Re: [PATCH 2/2] test-ww_mutex: Fix deadlock in test_cycle_work

Bradley Morgan <[email protected]>
Newsgroups org.kernel.vger.linux-kernel
Message-ID <[email protected]>
> +	workqueue_set_max_active(wq, cycle_ncpus + 1);
> +	workqueue_set_min_active(wq, cycle_ncpus + 1);

Works on any topology. min_active is the floor of that clamp, so
every node ends up with N + 1 slots.

One thing... set_max before set_min is load bearing, not style. heh
workqueue_set_min_active() clamps against saved_max_active:

	saved_min_active = clamp(min_active, 0, wq->saved_max_active);

so set_min first would silently cap at the old 1024. The comment
tells the story the other way around, heh. Something like... Feel free to
bikeshed.

	/*
	 * test_cycle_work() needs all N + 1 works running at once.
	 * Raise max first, min_active is clamped to it.
	 */

I mean, this doesn't particularly kill the patch. Please add

Reviewed-by: Bradley Morgan <[email protected]>

Thanks!
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.