On Mon, 2024-12-23 at 20:05 +0000, Al Viro wrote:
> On Mon, Dec 23, 2024 at 02:52:12PM -0500, James Bottomley wrote:
> >
> > +static int efivarfs_file_release(struct inode *inode, struct file
> > *file)
> > +{
> > + inode_lock(inode);
> > + if (i_size_read(inode) == 0 && !d_unhashed(file-
> > >f_path.dentry)) {
> > + drop_nlink(inode);
> > + d_delete(file->f_path.dentry);
> > + dput(file->f_path.dentry);
> > + }
> > + inode_unlock(inode);
> > + return 0;
> > +}
>
> This is wrong; so's existing logics for removal from write(). Think
> what happens if you open the sucker, have something bound on top of
> it and do that deleting write().
Shouldn't the bind have taken a dentry reference? in which case we'll
just drop the dentry but it won't be the final put, so it will still
hang around.
> Let me look into that area...
Thanks; as you say, delete from write has been around for over a decade
in this filesystem. We can defer the delete, but it has to happen
somewhere if a write causes an EFI variable to be removed.
Regards,
James
