[openib-general] Re: Disabling IRQ #201 message
Michael S. Tsirkin
mst at mellanox.co.il
Wed Feb 2 11:59:08 PST 2005
Quoting r. Roland Dreier <roland at topspin.com>:
> Subject: Re: [openib-general] Re: Disabling IRQ #201 message
>
> Michael> Why does it "totally break" it? We did suppress the
> Michael> interrupt (write to clear interrupt register) so we
> Michael> return "HANDLED". Its unavoidable that hardware may
> Michael> re-assert it while the handler is still running.
>
> Because the purpose of the IRQ_HANDLED return value is to tell the
> kernel that our device did indeed generate the interrupt. If the
> interrupt is shared with another device with a stuck interrupt then
> always returning IRQ_HANDLED from our handler will prevent the kernel
> from detecting the stuck interrupt.
>
> - R.
>
Roland, I know, but I believe thats a far cry from "totally break it".
Reson being, after the kernel disables a stuck interrupt, any device sharing it
will not function properly anyway.
--
MST - Michael S. Tsirkin
More information about the general
mailing list