From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S264126AbUARWEI (ORCPT ); Sun, 18 Jan 2004 17:04:08 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S264141AbUARWEI (ORCPT ); Sun, 18 Jan 2004 17:04:08 -0500 Received: from dbl.q-ag.de ([213.172.117.3]:6060 "EHLO dbl.q-ag.de") by vger.kernel.org with ESMTP id S264126AbUARWEG (ORCPT ); Sun, 18 Jan 2004 17:04:06 -0500 Message-ID: <400B02CE.5060907@colorfullife.com> Date: Sun, 18 Jan 2004 23:03:58 +0100 From: Manfred Spraul User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.4) Gecko/20030624 X-Accept-Language: en-us, en MIME-Version: 1.0 To: caszonyi@rdslink.ro CC: linux-kernel@vger.kernel.org Subject: Re: Slab coruption and oops with 2.6.1-mm4 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Calin wrote: >Then usual oops but *before that* something which hasn't happened before. >Slab corruption: start=c57c2000, len=4096 >000: 6e 72 6d 71 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b > > Is that reproducable? The selftest code in slab noticed a change in a buffer that is not in use: All values should be 0x6b, but the first 4 bytes are different. Could you try to boot with CONFIG_DEBUG_PAGEALLOC enabled? Perhaps that identifies who corrupts the page. Unfortunately it can't detect DMA writes. >------------ kernel BUG at include/linux/mm.h:275! > put_page_testzero noticed that the page count was already 0. It seems that someone uses memory that was already freed. -- Manfred