From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752519Ab1AUJ6Q (ORCPT ); Fri, 21 Jan 2011 04:58:16 -0500 Received: from TYO201.gate.nec.co.jp ([202.32.8.193]:40298 "EHLO tyo201.gate.nec.co.jp" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750955Ab1AUJ6P (ORCPT ); Fri, 21 Jan 2011 04:58:15 -0500 Date: Fri, 21 Jan 2011 18:56:04 +0900 From: Naoya Horiguchi To: Jin Dongming Cc: Andi Kleen , Huang Ying , Hidetoshi Seto , LKLM Subject: Re: [PATCH 3/3 -v2] Fix the wrong corrupted page information of hugetlb page. Message-ID: <20110121095604.GC8897@spritzera.linux.bs1.fc.nec.co.jp> References: <4D38E04A.8050402@np.css.fujitsu.com> <20110121051251.GA16406@spritzera.linux.bs1.fc.nec.co.jp> <4D394120.3020509@np.css.fujitsu.com> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-2022-jp Content-Disposition: inline In-Reply-To: <4D394120.3020509@np.css.fujitsu.com> User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, Jan 21, 2011 at 05:17:36PM +0900, Jin Dongming wrote: > When more than on tail page of the same mapped hugetlb page are > poisoned at the same time, the sum of corrupted page should be > the size of one hugetlb page(512 pages). But HardwareCorrupted > information in /proc/meminfo is 1024 pages or more. > > That is because when the above condition happened, > __memory_failure() will be called many times and pages in hugetlb > page are accounted more than one time. > > This patch fixed it by incrementing the number of corrupted page > one by one. > > -v2 fix checking for "just unpoisoned". > > Signed-off-by: Jin Dongming > Reviewed-by: Hidetoshi Seto Reviewed-by: Naoya Horiguchi