Currently, registering a new binary type pins the binfmt_misc
filesystem. Specifically, this means that as long as there is at least
one binary type registered the binfmt_misc filesystem survives all
umounts, i.e. the superblock is not destroyed. Meaning that a umount
followed by another mount will end up with the same superblock and the
same binary type handlers. This is a behavior we tend to discourage for
any new filesystems (apart from a few special filesystems such as e.g.
configfs or debugfs). A umount operation without the filesystem being
pinned - by e.g. someone holding a file descriptor to an open file -
should usually result in the destruction of the superblock and all
associated resources. This makes introspection easier and leads to
clearly defined, simple and clean semantics. An administrator can rely
on the fact that a umount will guarantee a clean slate making it
possible to reinitialize a filesystem. Right now all binary types would
need to be explicitly deleted before that can happen.

This allows us to remove the heavy-handed calls to simple_pin_fs() and
simple_release_fs() when creating and deleting binary types. This in
turn allows us to replace the current brittle pinning mechanism abusing
dget() which has caused a range of bugs judging from prior fixes in [2]
and [3]. The additional dget() in load_misc_binary() pins the dentry but
only does so for the sake to prevent ->evict_inode() from freeing the
node when a user removes the binary type and kill_node() is run. Which
would mean ->interpreter and ->interp_file would be freed causing a UAF.

This isn't really nicely documented nor is it very clean because it
relies on simple_pin_fs() pinning the filesystem as long as at least one
binary type exists. Otherwise it would cause load_misc_binary() to hold
on to a dentry belonging to a superblock that has been shutdown.
Replace that implicit pinning with a clean and simple per-node refcount
and get rid of the ugly dget() pinning. A similar mechanism exists for
e.g. binderfs (cf. [4]). All the cleanup work can now be done in
->evict_inode().

In a follow-up patch we will make it possible to use binfmt_misc in
sandboxes. We will use the cleaner semantics where a umount for the
filesystem will cause the superblock and all resources to be
deallocated. In preparation for this apply the same semantics to the
initial binfmt_misc mount. Note, that this is a user-visible change and
as such a uapi change but one that we can reasonably risk. We've
discussed this in earlier versions of this patchset (cf. [1]).

The main user and provider of binfmt_misc is systemd. Systemd provides
binfmt_misc via autofs since it is configurable as a kernel module and
is used by a few exotic packages and users. As such a binfmt_misc mount
is triggered when /proc/sys/fs/binfmt_misc is accessed and is only
provided on demand. Other autofs on demand filesystems include EFI ESP
which systemd umounts if the mountpoint stays idle for a certain amount
of time. This doesn't apply to the binfmt_misc autofs mount which isn't
touched once it is mounted meaning this change can't accidently wipe
binary type handlers without someone having explicitly unmounted
binfmt_misc. After speaking to systemd folks they don't expect this
change to affect them.

In line with our general policy, if we see a regression for systemd or
other users with this change we will switch back to the old behavior for
the initial binfmt_misc mount and have binary types pin the filesystem
again. But while we touch this code let's take the chance and let's
improve on the status quo.

[1]: https://lore.kernel.org/r/20191216091220.465626-2-laurent@vivier.eu
[2]: commit 43a4f261 ("exec: binfmt_misc: fix race between load_misc_binary() and kill_node()"
[3]: commit 83f91827 ("exec: binfmt_misc: shift filp_close(interp_file) from kill_node() to bm_evict_inode()")
[4]: commit f0fe2c0f ("binder: prevent UAF for binderfs devices II")

Link: https://lore.kernel.org/r/20211028103114.2849140-1-brauner@kernel.org (v1)
Cc: Sargun Dhillon <sargun@sargun.me>
Cc: Serge Hallyn <serge@hallyn.com>
Cc: Jann Horn <jannh@google.com>
Cc: Henning Schild <henning.schild@siemens.com>
Cc: Andrei Vagin <avagin@gmail.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Laurent Vivier <laurent@vivier.eu>
Cc: linux-fsdevel@vger.kernel.org
Acked-by: Serge Hallyn <serge@hallyn.com>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
---
/* v2 */
- Christian Brauner <christian.brauner@ubuntu.com>:
  - Add more comments that explain what's going on.
  - Rename functions while changing them to better reflect what they are
    doing to make the code easier to understand.
  - In the first version when a specific binary type handler was removed
    either through a write to the entry's file or all binary type
    handlers were removed by a write to the binfmt_misc mount's status
    file all cleanup work happened during inode eviction.
    That includes removal of the relevant entries from entry list. While
    that works fine I disliked that model after thinking about it for a
    bit. Because it means that there was a window were someone has
    already removed a or all binary handlers but they could still be
    safely reached from load_misc_binary() when it has managed to take
    the read_lock() on the entries list while inode eviction was already
    happening. Again, that perfectly benign but it's cleaner to remove
    the binary handler from the list immediately meaning that ones the
    write to then entry's file or the binfmt_misc status file returns
    the binary type cannot be executed anymore. That gives stronger
    guarantees to the user.