[PATCH 1/1] security: avoid a leak in securityfs_remove()

[Paul Moore]

On Thu, Jun 2, 2022 at 3:31 AM Yuanjun Gong <ruc_gongyuanjun at 163.com> wrote:
>
> Delete the dentry in securityfs_remove() to make sure the
> dentry is not used by another thread and live longer than
> the call of securityfs_remove().
>
> Signed-off-by: Yuanjun Gong <ruc_gongyuanjun at 163.com>
> ---
>  security/inode.c | 1 +
>  1 file changed, 1 insertion(+)
>
> diff --git a/security/inode.c b/security/inode.c
> index 6c326939750d..606f390d21d2 100644
> --- a/security/inode.c
> +++ b/security/inode.c
> @@ -306,6 +306,7 @@ void securityfs_remove(struct dentry *dentry)
>                         simple_rmdir(dir, dentry);
>                 else
>                         simple_unlink(dir, dentry);
> +               d_delete(dentry);
>                 dput(dentry);

Doesn't the dput() call take care of this?

Do you have a reproducer where you can show that we are leaking dentries?

>         }
>         inode_unlock(dir);
> --
> 2.17.1

-- 
paul-moore.com