From: Christoph Hellwig Date: Wed, 15 Jan 2020 07:48:27 +0100 Subject: [Cluster-devel] [PATCH 08/12] ext4: hold i_rwsem until AIO completes In-Reply-To: <20200114215023.GH140865@mit.edu> References: <20200114161225.309792-1-hch@lst.de> <20200114161225.309792-9-hch@lst.de> <20200114215023.GH140865@mit.edu> Message-ID: <20200115064827.GA21176@lst.de> List-Id: To: cluster-devel.redhat.com MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit On Tue, Jan 14, 2020 at 04:50:23PM -0500, Theodore Y. Ts'o wrote: > I note that you've dropped the inode_dio_wait() in ext4's ZERO_RANGE, > COLLAPSE_RANGE, INSERT_RANGE, etc. We had added these to avoid > problems when various fallocate operations which modify the inode's > logical->physical block mapping racing with direct I/O (both reads or > writes). > > I don't see a replacement protection in this patch series. How does > are file systems supported to protect against such races? By holding i_rwsem until the direct I/O operations are finished, and not just until they are sumbitted.