Re: [PATCH] afs: proc cells and rootcell are writeable

[Al Viro <viro@ZenIV.linux.org.uk>]

On Tue, Jan 28, 2014 at 08:20:12PM +0000, David Howells wrote:
> Linus Torvalds <torvalds@linux-foundation.org> wrote:
> 
> > > -       p = proc_create("cells", 0, proc_afs, &afs_proc_cells_fops);
> > > +       p = proc_create("cells", S_IFREG | S_IRUGO | S_IWUSR, proc_afs, &afs_proc_cells_fops);
> > > -       p = proc_create("rootcell", 0, proc_afs, &afs_proc_rootcell_fops);
> > > +       p = proc_create("rootcell", S_IFREG | S_IRUGO | S_IWUSR, proc_afs, &afs_proc_rootcell_fops);
> > 
> > So the S_IFREG isn't necessary.
> 
> True.  Is it worth creating proc_create_special() that can create a non-regular
> file and then making proc_create() only permit regular files (and complain if
> the S_IFMT field is not zero)?

We already do: in proc_create_data() we have
        struct proc_dir_entry *pde;
        if ((mode & S_IFMT) == 0)
                mode |= S_IFREG;

        if (!S_ISREG(mode)) {
                WARN_ON(1);     /* use proc_mkdir() */
                return NULL;
        }

proc_mkdir{,_data,_mode} are there for purpose.  Nobody had been insane
enough to put FIFOs or sockets in procfs and anything else would need
additional data anyway.  proc_symlink() is there, proc_mknod() isn't and
nobody has complained yet.  Let's keep it that way, plese...