Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S965482AbdCXSYl (ORCPT ); Fri, 24 Mar 2017 14:24:41 -0400 Received: from mga05.intel.com ([192.55.52.43]:39038 "EHLO mga05.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S935777AbdCXSXo (ORCPT ); Fri, 24 Mar 2017 14:23:44 -0400 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.36,216,1486454400"; d="scan'208";a="64644356" Message-ID: <1490379805.2733.133.camel@linux.intel.com> Subject: Re: [PATCH v4 01/11] mm: x86: move _PAGE_SWP_SOFT_DIRTY from bit 7 to bit 1 From: Tim Chen To: Zi Yan , linux-kernel@vger.kernel.org, linux-mm@kvack.org Cc: kirill.shutemov@linux.intel.com, akpm@linux-foundation.org, minchan@kernel.org, vbabka@suse.cz, mgorman@techsingularity.net, mhocko@kernel.org, n-horiguchi@ah.jp.nec.com, khandual@linux.vnet.ibm.com, zi.yan@cs.rutgers.edu, dnellans@nvidia.com Date: Fri, 24 Mar 2017 11:23:25 -0700 In-Reply-To: <20170313154507.3647-2-zi.yan@sent.com> References: <20170313154507.3647-1-zi.yan@sent.com> <20170313154507.3647-2-zi.yan@sent.com> Content-Type: text/plain; charset="UTF-8" X-Mailer: Evolution 3.18.5.2 (3.18.5.2-1.fc23) Mime-Version: 1.0 Content-Transfer-Encoding: 8bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2417 Lines: 50 On Mon, 2017-03-13 at 11:44 -0400, Zi Yan wrote: > From: Naoya Horiguchi > > pmd_present() checks _PAGE_PSE along with _PAGE_PRESENT to avoid > false negative return when it races with thp spilt > (during which _PAGE_PRESENT is temporary cleared.) I don't think that > dropping _PAGE_PSE check in pmd_present() works well because it can > hurt optimization of tlb handling in thp split. > In the current kernel, bits 1-4 are not used in non-present format > since commit 00839ee3b299 ("x86/mm: Move swap offset/type up in PTE to > work around erratum"). So let's move _PAGE_SWP_SOFT_DIRTY to bit 1. > Bit 7 is used as reserved (always clear), so please don't use it for > other purpose. > > Signed-off-by: Naoya Horiguchi > Signed-off-by: Zi Yan > --- >  arch/x86/include/asm/pgtable_64.h    | 12 +++++++++--- >  arch/x86/include/asm/pgtable_types.h | 10 +++++----- >  2 files changed, 14 insertions(+), 8 deletions(-) > > diff --git a/arch/x86/include/asm/pgtable_64.h b/arch/x86/include/asm/pgtable_64.h > index 73c7ccc38912..a5c4fc62e078 100644 > --- a/arch/x86/include/asm/pgtable_64.h > +++ b/arch/x86/include/asm/pgtable_64.h > @@ -157,15 +157,21 @@ static inline int pgd_large(pgd_t pgd) { return 0; } >  /* >   * Encode and de-code a swap entry >   * > - * |     ...            | 11| 10|  9|8|7|6|5| 4| 3|2|1|0| <- bit number > - * |     ...            |SW3|SW2|SW1|G|L|D|A|CD|WT|U|W|P| <- bit names > - * | OFFSET (14->63) | TYPE (9-13)  |0|X|X|X| X| X|X|X|0| <- swp entry > + * |     ...            | 11| 10|  9|8|7|6|5| 4| 3|2| 1|0| <- bit number > + * |     ...            |SW3|SW2|SW1|G|L|D|A|CD|WT|U| W|P| <- bit names > + * | OFFSET (14->63) | TYPE (9-13)  |0|0|X|X| X| X|X|SD|0| <- swp entry >   * >   * G (8) is aliased and used as a PROT_NONE indicator for >   * !present ptes.  We need to start storing swap entries above >   * there.  We also need to avoid using A and D because of an >   * erratum where they can be incorrectly set by hardware on >   * non-present PTEs. > + * > + * SD (1) in swp entry is used to store soft dirty bit, which helps us > + * remember soft dirty over page migration > + * > + * Bit 7 in swp entry should be 0 because pmd_present checks not only P, > + * but also G. but also L and G.