Re: Multiple MSI

[Benjamin Herrenschmidt <benh@kernel.crashing.org>]

> Some years ago, we had discussions about getting rid of IRQ numbers
> altogether, or at least the requirement to have device drivers know
> about them. Does anyone remember what happened to that idea?

I think it's not totally dead. Last I heard, someone (jgarzik ?) was
slowly, bit by bit, removing the dependencies on the irq argument on irq
handlers which is one step in the direction.

> I think the concept was that you pass around struct irq_desc pointers
> that may or may not be dynamically allocated by the interrupt controller
> code.

Yup. There are still a few hard dependencies on numbers left and right
tho. The main issue is old userspace tied to the layout of things
like /proc/interrupts though I'd be happy to special case the 16
"legacy" interrupts (like we do on powerpc in our remapping layer) and
only show these here ...

> Another simplification that should really help here is encapsulating
> request_irq() per subsystem so that you can do something like
> 
> int my_irq(struct pci_dev *dev);
> err = pci_request_irq(pci_dev, &my_irq, IRQF_SHARED);
> 
> Most PCI drivers should be trivial to convert to this model. If you want
> to have multiple MSI/MSI-X interrupts for one PCI device with this model,
> you'd need to introduce the number back as an offset, I guess.

Might indeed be a good idea.

Cheers,
Ben.