> > Turns out that there are many users that give a dynamic string as the > > workqueue name (only the first three are relevant for the problem at > > hand because the others are single-threaded): > > I'm not sure the single threadedness of a workqueue matters here.
Oh I should have explained this, missed it earlier. No, the single-threadedness doesn't really matter for the problem, but the warning Dave got couldn't possibly have triggered for a single-threaded workqueue because it was from the CPU up/down code that spawns/stops the new workqueue thread on a new/dying CPU. johannes
signature.asc
Description: This is a digitally signed message part