From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932481AbXCFMp5 (ORCPT ); Tue, 6 Mar 2007 07:45:57 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S933243AbXCFMp5 (ORCPT ); Tue, 6 Mar 2007 07:45:57 -0500 Received: from village.ehouse.ru ([193.111.92.18]:1305 "EHLO mail.ehouse.ru" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932481AbXCFMp4 (ORCPT ); Tue, 6 Mar 2007 07:45:56 -0500 From: "Alexander Y. Fomichev" Reply-To: "Alexander Y. Fomichev" To: Bill Irwin Subject: Re: kernel BUG at arch/x86_64/mm/../../i386/mm/hugetlbpage.c:140! Date: Tue, 6 Mar 2007 15:45:23 +0300 User-Agent: KMail/1.9.6 Cc: Linux Kernel Mailing List , admin@list.net.ru References: <200703021651.16050.gluk@php4.ru> <20070302214111.GL10643@holomorphy.com> In-Reply-To: <20070302214111.GL10643@holomorphy.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200703061545.23470.gluk@php4.ru> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Saturday 03 March 2007, Bill Irwin wrote: > On Fri, Mar 02, 2007 at 04:51:15PM +0300, Alexander Y. Fomichev wrote: > > I'm hit a bug on 2.6.21-rc1 at startup of mysql with 'large-pages' flag > > set. (at this point mysql trying to allocate pages from hugetlb pool by > > sysv shm syscalls). Seems like it could be triggered by previous badness > > and probably hugetlb itself is not related. Anyway i couldn't reproduce > > it by now with 2.6.21-rc2 git commit > > 562aa1d4c6a874373f9a48ac184f662fbbb06a04. Very likely it has been fixed > > somwhere between 2.6.21-rc1 and -rc2, but i couldn't find something > > related by git log so any comments are welcome. > > If you have a known-working kernel version, git-bisect might help you > track down where it was introduced. Given the messages prior to the > hugetlbpage.c BUG_ON I'd say that this is something else besides the > specific code listed by line number, Nop, seems like it is hugetlb bug fixed by Adam commit 516dffdcd8827a40532798602830dfcfc672294c [PATCH] Fix get_unmapped_area and fsync for hugetlb shm segments I've disregard it thoughtlessly so Adam marks it specific for powerpc/ia64. (hmm.. i donno realy why, huge pages should be alined anyway and if i understend it correctly it may not be if is_file_hugepages() fails) As of "junk" just before the BUG this again IIUC because o-direct staff catches this issue before hugetlb. > though I wouldn't rule out hugetlb > having tripped over itself before the actual BUG. > Yep, it is. > -- wli -- Best regards. Alexander Y. Fomichev Public PGP key: http://sysadminday.org.ru/gluk.asc