From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from bombadil.infradead.org ([65.50.211.133]:46086 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751921AbdDLKaQ (ORCPT ); Wed, 12 Apr 2017 06:30:16 -0400 Received: from clnet-p099-196.ikbnet.co.at ([83.175.99.196] helo=localhost) by bombadil.infradead.org with esmtpsa (Exim 4.87 #1 (Red Hat Linux)) id 1cyFXU-0004j7-1H for linux-xfs@vger.kernel.org; Wed, 12 Apr 2017 10:30:16 +0000 From: Christoph Hellwig Subject: split the reflink remap from the block allocation path V3 Date: Wed, 12 Apr 2017 12:30:07 +0200 Message-Id: <20170412103013.9355-1-hch@lst.de> Sender: linux-xfs-owner@vger.kernel.org List-ID: List-Id: xfs To: linux-xfs@vger.kernel.org We've run into various problems due to the fact that the reflink remap code reuses the xfs_bmapi_write codepath for setting up the extent list entries and abuses the firstblock field for that purpose. This series fixes this by creating an entirely separate xfs_bmapi_remap path that is much simpler than xfs_bmapi_write and does not need to overload the firstblock field. Changes since V2: - add a missing call to xfs_iread_extents Changes since V1: - warn when encountering invalid forks in bmap item recovery - tidy up the xfs_bmap_add_extent_hole_real declaration - trivial indentation change in xfs_bmapi_remap