From: Christoph Hellwig Date: Tue, 25 Jun 2019 11:57:07 +0200 Subject: [Cluster-devel] [PATCH] fs: Move mark_inode_dirty out of __generic_write_end In-Reply-To: <20190624182243.22447-1-agruenba@redhat.com> References: <20190618144716.8133-1-agruenba@redhat.com> <20190624065408.GA3565@lst.de> <20190624182243.22447-1-agruenba@redhat.com> Message-ID: <20190625095707.GA1462@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 Mon, Jun 24, 2019 at 08:22:43PM +0200, Andreas Gruenbacher wrote: > That would work, but I don't like how this leaves us with a vfs function > that updates i_size without bothering to dirty the inode very much. This isn't a VFS function, it is a helper library. > > How about if we move the __generic_write_end call into the page_done > callback and leave special handling to the filesystem code if needed > instead? The below patch seems to work for gfs2. That seems way more complicated. I'd much rather go with something like may patch plus maybe a big fat comment explaining that persisting the size update is the file systems job. Note that a lot of the modern file systems don't use the VFS inode tracking for that, besides XFS that includes at least btrfs and ocfs2 as well.