On Wed, Sep 21, 2022 at 04:58:38PM +0200, Miklos Szeredi wrote: > On Wed, 21 Sept 2022 at 11:08, Christian Brauner <brauner@xxxxxxxxxx> wrote: > > > > On Tue, Sep 20, 2022 at 09:36:31PM +0200, Miklos Szeredi wrote: > > > This is in preparation for adding tmpfile support to fuse, which requires > > > that the tmpfile creation and opening are done as a single operation. > > > > > > Replace the 'struct dentry *' argument of i_op->tmpfile with > > > 'struct file *'. > > > > > > Call finish_open_simple() as the last thing in ->tmpfile() instances (may > > > be omitted in the error case). > > > > > > Change d_tmpfile() argument to 'struct file *' as well to make callers more > > > readable. > > > > > > Signed-off-by: Miklos Szeredi <mszeredi@xxxxxxxxxx> > > > --- > > > > Seems fine to me. Fwiw, it feels like all the file->f_path.dentry derefs > > could be wrapped in a helper similar to file_inode(). I know we have > > file_dentry() but that calls d_real() so not sure if that'll be correct > > for all updated callers, > > I don't think file_dentry() should be used for this. > > file_dentry() is basically a hack for overlayfs's "fake path" thing. > It should only be used where strictly necessary. At one point it > would be good to look again at cleaning this mess up. Yeah, that's what I was getting at. The file_dentry() helper would ideally just be as simple as file_inode() and then we'd have file_dentry_real() for the stacking filesystem scenarios.
