From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753791AbaFQBAb (ORCPT ); Mon, 16 Jun 2014 21:00:31 -0400 Received: from mail-pa0-f51.google.com ([209.85.220.51]:57900 "EHLO mail-pa0-f51.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753423AbaFQBA3 (ORCPT ); Mon, 16 Jun 2014 21:00:29 -0400 Date: Mon, 16 Jun 2014 17:59:07 -0700 (PDT) From: Hugh Dickins X-X-Sender: hugh@eggly.anvils To: Naoya Horiguchi cc: Hugh Dickins , Andrew Morton , Christoph Lameter , linux-kernel@vger.kernel.org, linux-mm@kvack.org Subject: Re: [PATCH] hugetlb: fix copy_hugetlb_page_range() to handle migration/hwpoisoned entry In-Reply-To: <20140616195950.GA21801@nhori.bos.redhat.com> Message-ID: References: <1402081620-1247-1-git-send-email-n-horiguchi@ah.jp.nec.com> <20140616195950.GA21801@nhori.bos.redhat.com> User-Agent: Alpine 2.11 (LSU 23 2013-08-11) MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Mon, 16 Jun 2014, Naoya Horiguchi wrote: > On Sun, Jun 15, 2014 at 05:19:29PM -0700, Hugh Dickins wrote: > > > > Hold on, that restriction of hugepage migration was marked for stable > > 3.12+, whereas this is marked for stable 2.6.36+ (a glance at my old > > trees suggests 2.6.37+, but you may know better - perhaps hugepage > > migration got backported to 2.6.36-stable, though hardly seems > > stable material). > > Sorry, I misinterpreted one thing. > I thought hugepage migration was merged at 2.6.36 because git-describe > shows v2.6.36-rc7-73-g290408d4a2 for commit 290408d4a2 "hugetlb: hugepage > migration core." But actually that's merged at commit f1ebdd60cc, or > v2.6.36-5792-gf1ebdd60cc73. So this is 2.6.37 stuff. > > Originally hugepage migration was used only for soft offlining in > mm/memory-failure.c which is available only in x86_64, so we implicitly > assumed that hugepage migration was restricted to x86_64. > At 3.12, hugepage migration became available for numa APIs like mbind(), > which are used for other architectures, so the restriction with > hugepage_migration_supported() became necessary since then. > This is the reason why the disablement was marked for 3.12+. > This patch are helpful before extension in 3.12, so it should be marked 2.6.37+. That all makes sense to me now: thanks a lot for explaining the history. Hugh