Re: [PATCH v3] fork: free vmapped stacks in cache when cpus are offline

[Thomas Gleixner <tglx@linutronix.de>]

On Sat, 11 Feb 2017, Hoeun Ryu wrote:

> On Fri, Feb 10, 2017 at 11:41 PM, Michal Hocko <mhocko@kernel.org> wrote:
> > On Fri 10-02-17 23:31:41, Hoeun Ryu wrote:
> >> On Fri, Feb 10, 2017 at 9:05 PM, Michal Hocko <mhocko@kernel.org> wrote:
> >> > On Fri 10-02-17 17:32:07, Hoeun Ryu wrote:
> > [...]
> >> >>  static unsigned long *alloc_thread_stack_node(struct task_struct *tsk, int node)
> >> >> @@ -456,6 +474,11 @@ void __init fork_init(void)
> >> >>       for (i = 0; i < UCOUNT_COUNTS; i++) {
> >> >>               init_user_ns.ucount_max[i] = max_threads/2;
> >> >>       }
> >> >> +
> >> >> +#ifdef CONFIG_VMAP_STACK
> >> >> +     cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "vm_stack_cache",
> >> >> +                       NULL, free_vm_stack_cache);
> >> >> +#endif
> >> >
> >> > I am not familiar the new hotplug infrastructure so I might be missing
> >> > something. CPUHP_AP_ONLINE_DYN will allocate a state which is has only
> >> > 30 slots available.

That's enough and when we hit that limit we just up it.

> >> > The name also suggests this will be called on an online event.

The states are symmetric.