Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754733AbcKJJMS convert rfc822-to-8bit (ORCPT ); Thu, 10 Nov 2016 04:12:18 -0500 Received: from TYO202.gate.nec.co.jp ([210.143.35.52]:55853 "EHLO tyo202.gate.nec.co.jp" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751664AbcKJJMP (ORCPT ); Thu, 10 Nov 2016 04:12:15 -0500 From: Naoya Horiguchi To: Anshuman Khandual CC: "linux-mm@kvack.org" , "Kirill A. Shutemov" , Hugh Dickins , "Andrew Morton" , Dave Hansen , Andrea Arcangeli , Mel Gorman , Michal Hocko , "Vlastimil Babka" , Pavel Emelyanov , Zi Yan , Balbir Singh , "linux-kernel@vger.kernel.org" , "Naoya Horiguchi" Subject: Re: [PATCH v2 03/12] mm: thp: introduce separate TTU flag for thp freezing Thread-Topic: [PATCH v2 03/12] mm: thp: introduce separate TTU flag for thp freezing Thread-Index: AQHSOU8siXhj9QojUE2xgpJISRcw2qDRUEwAgAAKYAA= Date: Thu, 10 Nov 2016 09:09:05 +0000 Message-ID: <20161110090904.GA9173@hori1.linux.bs1.fc.nec.co.jp> References: <1478561517-4317-1-git-send-email-n-horiguchi@ah.jp.nec.com> <1478561517-4317-4-git-send-email-n-horiguchi@ah.jp.nec.com> <5824307C.7070105@linux.vnet.ibm.com> In-Reply-To: <5824307C.7070105@linux.vnet.ibm.com> Accept-Language: en-US, ja-JP Content-Language: ja-JP X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [10.128.101.28] Content-Type: text/plain; charset="iso-2022-jp" Content-ID: <28695A655618C844BDCCAB546C54D2AD@gisp.nec.co.jp> Content-Transfer-Encoding: 8BIT MIME-Version: 1.0 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1955 Lines: 51 On Thu, Nov 10, 2016 at 02:01:56PM +0530, Anshuman Khandual wrote: > On 11/08/2016 05:01 AM, Naoya Horiguchi wrote: > > TTU_MIGRATION is used to convert pte into migration entry until thp split > > completes. This behavior conflicts with thp migration added later patches, > > Hmm, could you please explain why it conflicts with the PMD based > migration without split ? Why TTU_MIGRATION cannot be used to > freeze/hold on the PMD while it's being migrated ? try_to_unmap() is used both for thp split (via freeze_page()) and page migration (via __unmap_and_move()). In freeze_page(), ttu_flag given for head page is like below (assuming anonymous thp): (TTU_IGNORE_MLOCK | TTU_IGNORE_ACCESS | TTU_RMAP_LOCKED | \ TTU_MIGRATION | TTU_SPLIT_HUGE_PMD) and ttu_flag given for tail pages is: (TTU_IGNORE_MLOCK | TTU_IGNORE_ACCESS | TTU_RMAP_LOCKED | \ TTU_MIGRATION) __unmap_and_move() calls try_to_unmap() with ttu_flag: (TTU_MIGRATION | TTU_IGNORE_MLOCK | TTU_IGNORE_ACCESS) Now I'm trying to insert a branch for thp migration at the top of try_to_unmap_one() like below static int try_to_unmap_one(struct page *page, struct vm_area_struct *vma, unsigned long address, void *arg) { ... if (flags & TTU_MIGRATION) { if (!PageHuge(page) && PageTransCompound(page)) { set_pmd_migration_entry(page, vma, address); goto out; } } , so try_to_unmap() for tail pages called by thp split can go into thp migration code path (which converts *pmd* into migration entry), while the expectation is to freeze thp (which converts *pte* into migration entry.) I detected this failure as a "bad page state" error in a testcase where split_huge_page() is called from queue_pages_pte_range(). Anyway, I'll add this explanation into the patch description in the next post. Thanks, Naoya Horiguchi