On 10/1/07, Benjamin Herrenschmidt <benh@kernel.crashing.org> wrote:
quoted text
>
> On Sun, 2007-09-30 at 16:57 -0600, Grant Likely wrote:
> >         val |= ACE_CTRL_DATABUFRDYIRQ | ACE_CTRL_ERRORIRQ;
> >         ace_out(ace, ACE_CTRL, val);
> >
> > +       /* Now we can hook up the irq handler */
> > +       if (ace->irq != NO_IRQ) {
> > +               rc = request_irq(ace->irq, ace_interrupt, 0,
> > "systemace", ace);
> > +               if (rc) {
> > +                       /* Failure - fall back to polled mode */
> > +                       dev_err(ace->dev, "request_irq failed\n");
> > +                       ace->irq = NO_IRQ;
> > +               }
> > +       }
> > +
>
> I don't know the HW but from the above, it looks like you enable
> interrupt emission on the HW before you register the handler, which is
> wrong. You should make sure on the contrary that IRQs on the HW are
> disabled until after you have registered a handler.
>
> Only really a problem if you have shared interrupts but still...

Yeah, you're right.  Fortunately all current in-tree platforms which
use this do not have shared interrupts, but I'd like to be correct on
this.

I'll tidy this up and send a fixup patch.

Thanks,
g.

-- 
Grant Likely, B.Sc., P.Eng.
Secret Lab Technologies Ltd.
grant.likely@secretlab.ca
(403) 399-0195
-
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/