From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from ipmail06.adl6.internode.on.net ([150.101.137.145]:55850 "EHLO ipmail06.adl6.internode.on.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750957AbdB1XNz (ORCPT ); Tue, 28 Feb 2017 18:13:55 -0500 Date: Wed, 1 Mar 2017 09:46:45 +1100 From: Dave Chinner Subject: Re: lockdep splat in today's master in generic/070 Message-ID: <20170228224645.GG17542@dastard> References: <0e8e52ac-36e4-c8f9-99c7-88e2c9bcb782@suse.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <0e8e52ac-36e4-c8f9-99c7-88e2c9bcb782@suse.com> Sender: linux-xfs-owner@vger.kernel.org List-ID: List-Id: xfs To: Nikolay Borisov Cc: linux-xfs@vger.kernel.org On Tue, Feb 28, 2017 at 04:56:01PM +0200, Nikolay Borisov wrote: > Hello, > > I've been running xfstest and I can reliably reproduce the following lockdep splat: vm_map_ram() has hard coded GFP_KERNEL allocations, it's being called in GFP_NOFS context, and lockdep doesn't know about memalloc_noio_save() used by _xfs_buf_map_pages() to work around the problems in vm_map_ram(). See the comment in _xfs_buf_map_pages(). -Dave. -- Dave Chinner david@fromorbit.com