On Sat, Nov 09, 2019 at 03:13:33AM +0000, Al Viro wrote: > Does anyone see objections to the following patch? Christoph, that seems to > be your code; am I missing something subtle here? AFAICS, that goes back to > 2007 or so... This goes back to way before that, that series jut factored out proper export operations from the two inode or superblock methods we had before with the rest handled in core code somewhere that made a complete mess of file systems with 64-bit inode numbers. Otherwise this looks fine, although splitting the refactoring from the actual change would make for a much more readable series.