Re: [PATCH] mm/secretmem: Avoid letting secretmem_users drop to zero

From: Dmitry Vyukov
Date: Thu Oct 21 2021 - 11:49:13 EST


On Thu, 21 Oct 2021 at 17:40, Kees Cook <keescook@xxxxxxxxxxxx> wrote:
>
> Quoting Dmitry: "refcount_inc() needs to be done before fd_install().
> After fd_install() finishes, the fd can be used by userspace and we can
> have secret data in memory before the refcount_inc().
>
> A straightforward mis-use where a user will predict the returned fd
> in another thread before the syscall returns and will use it to store
> secret data is somewhat dubious because such a user just shoots themself
> in the foot.
>
> But a more interesting mis-use would be to close the predicted fd and
> decrement the refcount before the corresponding refcount_inc, this way
> one can briefly drop the refcount to zero while there are other users
> of secretmem."
>
> Move fd_install() after refcount_inc().
>
> Cc: Mike Rapoport <rppt@xxxxxxxxxx>
> Cc: Jordy Zomer <jordy@pwning.systems>
> Cc: linux-mm@xxxxxxxxx
> Reported-by: Dmitry Vyukov <dvyukov@xxxxxxxxxx>
> Link: https://lore.kernel.org/lkml/CACT4Y+b1sW6-Hkn8HQYw_SsT7X3tp-CJNh2ci0wG3ZnQz9jjig@xxxxxxxxxxxxxx
> Fixes: 9a436f8ff631 ("PM: hibernate: disable when there are active secretmem users")
> Signed-off-by: Kees Cook <keescook@xxxxxxxxxxxx>

Reviewed-by: Dmitry Vyukov <dvyukov@xxxxxxxxxx>

> ---
> mm/secretmem.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/mm/secretmem.c b/mm/secretmem.c
> index 1fea68b8d5a6..924d84ba481f 100644
> --- a/mm/secretmem.c
> +++ b/mm/secretmem.c
> @@ -217,8 +217,8 @@ SYSCALL_DEFINE1(memfd_secret, unsigned int, flags)
>
> file->f_flags |= O_LARGEFILE;
>
> - fd_install(fd, file);
> refcount_inc(&secretmem_users);
> + fd_install(fd, file);
> return fd;
>
> err_put_fd:
> --
> 2.30.2
>