Re: Problem with 2.6.17-mm2

From: Ingo Molnar
Date: Sun Jun 25 2006 - 15:50:21 EST



* Thomas Gleixner <tglx@xxxxxxxxxxxxx> wrote:

> On Sun, 2006-06-25 at 10:32 -0700, Andrew Morton wrote:
> > > So in fact this just silently acks spurious interrupts which have an
> > > hw_irq_controller assigned. If there is no action, then nothing has
> > > called setup_irq/request_irq for this interrupt line and therefor it is
> > > an spurious interrupt which should not happen.
> > >
> > >
> > > genirq makes these visible and informs noisily about those events.
> > >
> >
> > hm, OK. I guess we can let it ride for now. Later we can decide whether
> > we need to shut that warning up. I suspect we should, if the machine's
> > working OK.
>
> We can make it once per IRQ.

yeah. A bit more sophisticated method would be to use a new sticky
IRQ_SPURIOUS bit and only print a warning if it goes from 0 to 1.
Whenever a real handler is installed the bit gets cleared. This will
make behavior a bit more deterministic than 'once per bootup', and it
will still not spam the box with printks. (Or never let that bit go from
1 to 0 - this effectively implements the once-per-bootup warning.)

> In fact I think the original behaviour is a BUG. You have no chance to
> notice that your box gets flooded by such interrupts. With my
> willingly asserted spurious interrupts the box simply stalls in a
> flood of interrupts without any notice.

hm, doesnt note_interrupt()/irqpoll detect and handle this to a certain
degree?

Ingo
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/