Re: [PATCH 0/4] Workqueue: rename system workqueue and add WQ_PERCPU

[Michal Hocko]

On Mon 05-05-25 08:51:21, Sebastian Andrzej Siewior wrote:
> On 2025-05-03 10:28:30 [+0200], Marco Crivellari wrote:
> > Hi!
> Hi,
> 
> > This series is the follow up of the discussion from:
> > 	"workqueue: Always use wq_select_unbound_cpu() for WORK_CPU_UNBOUND."
> > 	https://lore.kernel.org/all/20250221112003.1dSuoGyc@xxxxxxxxxxxxx/
> > 
> > 1)  [P 1-2] system workqueue rename:
> >  
> > 	system_wq is a per-CPU workqueue, but his name is not clear.
> > 	system_unbound_wq is to be used when locality is not required.
> > 	
> > 	system_wq renamed in system_percpu_wq, while system_unbound_wq
> > 	became system_dfl_wq.
> > 
> > 2)  [P 3] Introduction of WQ_PERCPU.
> > 
> > 	This patch adds a new WQ_PERCPU flag to explicitly request the legacy
> > 	per-CPU behavior. WQ_UNBOUND will be removed once the migration is
> > 	complete.
> > 
> > 	Every alloc_workqueue() caller should use one among WQ_PERCPU or
> > 	WQ_UNBOUND. This is actually enforced warning if both or none of them
> > 	are present at the same time.
> > 
> > 3)  [P 4] alloc_workqueue() callee should pass explicitly WQ_PERCPU.
> > 
> > 	This patch ensures that every caller that needs per-cpu workqueue
> > 	will explicitly require it, using the WQ_PERCPU flag.
> 
> Sounds like a plan.
> I assume the huge patches were made with coccinelle?

Yes, this makes a lot of sense. I think it is worth mentioning why do we
want/need to go through this major refactoring. From my POV this will
help cpu isolation in a long term because it reduces unpredictable
interference from pcp workers. 
-- 
Michal Hocko
SUSE Labs