Re: PCI patch for 2.3.21

Linus Torvalds (torvalds@transmeta.com)
Thu, 14 Oct 1999 07:57:01 -0700 (PDT)

On Thu, 14 Oct 1999, Pavel Machek wrote:
>
> Even when PCI device is trying to interrupt you (level triggered) you
> can disable that interrupt. userirq driver is using this kind of dirty
> trick, and yes, I works.
>
> So, if uhci receives interrupts it does not know how to shut up, it
> can disable that interrupt to avoid hard lockup. [Well, with some
> sideeffects: like killing all devices on shared interrupt.]

That is STILL wrong. You'd have a dead mouse, just because the PCI
subsystem (incorrectly) enabled an interrupt route that it should not have
enabled.

Trust me, the rigth thing is to NOT have the PCI subsystem change things
it does not understand. Problem solved.

> We could do something like
>
> if (no driver handler interrupt, and 1000 of such interrupts came one
> after each other)
> printk( "Someone is interrupting us and we don't know what to do
> with it\n" );
> disable_irq(offending_irq);

Sure. Dead mouse.

Linus

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