o Don't hold lock over bus_teardown_intr when we're waiting for the

thread to die, since that can cause deadlock with the ISR (in theory).
o Minor cleanup
o comment nit
This commit is contained in:
Warner Losh 2007-05-16 19:53:49 +00:00
parent af3bdf60d5
commit 26432f7ba2

View File

@ -332,13 +332,13 @@ cbb_detach(device_t brdev)
/*
* Wait for the thread to die. kthread_exit will do a wakeup
* on the event thread's struct thread * so that we know it is
* save to proceed. IF the thread is running, set the please
* safe to proceed. IF the thread is running, set the please
* die flag and wait for it to comply. Since the wakeup on
* the event thread happens only in kthread_exit, we don't
* need to loop here.
*/
mtx_lock(&sc->mtx);
bus_teardown_intr(brdev, sc->irq_res, sc->intrhand);
mtx_lock(&sc->mtx);
sc->flags |= CBB_KTHREAD_DONE;
while (sc->flags & CBB_KTHREAD_RUNNING) {
DEVPRINTF((sc->dev, "Waiting for thread to die\n"));
@ -469,8 +469,8 @@ cbb_event_thread(void *arg)
mtx_lock(&sc->mtx);
sc->flags |= CBB_KTHREAD_RUNNING;
mtx_unlock(&sc->mtx);
while ((sc->flags & CBB_KTHREAD_DONE) == 0) {
mtx_unlock(&sc->mtx);
/*
* We take out Giant here because we need it deep,
* down in the bowels of the vm system for mapping the
@ -527,10 +527,8 @@ cbb_event_thread(void *arg)
while (err != EWOULDBLOCK &&
(sc->flags & CBB_KTHREAD_DONE) == 0)
err = cv_timedwait(&sc->cv, &sc->mtx, hz / 4);
mtx_unlock(&sc->mtx);
}
DEVPRINTF((sc->dev, "Thread terminating\n"));
mtx_lock(&sc->mtx);
sc->flags &= ~CBB_KTHREAD_RUNNING;
mtx_unlock(&sc->mtx);
kthread_exit(0);