Messages in this thread | | | Date | Tue, 22 Jan 2013 17:07:58 -0800 | From | Tejun Heo <> | Subject | Re: [PATCHSET] workqueue: remove gcwq and make worker_pool the only backend abstraction |
| |
Hey.
On Tue, Jan 22, 2013 at 02:37:02PM +0900, Joonsoo Kim wrote: > > It seems like we'll need to support worker pools with custom > > attributes, which is planned to be implemented as extra worker_pools > > for the unbound CPU. Removing gcwq and having worker_pool as the top > > level abstraction makes things much simpler for such designs. Also, > > there's scalability benefit to not sharing locking and busy hash among > > different worker pools as worker pools w/ custom attributes are likely > > to have widely different memory / cpu locality characteristics. > > Could you tell me why extra worker_pools with custom attributes are needed? > Or could you give a reference link for this?
Currently, there are two expected users - writeback and crypto. The former currently implements its own worker pool and the latter is using per-cpu workqueue but not particularly happy with it. Being bound to the issuing CPU seems a bit too limiting.
Thanks.
-- tejun
| |