On Thursday 14 March 2013 12:13:59 Ming Lei wrote: > On Wed, Mar 13, 2013 at 11:15 PM, Oliver Neukum <oneukum@xxxxxxx> wrote: > > On Wednesday 13 March 2013 23:02:32 Ming Lei wrote: > >> On Wed, Mar 13, 2013 at 10:55 PM, Oliver Neukum <oneukum@xxxxxxx> wrote: > >> > > >> > But it will be run when the next read() call is made. The question is what happens > >> > >> When next read() is called, looks it won't be reached if > >> 'ongoing_read' is set, and > >> it needn't run without ongoing URBs. > > > > If ongoing_read is not set, IO may be started and its completion would > > not be waited for, as the IO that had been started, but wasn't waited > > for due to an interrupt, has completed. > > OK, got it, so how about INIT_COMPLETION(&dev->bulk_in_completion) > in case of !ongoing_read? That means replacing 'if (!dev->processed_urb)' > with INIT_COMPLETION(&dev->bulk_in_completion). It is possible, if you also change locking in skel_read_bulk_callback() and skel_read(). But then it makes no sense and you'd be better of with a waitqueue instead of a completion. Regards Oliver -- 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
