On Tue, Dec 11, 2018 at 1:36 PM Nicolas Saenz Julienne <nsaenzjulienne@xxxxxxx> wrote: > On Mon, 2018-12-10 at 22:11 +0100, Arnd Bergmann wrote: > > @@ -447,26 +444,26 @@ remote_event_wait(VCHIQ_STATE_T *state, > > REMOTE_EVENT_T *event) > > } > > > > static inline void > > -remote_event_signal_local(VCHIQ_STATE_T *state, REMOTE_EVENT_T > > *event) > > +remote_event_signal_local(wait_queue_head_t *wq, REMOTE_EVENT_T > > *event) > > { > > event->armed = 0; > > - complete((struct completion *)((char *)state + event->event)); > > + wake_up_all(wq); > > Shouldn't this just be "wake_up(wq)"? I wasn't entirely sure if we could get with more than one thread waiting for the wakeup. With the semaphore or completion that would already be broken because we'd only wake up one of them, but I was hoping to stay on the safe side with wake_up_all(). Arnd _______________________________________________ devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxx http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel