Re: [PATCH v2] mm/kmemleak: avoid soft lockup when scanning task stacks

[Catalin Marinas]

Hi Breno,

On Mon, Jun 15, 2026 at 02:27:13AM -0700, Breno Leitao wrote:
> On Fri, Jun 12, 2026 at 06:11:40PM +0100, Catalin Marinas wrote:
> > 	if (kmemleak_stack_scan) {
> > 		struct pid *pid;
> > 		int nr = 1;
> > 
> > 		do {
> > 			struct task_struct *p = NULL;
> > 
> > 			rcu_read_lock();
> > 			pid = find_ge_pid(nr, &init_pid_ns);
> > 			if (pid) {
> > 				nr = pid_nr(pid) + 1;
> > 				p = pid_task(pid, PIDTYPE_PID);
> > 				if (p)
> > 					get_task_struct(p);
> > 			}
> > 			rcu_read_unlock();
> > 
> > 			if (p) {
> > 				void *stack = try_get_task_stack(p);
> > 
> > 				if (stack) {
> > 					scan_block(stack, stack + THREAD_SIZE,
> > 							NULL);
> > 					put_task_stack(p);
> > 				}
> > 				put_task_struct(p);
> > 			}
> 
> Should we add a scan_should_stop() check here to allow early
> termination?

Ah, yes, I think we should, maybe as 'while (pid && !scan_should_stop())'.

scan_block() already bails out early but it will still go quickly
through all the stacks which we don't need. As a later clean-up, I
wonder whether we should change scan_block() to return non-zero and use
that.

-- 
Catalin