From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1759108Ab3K1Hsa (ORCPT ); Thu, 28 Nov 2013 02:48:30 -0500 Received: from LGEMRELSE1Q.lge.com ([156.147.1.111]:59593 "EHLO LGEMRELSE1Q.lge.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754031Ab3K1Hql (ORCPT ); Thu, 28 Nov 2013 02:46:41 -0500 X-AuditID: 9c93016f-b7b6aae000005fae-d2-5296f4dfab77 From: Joonsoo Kim To: Andrew Morton Cc: Mel Gorman , Hugh Dickins , Rik van Riel , Ingo Molnar , Naoya Horiguchi , Hillf Danton , linux-kernel@vger.kernel.org, linux-mm@kvack.org, Joonsoo Kim , Joonsoo Kim Subject: [PATCH 3/9] mm/rmap: factor lock function out of rmap_walk_anon() Date: Thu, 28 Nov 2013 16:48:40 +0900 Message-Id: <1385624926-28883-4-git-send-email-iamjoonsoo.kim@lge.com> X-Mailer: git-send-email 1.7.9.5 In-Reply-To: <1385624926-28883-1-git-send-email-iamjoonsoo.kim@lge.com> References: <1385624926-28883-1-git-send-email-iamjoonsoo.kim@lge.com> X-Brightmail-Tracker: AAAAAA== Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org When we traverse anon_vma, we need to take a read-side anon_lock. But there is subtle difference in the situation so that we can't use same method to take a lock in each cases. Therefore, we need to make rmap_walk_anon() taking difference lock function. This patch is the first step, factoring lock function for anon_lock out of rmap_walk_anon(). It will be used in case of removing migration entry and in default of rmap_walk_anon(). Signed-off-by: Joonsoo Kim diff --git a/mm/rmap.c b/mm/rmap.c index e6d532c..916f2ed 100644 --- a/mm/rmap.c +++ b/mm/rmap.c @@ -1683,6 +1683,24 @@ void __put_anon_vma(struct anon_vma *anon_vma) } #ifdef CONFIG_MIGRATION +static struct anon_vma *rmap_walk_anon_lock(struct page *page) +{ + struct anon_vma *anon_vma; + + /* + * Note: remove_migration_ptes() cannot use page_lock_anon_vma_read() + * because that depends on page_mapped(); but not all its usages + * are holding mmap_sem. Users without mmap_sem are required to + * take a reference count to prevent the anon_vma disappearing + */ + anon_vma = page_anon_vma(page); + if (!anon_vma) + return NULL; + + anon_vma_lock_read(anon_vma); + return anon_vma; +} + /* * rmap_walk() and its helpers rmap_walk_anon() and rmap_walk_file(): * Called by migrate.c to remove migration ptes, but might be used more later. @@ -1695,16 +1713,10 @@ static int rmap_walk_anon(struct page *page, int (*rmap_one)(struct page *, struct anon_vma_chain *avc; int ret = SWAP_AGAIN; - /* - * Note: remove_migration_ptes() cannot use page_lock_anon_vma_read() - * because that depends on page_mapped(); but not all its usages - * are holding mmap_sem. Users without mmap_sem are required to - * take a reference count to prevent the anon_vma disappearing - */ - anon_vma = page_anon_vma(page); + anon_vma = rmap_walk_anon_lock(page); if (!anon_vma) return ret; - anon_vma_lock_read(anon_vma); + anon_vma_interval_tree_foreach(avc, &anon_vma->rb_root, pgoff, pgoff) { struct vm_area_struct *vma = avc->vma; unsigned long address = vma_address(page, vma); -- 1.7.9.5