Re: [RFC PATCH] vfs: syscalls: Add create_automount() and remove_automount()

From: Linus Torvalds
Date: Wed Feb 19 2020 - 13:56:17 EST

On Wed, Feb 19, 2020 at 9:55 AM David Howells <dhowells@xxxxxxxxxx> wrote:
> There's a file type beyond file, dir and symlink that AFS supports:
> mountpoint. It appears as a directory with no lookup op in Linux - though it
> does support readlink. When a client walks over it, it causes an automount of
> the volume named by the content of the mountpoint "file" on that point. NFS
> and CIFS have similar things.

Honestly, AFS isn't common or relevant enough to make this worth a new
special system call etc.

Why don't you just use mkdir with S_ISVTX set, or something like that?
Maybe we can even add a new mode bit and allow you to set that one.

And why would removal be any different from rmdir()?

Or just do a perfectly regular mkdir(), followed by a ioctl().

> Directory, not file. You can do mkdir (requiring write and execute), for
> example, in a directory you cannot open (which would require read). If you
> cannot open it, you cannot do ioctl on it.

Honestly, who cares?

Seriously. Just make the rule be that you need read permission on the
directory too in order to do that ioctl() that is your magical "make
special node".

What makes this all *SO* special, and *SO* important that you need to
follow somebody elses rules that absolutely nobody cares about?