Re: [PATCH] mm, page_alloc: Use static global work_struct for draining per-cpu pages

[Tejun Heo]

Hello,

On Thu, Jan 26, 2017 at 10:47:32AM +0000, Mel Gorman wrote:
> On Wed, Jan 25, 2017 at 04:08:02PM -0800, Andrew Morton wrote:
> > > +	for_each_cpu(cpu, &cpus_with_pcps) {
> > > +		struct work_struct *work = per_cpu_ptr(&pcpu_drain, cpu);
> > > +		INIT_WORK(work, drain_local_pages_wq);
> > 
> > It's strange to repeatedly run INIT_WORK() in this fashion. 
> > Overwriting an atomic_t which should already be zero, initializing a
> > list_head which should already be in the initialized state...
> > 
> > Can we instead do this a single time in init code?
> > 
> 
> INIT_WORK does different things depending on whether LOCKDEP is enabled or
> not and also whether object debugging is enabled. I'd worry that it's not
> functionally equivalent or some future change would break the assumptions
> about what INIT_WORK does internally. The init cost is there, but it's
> insignicant in comparison to the whole workqueue operation or the old
> cost of sending IPIs for that matter.

Both initing once or per each invocation are perfectly valid and
guaranteed to work.  idk, I don't have a strong opinion hereag.

Thanks.

-- 
tejun