM */ >> struct page *node_mem_map; >> #ifdef CONFIG_MEMCG > > looks like it's indexed the same way regardless of which node it is in. > In other words, if there are two nodes, at least half of it is wasted, > and 3/4 if there are four nodes. That seems a bit suboptimal. > You're right, I have not addressed that problem in this initial RFC. Thanks for pointing it out! Going forward, we can surely optimize the way we deal with memory regions on NUMA systems, using some of the sparsemem techniques. > Could you remind us of the logic for leaving sparsemem out of the > equation here? > Nothing, its just that in this first RFC I was more focussed towards getting the overall design right, in terms of having an acceptable way of tracking pages belonging to different regions within the page allocator (freelists) and using it to influence page allocation decisions. And also to compare the merits of this approach over the previous "Hierarchy" design, in a broad ("big picture") sense. I'll add the above point you raised in my todo-list and address it in subsequent versions of the patchset. Thank you very much for the quick feedback! Regards, Srivatsa S. Bhat