From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from bombadil.infradead.org ([198.137.202.9]:60852 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751798AbcLFNt1 (ORCPT ); Tue, 6 Dec 2016 08:49:27 -0500 Date: Tue, 6 Dec 2016 05:48:47 -0800 From: Christoph Hellwig Subject: Re: [BUG] xfs/109 crashed 2k block size reflink enabled XFS Message-ID: <20161206134847.GA10181@infradead.org> References: <20161205092112.GS29149@eguan.usersys.redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20161205092112.GS29149@eguan.usersys.redhat.com> Sender: linux-xfs-owner@vger.kernel.org List-ID: List-Id: xfs To: Eryu Guan Cc: linux-xfs@vger.kernel.org Hi Eryu, my attempts at reproducing this are still not very reliable, someimes it takes three iterations, sometimes it's doing fine for 100+. Can you try reverting commit 0af32fb46 ("xfs: fix bogus space reservation in xfs_iomap_write_allocate") and give this a spin on your test setup? I wonder if somehow our accounting for the case where we convert from inline extents to the btree format is not correct, and the code before that commit was papering over that fact.