Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S948397AbcKAAxD (ORCPT ); Mon, 31 Oct 2016 20:53:03 -0400 Received: from mail-pf0-f193.google.com ([209.85.192.193]:33529 "EHLO mail-pf0-f193.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S948334AbcKAAxC (ORCPT ); Mon, 31 Oct 2016 20:53:02 -0400 From: Naoya Horiguchi To: Andrew Morton Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org, Naoya Horiguchi , Naoya Horiguchi Subject: [PATCH v1] mm: hwpoison: fix thp split handling in memory_failure() Date: Tue, 1 Nov 2016 09:52:57 +0900 Message-Id: <1477961577-7183-1-git-send-email-n-horiguchi@ah.jp.nec.com> X-Mailer: git-send-email 2.7.0 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1874 Lines: 51 When memory_failure() runs on a thp tail page after pmd is split, we trigger the following VM_BUG_ON_PAGE(): [ 619.550520] page:ffffd7cd819b0040 count:0 mapcount:0 mapping: (null) index:0x1 [ 619.555486] flags: 0x1fffc000400000(hwpoison) [ 619.556408] page dumped because: VM_BUG_ON_PAGE(!page_count(p)) [ 619.558998] ------------[ cut here ]------------ [ 619.561388] kernel BUG at /src/linux-dev/mm/memory-failure.c:1132! memory_failure() passed refcount and page lock from tail page to head page, which is not needed because we can pass any subpage to split_huge_page(). Signed-off-by: Naoya Horiguchi Fixes: 61f5d698cc97 ("mm: re-enable THP") Cc: stable@vger.kernel.org # 4.5+ --- mm/memory-failure.c | 12 +++++------- 1 file changed, 5 insertions(+), 7 deletions(-) diff --git v4.8-rc8-mmotm-2016-09-27-16-08/mm/memory-failure.c v4.8-rc8-mmotm-2016-09-27-16-08_patched/mm/memory-failure.c index de88f33..19e796d 100644 --- v4.8-rc8-mmotm-2016-09-27-16-08/mm/memory-failure.c +++ v4.8-rc8-mmotm-2016-09-27-16-08_patched/mm/memory-failure.c @@ -1112,10 +1112,10 @@ int memory_failure(unsigned long pfn, int trapno, int flags) } if (!PageHuge(p) && PageTransHuge(hpage)) { - lock_page(hpage); - if (!PageAnon(hpage) || unlikely(split_huge_page(hpage))) { - unlock_page(hpage); - if (!PageAnon(hpage)) + lock_page(p); + if (!PageAnon(p) || unlikely(split_huge_page(p))) { + unlock_page(p); + if (!PageAnon(p)) pr_err("Memory failure: %#lx: non anonymous thp\n", pfn); else @@ -1126,9 +1126,7 @@ int memory_failure(unsigned long pfn, int trapno, int flags) put_hwpoison_page(p); return -EBUSY; } - unlock_page(hpage); - get_hwpoison_page(p); - put_hwpoison_page(hpage); + unlock_page(p); VM_BUG_ON_PAGE(!page_count(p), p); hpage = compound_head(p); } -- 2.7.0