Re: [PATCH 07/11] pidns: Wait in zap_pid_ns_processes until pid_ns->nr_hashed == 1

[Oleg Nesterov <oleg-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>]

On 12/21, Oleg Nesterov wrote:
>
> And in fact I think this is not strange, but simply wrong.
>
> Please consider the XXX case above. Suppose that free_pid(P) happens
> after ns->child_reaper exits and thus this pointer points to nowhere.
> Suppose also that there is another injected pid so nr_hashed == 2.
> In this case wake_up_process(ns->child_reaper) means use-after-free,
> no?

Hmm. And another minor problem, unless I missed something.

Once again, the parent namespace injects the task T after ns->reaper
sees nr_hashed == 1 and returns. Suppose that reaper's parent does
do_wait() and free_pidmap() clears the bit == 1.

Now, what if T doesn't exit but forks? We must not re-create the
task with pid_nr == 1 in the dead namespace. Normally this can't
happen, RESERVED_PIDS logic in alloc_pidmap() saves us. But it
seems that we need

	- .extra1 = &zero,
	+ .extra1 = &one,

in pid_ns_ctl_table.

Oleg.