From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753633AbaJBNa5 (ORCPT ); Thu, 2 Oct 2014 09:30:57 -0400 Received: from cantor2.suse.de ([195.135.220.15]:35414 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753166AbaJBN3Y (ORCPT ); Thu, 2 Oct 2014 09:29:24 -0400 From: Mel Gorman To: Dave Jones Cc: Linus Torvalds , Hugh Dickins , Al Viro , Rik van Riel , Ingo Molnar , Peter Zijlstra , Aneesh Kumar , Michel Lespinasse , Kirill A Shutemov , Mel Gorman , Linux Kernel Subject: [PATCH 2/4] mm: migrate: Close race between migration completion and mprotect Date: Thu, 2 Oct 2014 14:29:16 +0100 Message-Id: <1412256558-9995-3-git-send-email-mgorman@suse.de> X-Mailer: git-send-email 1.8.4.5 In-Reply-To: <1412256558-9995-1-git-send-email-mgorman@suse.de> References: <1412256558-9995-1-git-send-email-mgorman@suse.de> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org A migration entry is marked as write if pte_write was true at the time the entry was created. The VMA protections are not double checked when migration entries are being removed as mprotect marks write-migration-entries as read. It means that potentially we take a spurious fault to mark PTEs write again but it's straight-forward. However, there is a race between write migrations being marked read and migrations finishing. This potentially allows a PTE to be write that should have been read. Close this race by double checking the VMA permissions when migration completes. Signed-off-by: Mel Gorman --- mm/migrate.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/mm/migrate.c b/mm/migrate.c index f78ec9b..f59b5de 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -146,7 +146,9 @@ static int remove_migration_pte(struct page *new, struct vm_area_struct *vma, pte = pte_mkold(mk_pte(new, vma->vm_page_prot)); if (pte_swp_soft_dirty(*ptep)) pte = pte_mksoft_dirty(pte); - if (is_write_migration_entry(entry)) + + /* Recheck VMA as permissions can change during migration started */ + if (is_write_migration_entry(entry) && (vma->vm_flags & (VM_WRITE))) pte = pte_mkwrite(pte); #ifdef CONFIG_HUGETLB_PAGE if (PageHuge(new)) { -- 1.8.4.5