Re: [RFC PATCH 3/8] perf: Init/fini PMU specific data

[Liang, Kan]

On 12/2/2019 3:25 PM, Andi Kleen wrote:
> Looks reasonable to me.
>
> > //get current number of threads
> > read_lock(&tasklist_lock);
> > for_each_process_thread(g, p)
> > 	num_thread++;
> > read_unlock(&tasklist_lock);
>
> I'm sure we have that count somewhere.

It looks like we can get the number from global variable "nr_threads"
I will use it in v2.

> > //allocate the space for them
> > for (i = 0; i < num_thread; i++)
> > 	data[i] = kzalloc(ctx_size, flags);
> > i = 0;
> >
> > /*
> >   * Assign the space to tasks
> >   * There may be some new threads created when we allocate space.
> >   * new_task will track its number.
> >   */
> > raw_spin_lock_irqsave(&task_data_events_lock, flags);
> >
> > if (atomic_inc_return(&nr_task_data_events) > 1)
> > 	goto unlock;
> >
> > for_each_process_thread(g, p) {
> > 	if (i < num_thread)
> > 		p->perf_ctx_data = data[i++];
> > 	else
> > 		new_task++;
> > }
> > raw_spin_unlock_irqrestore(&task_data_events_lock, flags);
>
> Is that lock taken in the context switch? >
> If not could be a normal spinlock, thus be more RT friendly.

It's not in context switch. I will use the normal spinlock to instead.

Thanks,
Kan