On Fri, Nov 13, 2020 at 1:28 PM Al Viro <viro@xxxxxxxxxxxxxxxxxx> wrote: > > On Fri, Nov 13, 2020 at 11:52:09AM +0100, Miklos Szeredi wrote: > > > It's the wrong interface, and we'll have to live with it forever if we > > go this route. > > > > Better get the interface right and *then* think about the > > implementation. I don't think adding ->atomic_tmpfile() would be that > > of a big deal, and likely other distributed fs would start using it in > > the future. > > Let me think about it; I'm very unhappy with the amount of surgery it has > taken to somewhat sanitize the results of ->atomic_open() introduction, so > I'd prefer to do it reasonably clean or not at all. The minimal VFS change for fuse to be able to do tmpfile with one request would be to pass open_flags to ->tmpfile(). That way the private data for the open file would need to be temporarily stored in the inode and ->open() would just pick it up from there without sending another request. Not the cleanest, but I really don't care as long as the public interface is the right one. Thanks, Miklos
