Messages in this thread | | | Subject | Re: workqueue thing | From | Peter Zijlstra <> | Date | Tue, 22 Dec 2009 12:10:20 +0100 |
| |
On Tue, 2009-12-22 at 09:00 +0900, Tejun Heo wrote: > > Yeah, sure, it's not suited for offloading CPU intensive workloads > (checksumming and encryption basically). Workqueue interface was > never meant to be used by them - it has strong cpu affinity. We need > something which is more parallelism aware for those.
Right, so what about cleaning that up first, then looking how many workqueues can be removed by converting to threaded interrupts and then maybe look again at some of your async things?
As to the SCSI/ATA error, have to wait for a year on hardware threads, why not simply use the existing work to create a one-time (non-affine) thread for that? Its not like it'll ever happen much.
| |