From: Christoph Hellwig Subject: Re: [BUG v4.10-rc1] fzero returns EIO on DAX mount Date: Fri, 13 Jan 2017 05:21:31 -0800 Message-ID: <20170113132131.GA19088@infradead.org> References: <20170106103239.GY1859@eguan.usersys.redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: linux-block@vger.kernel.org, chaitanya.kulkarni@hgst.com, linux-xfs@vger.kernel.org, linux-ext4@vger.kernel.org To: Eryu Guan Return-path: Received: from bombadil.infradead.org ([198.137.202.9]:38514 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751763AbdAMNVn (ORCPT ); Fri, 13 Jan 2017 08:21:43 -0500 Content-Disposition: inline In-Reply-To: <20170106103239.GY1859@eguan.usersys.redhat.com> Sender: linux-ext4-owner@vger.kernel.org List-ID: The problem is that brd return -EIO due to alignment, and the old code waited for the bio and then ignored the error. The new code assumes if the submission is fine this is the method to pick and to move on. We could revert some of this for now, but I think the right thing to do is to stop abusing REQ_OP_DISCARD for zeroing now. I had planned to do this for 4.10 but it splitted, let's see if I can get the series small enough for this time in the merge window.