Re: [PATCH 1/7] work_on_cpu: helper for doing task on a CPU.

[Oleg Nesterov <oleg@tv-sign.ru>]

On 10/23, Rusty Russell wrote:
>
> +long work_on_cpu(unsigned int cpu, long (*fn)(void *), void *arg)
> +{
> +	struct work_for_cpu wfc;
> +
> +	INIT_WORK(&wfc.work, do_work_for_cpu);
> +	init_completion(&wfc.done);
> +	wfc.fn = fn;
> +	wfc.arg = arg;
> +	get_online_cpus();
> +	if (unlikely(!cpu_online(cpu))) {
> +		wfc.ret = -EINVAL;
> +		complete(&wfc.done);
> +	} else
> +		schedule_work_on(cpu, &wfc.work);

I do not claim this is wrong, but imho the code is a bit lisleading and
needs a comment (or the "fix", please see below).

Once we drop cpu_hotplug lock, CPU can go away and this work can migrate
to another cpu.

> +	put_online_cpus();
> +	wait_for_completion(&wfc.done);

Actually you don't need work_for_cpu->done, you can use flush_work().

IOW, I'd suggest

	long work_on_cpu(unsigned int cpu, long (*fn)(void *), void *arg)
	{
		struct work_for_cpu wfc;

		INIT_WORK(&wfc.work, do_work_for_cpu);
		wfc.fn = fn;
		wfc.arg = arg;
		wfc.ret = -EINVAL;

		get_online_cpus();
		if (likely(cpu_online(cpu))) {
			schedule_work_on(cpu, &wfc.work);
			flush_work(&wfc.work);
		}
		put_online_cpus();

		return wfc.ret;
	}

Oleg.