From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: with ECARTIS (v1.0.0; list xfs); Wed, 22 Oct 2008 14:08:57 -0700 (PDT) Received: from cuda.sgi.com (cuda1.sgi.com [192.48.168.28]) by oss.sgi.com (8.12.11.20060308/8.12.11/SuSE Linux 0.7) with ESMTP id m9ML8nUN020197 for ; Wed, 22 Oct 2008 14:08:49 -0700 Received: from ipmail01.adl6.internode.on.net (localhost [127.0.0.1]) by cuda.sgi.com (Spam Firewall) with ESMTP id 182331216714 for ; Wed, 22 Oct 2008 14:10:32 -0700 (PDT) Received: from ipmail01.adl6.internode.on.net (ipmail01.adl6.internode.on.net [203.16.214.146]) by cuda.sgi.com with ESMTP id QM45hcLP3HI8uC94 for ; Wed, 22 Oct 2008 14:10:32 -0700 (PDT) Date: Thu, 23 Oct 2008 08:10:28 +1100 From: Dave Chinner Subject: Re: BUG: sleeping function called from invalid context at kernel/rwsem.c:131 XFS? (was: Re: linux-next: Tree for October 17) Message-ID: <20081022211028.GO18495@disturbed> References: <20081017203710.GA27187@infradead.org> <20081017135510.7127c4e7@infradead.org> <20081020163327.GA15651@infradead.org> <20081020223549.GA21152@disturbed> <20081022075838.GK18495@disturbed> <20081022082550.GM18495@disturbed> <20081022101351.GB11313@infradead.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20081022101351.GB11313@infradead.org> Sender: xfs-bounce@oss.sgi.com Errors-to: xfs-bounce@oss.sgi.com List-Id: xfs To: Christoph Hellwig Cc: Alexander Beregalov , lachlan@sgi.com, Arjan van de Ven , xfs@oss.sgi.com, linux-next@vger.kernel.org, LKML On Wed, Oct 22, 2008 at 06:13:51AM -0400, Christoph Hellwig wrote: > On Wed, Oct 22, 2008 at 07:25:50PM +1100, Dave Chinner wrote: > > > Basically the above commit moved xfs_ilock() inside > > > radix_tree_preload()/radix_tree_preload_end(), which means we are > > > taking a rwsem() while we have an elevated preempt count. I'll > > > get a patch out to fix it. > > This really needs a warning. Then again I don't really understand this > as the point of radix_tree_preload was that we can do the actual > radix-tree under a lock, or not? Right - the preload allows us to do GFP_KERNEL allocations for radix tree nodes and use a spinlock for inserts into the tree. We could drop the preload stuff if we initialised the radix tree to use GFP_ATOMIC allocations for radix tree nodes, but that is more likely to lead to insert failures under low memory conditions compared to the preload method. Cheers, Dave. -- Dave Chinner david@fromorbit.com