Re: + cgroups-more-safe-tasklist-locking-in-cgroup_attach_proc.patch added to -mm tree

[Oleg Nesterov <oleg@redhat.com>]

> From: Ben Blum <bblum@andrew.cmu.edu>
>
> Fix unstable tasklist locking in cgroup_attach_proc.
>
> According to this thread - https://lkml.org/lkml/2011/7/27/243 - RCU is
> not sufficient to guarantee the tasklist is stable w.r.t.  de_thread and
> exit.  Taking tasklist_lock for reading, instead of rcu_read_lock, ensures
> proper exclusion.

I still think we should avoid the global lock.

In any case, with tasklist or siglock,

> -	rcu_read_lock();
> +	read_lock(&tasklist_lock);
>  	if (!thread_group_leader(leader)) {
>  		/*
>  		 * a race with de_thread from another thread's exec() may strip
> @@ -2036,7 +2036,7 @@ int cgroup_attach_proc(struct cgroup *cg
>  		 * throw this task away and try again (from cgroup_procs_write);
>  		 * this is "double-double-toil-and-trouble-check locking".
>  		 */
> -		rcu_read_unlock();
> +		read_unlock(&tasklist_lock);
>  		retval = -EAGAIN;

this check+comment becomes completely pointless and imho very confusing.

Oleg.