Re: [PATCH] workqueue: flush all pending jobs in destroy_workqueue()

[Alice Ryhl <aliceryhl@google.com>]

On Fri, Apr 25, 2025 at 09:25:57AM -1000, Tejun Heo wrote:
> Hello,
> 
> On Fri, Apr 25, 2025 at 09:33:54AM +0000, Alice Ryhl wrote:
> ...
> > Hmm. I think we would need to add a new field to delayed_work to keep
> > track of which list it has been added to.
> 
> Can't we use the same cpu that's already recorded in delayed_work->cpu?

Isn't that just going to be equal to WORK_CPU_UNBOUND most of the time?

Though I guess we could use the values NR_CPUS .. 2*NR_CPUS-1 to
remember which list is used when it is unbound.

> > Another option could be to add a boolean that disables the list. After
> > all, we never call destroy_workqueue() on system_wq so we don't need the
> > list for that workqueue.
> 
> It's not just system_wq tho. Any busy workqueue can hit scalability problems
> and the result would be usually subtle performance penalties. If we can keep
> it cheap enough, I'd prefer the behavior uniform across all workqueues.

Yeah ... that does make sense.

Alice