On Thu, Oct 10, 2013 at 06:29:59PM +0100, Peter Ujfalusi wrote:
On 10/10/2013 07:59 PM, Mark Rutland wrote:
No, they're not actually of much practical use to us at the minute but it was generally felt better to include the information and not use it so that if someone does come up with a use for them then the trees for deployed systems already have the information.
Sure, but if this device can generate multiple interrupts, we should make it possible to describe all of them, unambiguously.
This is why Jyri added them to the DT. They are not used by the Linux driver, but the HW have interrupt lines (two of them: TX and RX).
The binding only describes a single interrupt, and even if multiple interrupts were listed, there's no way to disambiguate them (e.g. interrupt-names).
It would be nice to remedy this.
Cheers, Mark.