Re: CLONE_PARENT after setns(CLONE_NEWPID)

[Oleg Nesterov <oleg@redhat.com>]

On 11/06, Eric W. Biederman wrote:
>
> --- a/kernel/fork.c
> +++ b/kernel/fork.c
> @@ -1174,7 +1174,7 @@ static struct task_struct *copy_process(unsigned long clone_flags,
>          * do not allow it to share a thread group or signal handlers or
>          * parent with the forking task.
>          */
> -       if (clone_flags & (CLONE_SIGHAND | CLONE_PARENT)) {
> +       if (clone_flags & (CLONE_SIGHAND)) {
>                 if ((clone_flags & (CLONE_NEWUSER | CLONE_NEWPID)) ||
>                     (task_active_pid_ns(current) !=
>                                 current->nsproxy->pid_ns_for_children))

OK, agreed. I failed to find any problem with CLONE_PARENT with
CLONE_NEWUSER or after setns. And the main point of 40a0d32d1eaf
was "make them consistent", not "tighten up".

Besides, this doesn't differ too much from setns + fork() && exit(),
the grandchild will have the new namespace and reparented.

Acked-by: Oleg Nesterov <oleg@redhat.com>

> Changing the CLONE_SIGHAND into CLONE_THREAD will need to happen in a
> separate patch.  It isn't stable material, and so far there is no
> compelling use case for it.

Yes. Again, 40a0d32d1eaf chose CLONE_SIGHAND to unify CLONE_NEWUSER/setns
cases, copy_process() used this check. And in fact I voted for CLONE_THREAD
from the very beginning, it was you who suggested to use CLONE_SIGHAND
instead ;) OTOH, it was probably right to not relax the restrictions we
already had.

Oleg.