From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-pa0-f53.google.com (mail-pa0-f53.google.com [209.85.220.53]) by kanga.kvack.org (Postfix) with ESMTP id DDB5F82F64 for ; Fri, 30 Oct 2015 15:42:20 -0400 (EDT) Received: by padhy1 with SMTP id hy1so76809531pad.0 for ; Fri, 30 Oct 2015 12:42:20 -0700 (PDT) Received: from mga03.intel.com (mga03.intel.com. [134.134.136.65]) by mx.google.com with ESMTP id td8si13112276pac.42.2015.10.30.12.42.19 for ; Fri, 30 Oct 2015 12:42:20 -0700 (PDT) From: "Luck, Tony" Subject: RE: [PATCH] mm: Introduce kernelcore=reliable option Date: Fri, 30 Oct 2015 19:42:17 +0000 Message-ID: <3908561D78D1C84285E8C5FCA982C28F32B64312@ORSMSX114.amr.corp.intel.com> References: <1444915942-15281-1-git-send-email-izumi.taku@jp.fujitsu.com> <3908561D78D1C84285E8C5FCA982C28F32B5A060@ORSMSX114.amr.corp.intel.com> <5628B427.3050403@jp.fujitsu.com> <3908561D78D1C84285E8C5FCA982C28F32B5C7AE@ORSMSX114.amr.corp.intel.com> <322B7BFA-08FE-4A8F-B54C-86901BDB7CBD@intel.com> <56330C0A.3060901@jp.fujitsu.com> In-Reply-To: <56330C0A.3060901@jp.fujitsu.com> Content-Language: en-US Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Sender: owner-linux-mm@kvack.org List-ID: To: Kamezawa Hiroyuki , "Izumi, Taku" Cc: "linux-kernel@vger.kernel.org" , "linux-mm@kvack.org" , "qiuxishi@huawei.com" , "mel@csn.ul.ie" , "akpm@linux-foundation.org" , "Hansen, Dave" , "matt@codeblueprint.co.uk" > If each memory controller has the same distance/latency, you (your firmwa= re) don't need > to allocate reliable memory per each memory controller. > If distance is problem, another node should be allocated. > > ...is the behavior(splitting zone) really required ? It's useful from a memory bandwidth perspective to have allocations spread across both memory controllers. Keeping a whole bunch of Xeon cores fed needs all the bandwidth you can get. Socket0 is also a problem. We want to mirror <4GB addresses because there is a bunch of critical stuff there (entire kernel text+data). But we can currently only mirror one block per memory controller, so we end up with just 2GB mirrored (the 2GB-4GB range is MMIO). This isn't enough for even a small machine (I have 128GB on node0 ... but that is really the bare minimum configuration ... 2GB is only enough to cover the "struct page" allocations for node0). I really have to allocate some more mirror from the other memory controller. -Tony -- 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