On Sat, 22 Feb 2014, Tejun Heo wrote: > On Sat, Feb 22, 2014 at 10:14:48AM -0500, Alan Stern wrote: > > Is the cancel_delayed_work_sync(&hub->init_work) call in hub_quiesce() > > going to get confused by all this? > > Yeah, you can't cancel a work item which hasn't been initialzed. > Maybe move init of the first work function there? I don't think it > really matters tho. > > > It's worth mentioning that the only reason for the hub_init_func3 stuff > > is, as the comment says, to avoid a long sleep (100 ms) inside a work > > routine. With all the changes to the work queue infrastructure, maybe > > this doesn't matter so much any more. If we got rid of it then there > > wouldn't be any multiplexing, and this whole issue would become moot. > > I don't really think that'd be necessary. Just sleeping synchronously > should be fine. How many threads are we talking about? One thread per hub (no more than 10 on a typical system). The code in question is part of the hub driver's probe sequence. On Sat, 22 Feb 2014, Peter Hurley wrote: > If a running hub init does not need to be single-threaded wrt > a different running hub init, I'm not quite sure what that means, but the hub init threads are indeed independent of each other. > then a single init work could be queued to > the system_unbound_wq which doesn't care about running times. This sort of thing sounds like the best approach. Tejun, do you want to rewrite the patch, getting rid of the hub_init_func3 and HUB_INIT3 business entirely? Or would you like me to do it? Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
