Re: [PATCH v2] mm: memory-failure: use rcu lock instead of tasklist_lock when collect_procs()

[Matthew Wilcox]

On Sat, Aug 26, 2023 at 09:46:53AM +0800, Tong Tiangen wrote:
> " the ``task_struct`` object is freed only after one or more
> grace periods elapse, with the help of call_rcu(), which is invoked via
> put_task_struct_rcu_user(). "
> 
> Combined with the code,when the task exits：
> 
> release_task()
> 	__exit_signal()
> 		__unhash_process()
> 			list_del_rcu(&p->tasks)
> 	
> 	put_task_struct_rcu_user()
> 		call_rcu(&task->rcu, delayed_put_task_struct);
> 			
> delayed_put_task_struct()
> 	put_task_struct()
> 		if (refcount_sub_and_test(nr, &t->usage))
> 			__put_task_struct()
> 				free_task()
> 	
> The code is consistent with the description in the document.
> 
> According to this understanding, i think for_each_process() under the
> protection of rcu locl is safe, that is, task_struct in the list will not be
> destroyed, and get_task_struct() is also safe.

Aha!  This is different from the usual pattern.  What I'm used to seeing
is:

if (refcount_sub_and_test()) {
	list_del_rcu();
	rcu_free();
}

and then on the read side you need a refcount_inc_not_zero(), which we
didn't have here.  Given this new information you've found, I withdraw
my objection.  It'd be nice to include some of this analysis in an
updated changelog (and maybe improved documentation for tasklist?).