From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ingo Molnar Subject: Re: [LOCKDEP] xfs: possible recursive locking detected Date: Tue, 4 Jul 2006 11:12:47 +0200 Message-ID: <20060704091247.GA15982@elte.hu> References: <20060704004116.GA7612@martell.zuzino.mipt.ru> <20060704011858.GG1605@parisc-linux.org> <20060704112503.H1495869@wobbly.melbourne.sgi.com> <20060704063225.GA2752@elte.hu> <20060704084143.GA12931@elte.hu> <20060704191100.C1497438@wobbly.melbourne.sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Alexey Dobriyan , Matthew Wilcox , linux-fsdevel@vger.kernel.org, xfs@oss.sgi.com, Arjan van de Ven Return-path: Received: from mx3.mail.elte.hu ([157.181.1.138]:62664 "EHLO mx3.mail.elte.hu") by vger.kernel.org with ESMTP id S1751228AbWGDJR0 (ORCPT ); Tue, 4 Jul 2006 05:17:26 -0400 To: Nathan Scott Content-Disposition: inline In-Reply-To: <20060704191100.C1497438@wobbly.melbourne.sgi.com> Sender: linux-fsdevel-owner@vger.kernel.org List-Id: linux-fsdevel.vger.kernel.org * Nathan Scott wrote: > > i'd really suggest to clean this up and to convert: > > xfs_ilock(ip, XFS_ILOCK_EXCL); > > to: > > down_write(&ip->i_lock); > > and: > > xfs_iunlock(ip, XFS_ILOCK_EXCL); > > to: > > up_write(&ip->i_lock); > > and eliminate all those layers of indirection. > > That would be good, but it doesn't work for all situations > unfortunately, and it would loose that debug-kernel sanity checking > that we have in there which validates ilock/iolock ordering rules. do you have anything in there that spinlock/mutex debugging or lockdep does not catch? If yes then i'll add it to the generic lock debugging code. Ingo